Comparison of A2E cytotoxicity and phototoxicity with all-trans-retinal in human retinal pigment epithelial cells

Article abstract of DOI:10.1111/j.1751-1097.2010.00750.x

All-trans-retinal is the precursor of A2E, a fluorophore within lipofuscin, which accumulates in human retinal pigment epithelial (hRPE) cells and contributes to age-related macular degeneration. Here we have compared the in vitro dark cytotoxicity and vi

Full text of DOI:10.1111/j.1751-1097.2010.00750.x

Photochemistry and Photobiology, 2010, 86: 781–791
Comparison of A2E Cytotoxicity and Phototoxicity with all-trans-Retinal in
Human Retinal Pigment Epithelial Cells†
Albert R. Wielgus*1, Colin F. Chignell1, Patricia Ceger1 and Joan E. Roberts2
1Laboratory of Pharmacology, NIEHS, Research Triangle Park, NC
2Department of Natural Sciences, Fordham University, New York, NY
Received 29 December 2009, accepted 30 March 2010, DOI: 10.1111 ⁄ j.1751-1097.2010.00750.x
rim of the photoreceptor disk (3). There, all-trans-retinal is
reduced to the less reactive all-trans-retinol by NADPH-
ABSTRACT
All-trans-retinal is the precursor of A2E, a fluorophore within
lipofuscin, which accumulates in human retinal pigment epithelial
(hRPE) cells and contributes to age-related macular degenera-
tion. Here we have compared the in vitro dark cytotoxicity and
visible-light-mediated photoreactivity of all-trans-retinal and
A2E in hRPE cells. All-trans-retinal caused distinct cytotoxicity
in hRPE cells measured with cell metabolic activity (MTS) and
lactate dehydrogenase release assays. Significant increases in
intracellular oxidized glutathione (GSSG), extracellular GSH
and GSSG levels and lipid hydroperoxide production were
observed in cells incubated in the dark with 25 and 50 lM all-
trans-retinal. Light modified all-trans-retinal’s harmful action
and decreased extracellular glutathione and hydroperoxide levels.
A2E (<25 lM) did not affect cell metabolism or cytoplasmic
membrane integrity in the dark or when irradiated. 25 lM A2E
raised the intracellular GSSG level in hRPE cells to a much
smaller extent than 25 lM all-trans-retinal. A2E did not induce
glutathione efflux or hydroperoxide generation in the dark or
after irradiation. These studies support our previous conclusions
that although A2E may be harmful at high concentrations or when
oxidized, its phototoxic properties are insignificant compared to
those of all-trans-retinal. The endogenous production of A2E may
serve as a protective mechanism to prevent damage to the retina
by free all-trans-retinal.
dependent all-trans-retinol dehydrogenases (RDHs) such as
RDH8, RDH11 and RDH12 (1,2,4,5,7,10). Photoreceptors are
unable to reisomerize all-trans-retinal or all-trans-retinol back
to 11-cis-retinal (11). Thus, all-trans-retinol is transported to
the retinal pigment epithelial (RPE) cells, esterified by lecithin-
retinol acyltransferase, recycled to 11-cis-retinal and translo-
cated back to the photoreceptors (5).
The RPE cells form a single-cell layer that is the most distal
layer of the retina. The role of the RPE is not limited to only
the recycling processes of the visual cycle. The cells also
transport ions, water and metabolic end products from the
subretinal space to the blood (11), and deliver nutrients such as
glucose, retinol and fatty acids in the opposite direction—from
the blood to the photoreceptors. Moreover, the RPE cells
phagocyte light-damaged photoreceptor outer segments in a
circadian regulated process, and then shed, digest, recycle and
transport the essential elements of their contents back to the
photoreceptors (11,12).
All-trans-retinal in its free form is a highly reactive
molecule. When it is released from photoactivated rhodopsin
it can react with phosphatidylethanolamine (PE), a primary
amine that is present in membranes of the photoreceptor outer
segments at relatively high abundance. A fraction of the all-
trans-retinal-PE adduct can undergo further condensation
leading to formation of 2-[2,6-dimethyl-8-(2,6,6-trimethyl-1-
cyclohexen-1-yl)-1E,3E,5E,7E-octatetraenyl]-1-(2-hydroxyeth-
yl)-4-[4-methyl-6-(2,6,6-trimethyl-1-cyclohexen-1-yl)-1E,3E,
5E-hexatrienyl]-pyridinium (A2E) (13). The precursor of the
A2E chromophore is transported to the RPE cells in phago-
cytosed photoreceptor outer segments. It can be also synthe-
sized in the retinal pigment epithelial cells and removed (14).
However, with age, it is observed that A2E and other
fluorescent chromophores (15) accumulate in human retinal
pigment epithelial (hRPE) cells in the form of granules called
lipofuscin. In advanced age, the concentration of A2E can
reach 20 lM (14,16,17). Lipofuscin accumulation increases
particularly in the elderly and is considered to be one of the
causative factors of age-related macular degeneration (AMD).
This disease is a major cause of visual impairment in the United
States and the leading cause of legal blindness for people over
the age of 65. It is estimated that more than 1.5% of
Caucasians at age more than 70 years old suffer advanced
INTRODUCTION
Retinoids—retinol, retinaldehyde and retinyl esters—are vita-
min A metabolites that are major constituents in the vertebrate
visual cycle (1–4). Light that reaches the rod or cone outer
segments in the retina induces photoisomerization of 11-cis-
retinal, a visual chromophore combined with rhodopsin to all-
trans-retinal, which accumulates in the photoreceptor outer
segments (4–6). Most of the all-trans-retinal (atRal) is released
from opsin into the disk lumen (5), after which the chromo-
phore is transported from the inside of disk membranes into
the cytoplasm (5,7) by the photoreceptor specific ATP-binding
cassette transporter (ABCA4 ⁄ ABCR) (5,8,9) localized in the
†This invited paper is part of the Symposium-in-Print: ‘‘Phototoxicity of the Skin
and Eye,’’ in honor of Dr. Colin Chignell.
*Corresponding author email: albert.wielgus@duke.edu (Albert R. Wielgus)
ꢀ 2010 U.S. Government. Journal Compilation. The American Society of Photobiology 0031-8655/10
781

782 Albert R. Wielgus et al.
(SSA) were purchased from Sigma (St. Louis, MO), ethanolamine,
HClO4, potassium chromate, triphenylphosphine (TPP) and xylenol
orange (XO) from Aldrich (Milwaukee, WI) and sodium carbonate
anhydrous from Mallinckrodt (Paris, KY). All reagents used for
experiments were at least of analytical grade.
Synthesis of A2E. A2E was synthesized as described previously (27).
Briefly, a mixture of all-trans-retinal and ethanolamine (2:1 ratio) was
incubated at room temperature for 2 days. The postsynthesis mixture
was purified by silica gel column chromatography and then by HPLC.
The purity of A2E was checked using HPLC and thin layer
chromatography (TLC) methods. The A2E concentration was deter-
mined by absorption spectroscopy (27) using a Hewlett Packard diode
array 8453 spectrophotometer (Hewlett Packard GmbH, Waldbronn,
Germany). Dry A2E samples were stored under argon at )70ꢁC for
further use.
Cell cultures. The human RPE cells were isolated from donor eyes
using the method of Hu et al. (30) and cultured in Falcon tissue culture
B10 cm dishes (Becton Dickinson Labware, Franklin Lakes, NJ) with
F-12 nutrient mixture containing 1.5 mM L-glutamine and supple-
mented with 10% fetal bovine serum (FBS), and 50 lg mL)1 genta-
micin (Gibco, Grand Island, NY). Cells were in the 15th–22nd
passage. They were detached by 0.125% trypsin solution (Gibco),
diluted 1:3–1:4 and plated for subculture in 96-well plates (Corning
Incorporated, Corning, NY) for cell viability and cytotoxicity assays,
and in the Falcon dishes for the remaining experiments.
AMD and another 10% have symptoms of early macular
degeneration (18).
Most of the all-trans-retinal molecules formed in the
photoreceptor outer segments are reduced to all-trans-retinol
or undergo condensation processes before they reach RPE cells
during phagocytosis. However, it was postulated that intensive
illumination would rapidly generate substantial amounts of
all-trans-retinal, which then would be present in phagocytosed
outer segment disks in the RPE cells (19). The results of the
most recent studies performed on Rdh8) ⁄ )Abca4) ⁄ ) mice
suggest that loss of the all-trans-RDHs, RDH8 and RDH12
enzymes, which facilitate conversion of all-trans-retinal to all-
trans-retinol in the RPE, can cause atRal to excessively
accumulate in the photoreceptors and diffuse into the RPE
(5), where it can be stored in lipofuscin (20). All-trans-retinal is
also synthesized inside the RPE cells. The oxidative cleavage of
b,b-carotene catalyzed by b,b-carotene-15,15¢-dioxygenase
leads to atRal formation under dark or photic conditions
directly in the RPE (21). Additionally, atRal is synthesized in
the RPE cells from all-trans-retinol of the retinylester pool (11)
and plays a role as a substrate of retinal G protein-coupled
receptor (RGR), a microsomal opsin present in the RPE cells
(21). It was shown that all-trans-retinol is oxidized efficiently
by a specific retinol dehydrogenase and that significant
amounts of all-trans-retinal accumulate in ARPE-hRGR and
bovine RPE cells in the absence of light (21).
The hRPE cells used in the present study consisted of one cell line of
pure culture of cells in active growth status. The purity of the cell line was
demonstrated by immunocytochemical methods: hRPE cells display
S-100 and cytokeratin, uveal melanocytes display S-100 antigen but not
cytokeratin, and fibroblasts display neither of these proteins (31).
The average cell density was (43.0
96-well plate and (2.8
5.6)Æ103 cells per well in the
0.8)Æ106 cells per dish.
Cell incubation with all-trans-retinal and A2E. All-trans-retinal and
A2E were reconstituted in DMSO and diluted with the medium to the
required concentration immediately before addition to the cell culture.
In most experiments, the cells were incubated in the dark with 10–
25 lM A2E, which mimics the highest concentration of accumulated
A2E in the elderly (16) or with 20–50 lM all-trans-retinal. Although the
concentrations of atRal and A2E were comparable in all experiments,
we also incubated the cells in medium containing 50 lM all-trans-
retinal. We used an atRal concentration twice as high as the maximal
concentration of A2E in order to test the potential significance of all-
trans-retinal condensation and subsequent A2E synthesis for protec-
tion of the RPE cells against cytotoxic and phototoxic effects of atRal,
as one A2E molecule is a condensation product of two all-trans-retinal
molecules. The DMSO concentration was 0.3% in all treatment
solutions including control samples, which were incubated in the
medium without any chromophore. Samples were maintained at 37ꢁC
in an air atmosphere containing 5% CO2 for 0–24 h to determine dark
toxicity and for 5 h prior to light exposure. The amounts of all-trans-
retinal and A2E incorporated into the cells and the total lipid contents
were determined by TLC.
Adult human retinal pigment epithelium and the anterior
layers of the retina constantly receive intense visible light
(>400 nm) (22). There are several endogenous RPE chromo-
phores including melanin, the mitochondrial enzyme cyto-
chrome-c oxidase and riboflavin that can act as potential
photosensitizers that may be responsible for visible light
short-wavelength damage to the RPE (23). However, their
photoreactivity is negligible compared with lipofuscin or all-
trans-retinal (3,24–26). A2E is one of the retinoid chromo-
phores isolated from ocular lipofuscin and has been reported to
be a potential photosensitizer (23). A2E absorbs in the visible
range of light and is characterized by two maxima at 336 and
439 nm (27). Although the absorbance spectrum of all-trans-
retinal is located mostly in UV with a maximum at 380 nm, it
partially absorbs in the visible light range up to ꢀ450 nm (25).
We report here a series of experiments comparing the
toxicity and phototoxicity of two retinal chromophores, A2E
and all-trans-retinal. The results of experiments carried out
in vitro proved that although both chromophores produce
superoxide anion during irradiation with visible light, A2E
does so less efficiently (26,28). Both chromophores exposed to
UV irradiation also generate singlet oxygen (14,26,28,29), but
Light exposure. The hRPE cells were irradiated directly in the
covered plates in which they had been grown by two fluorescent bulbs
(Philips F40AX50, 40 W, 5000 K; Advantage X, Somerset, NJ),
through a horizontal liquid filter (aqueous solution of NaNO2
[50 g L)1], K2CrO4 [0.2 g L)1] and Na2CO3 [1 g L)1]) with a 1 cm
pathlength transmitting wavelengths longer than 400 nm. Cells were
irradiated for 0–20 min at a light intensity of 6.1 W m)2
.
1
Cell metabolic activity and cell membrane permeability assays. After
the specified incubation time, the nutrient mixture containing 10%
FBS was exchanged for medium supplemented with 2% FBS during
the all-trans-retinal dark toxicity experiments or for PBS in the plates
containing cells assigned for irradiation. After light exposure, PBS was
replaced with the 2% FBS medium. The cells were incubated in the
dark at 37ꢁC in a 5% CO2 atmosphere for about 18 h. Cell metabolic
activity and cell membrane permeability were determined using MTS
(The CellTiter 96ꢂ AQueous Non-Radioactive Cell Proliferation Assay;
Promega Corp., Madison, WI) and lactate dehydrogenase (LDH)
assay (The CytoTox 96ꢂ Non-Radioactive Cytotoxicity Assay;
Promega Corp.), respectively.
the O2 photoproduction yield of A2E is at least one order of
magnitude lower than that of its precursor. In addition, it was
found that A2E has antioxidant properties. It quenches singlet
oxygen with about the same efficiency as vitamin E (14).
We present here a comparison of the in vitro visible-light
photoreactivity of A2E and its precursor all-trans-retinal and
dark cytotoxicity of both of these chromophores in hRPE cells.
MATERIALS AND METHODS
Absorption spectra. After 5 h incubation with 20, 30 or 50 lM atRal
or 25 lM A2E in 10 cm petri dishes, the medium was collected and
centrifuged at 140 g and 22ꢁC for 6 min. Each cell pellet was
Reagents. All-trans-retinal, bovine albumin (fraction V), butylated
hydroxytoluene (BHT), cumene hydroperoxide, dimethyl sulfoxide
(DMSO) sterile, FeSO4Æ7H2O, sodium nitrite and 5-sulfosalicylic acid

Photochemistry and Photobiology, 2010, 86 783
resuspended in 5 mL PBS and the cell suspensions were transferred to
the corresponding cell dishes. The cells were exposed to visible light or
stored in the dark for 20 min as described above. They were then
scraped off, transferred to 15 mL tubes and centrifuged at 3000 g and
4ꢁC for 10 min. The supernatant was removed and 1 mL of it was left in
each tube. A half milliliter of each cell suspension was extracted with
800 lL of CHCl3 ⁄ CH3OH (2:1) mixture for 4 min. The remaining cell
suspension was used for protein content determination. The entire
lower (organic) layer (ꢀ600 lL) was transferred to empty tubes and
evaporated in N2. Each sample was reconstituted in 500 lL CH3OH.
Absorption spectra were recorded in a 0.5 cm spectrophotometric cell
using a Hewlett Packard diode array 8453 spectrophotometer (Hewlett
Packard GmbH).
Preparation of cell samples for glutathione and hydroperoxide
determination. Each medium containing one of the chromophores
was replaced with 5 mL of PBS after incubation in the dark. Directly
after light exposure, 325 lL of PBS supernatant was taken from each
dish and transferred to a tube containing 325 lL of 5% solution of
SSA. The tube contents were mixed intensively and placed on ice for
protein precipitation. After 5 min incubation the samples were
centrifuged at 10 000 g and 4ꢁC for 5 min. Two 250 lL volumes of
each supernatant were placed in separate tubes for the extracellular
glutathione assay—total (GSx) and the oxidized form (GSSG),
respectively. The samples were stored at )80ꢁC until the assay. The
remaining volume of PBS was removed from each cell plate and
replaced with 0.5 mL HCl (10 mM). The cells were scraped off, placed
in separate tubes and diluted with the HCl solution to 1 mL. A half
milliliter of each cell suspension was transferred to separate tubes
containing 0.8 mL of extraction mixture (50 lM solution of BHT in
CHCl3 ⁄ CH3OH [2:1]), intensively shaken for a minute and centrifuged
as described previously. 700 lL of the lower (chloroform) phase was
transferred from each tube to an empty one, evaporated in a nitrogen
stream and stored at )80ꢁC until hydroperoxide determination. The
remaining volume of each cell suspension was sonicated (Ultrasonic
Homogenizer 4710 Series; Cole-Parmer Instruments Co., Chicago, IL)
for 10 s and centrifuged as described previously. 50 lL aliquots of each
supernatant were placed in empty tubes and the samples were stored at
)80ꢁC for the protein assay. To prepare samples for intracellular
glutathione determination, 325 lL of each cell supernatant was
transferred to separate tubes containing 325 lL of 5% SSA and
treated in the same way as the samples designated for external
glutathione determination.
the samples was measured at 560 nm in a GENios plate reader
(TECAN Austria GmbH, Grodig ⁄ Salzburg, Austria). Bovine albumin
was used as a standard.
Statistical analysis. Data in graphs are presented as average
of four to six measurements.
SD
RESULTS
Dark cytotoxicity
Preliminary experiments determined that all-trans-retinal
(atRal) and A2E were incorporated into human RPE cells as
monitored by TLC (data not shown). All-trans-retinal signif-
icantly impaired the metabolism of hRPE cells at concentra-
tions higher than 20 lM (Fig. 1a) in a dose-dependent manner
as measured by the MTS assay. After a 3 h treatment at 20 lM
atRal there was some decrease in viability (ꢀ20% after 24 h),
at 40 lM there was a larger (ꢀ40%) decrease and at 60 lM
metabolic activity was almost complete inhibited.
Determination of glutathione. Both extracellular and intracellular
concentrations of total and oxidized (GSSG) glutathione were deter-
mined using the assay described previously (32). The concentration of
the reduced form (GSH) was calculated in accordance with the
equation: GSH = GSx ) 2 GSSG. The cellular redox state of intra-
cellular glutathione was expressed as the ratio of GSH ⁄ GSSG.
Hydroperoxide assay. The modified ferric-xylenol orange (FOX)
assay was used for quantifying the concentrations of hydroperoxides
(33–35). Nonlipid hydroperoxides were determined after incubation of
samples with TPP (34,35).
Each sample of the evaporated chloroform phase was dissolved in
25 lL of methanol and stored on dry ice. Ten microliters of each
solution was added to 90 lL of methanol or 11.1 mM of TPP solution in
CH3OH to determine all ROOH and nonlipid hydroperoxides, respec-
tively. The samples were mixed and incubated for 30 min at room
temperature in the dark. To prepare the FOX reagent solution, 2.78 mg
of FeSO4Æ7H2O was dissolved in 1 mL of XO (2.5 mM) solution in
HClO4 (1.1 M). Then 18 mL of BHT (4.44 mM) solution in methanol,
1.5 mL of the XO solution in HClO4 and 0.5 mL of the iron (II) solution
were mixed together in a vial. At the end of the incubation time 0.9 mL
of the FOX reagent was added to each sample and stored for another
30 min. Thereafter, the samples were mixed and centrifuged at 2400 g
and 22ꢁC for 10 min. The absorbance of the samples was measured at
560 nm using a Hewlett Packard diode array 8453 spectrophotometer
(Hewlett Packard GmbH). Cumene hydroperoxide was used as a
standard for a calibration curve in each assay. Lipid hydroperoxide
(LOOH) concentration was calculated by subtracting the nonlipid
hydroperoxide concentration from the total ROOH concentration.
Protein determination. Protein concentration was determined using
a Protein Assay Kit (Pierce, Rockford, IL) containing bicinchoninic
acid. The samples assigned for protein determination were diluted with
PBS. The assays were carried out in 96-well plates. The absorbance of
Figure 1. Cytotoxic effect of all-trans-retinal (atRal) on human RPE
cells. Cells cultured in a 96-well plate were incubated in the dark up to
24 h in medium supplemented with 0–60 lM atRal. Metabolic activity
of the cells and cytoplasmic membrane integrity were monitored using
the MTS assay (a) and lactate dehydrogenase (LDH) assay (b).
Metabolic activity of the cells incubated with atRal was normalized to
the sample containing no retinal and incubated for the same time. Cell
membrane permeability expressed as the concentration of LDH
released from the cells was normalized to the total concentration of
LDH in cells treated in the same way and lysed before the assay.

784 Albert R. Wielgus et al.
All-trans-retinal treatment of the hRPE cells did induce
release of LDH from the cells (Fig. 1b) and was also dose
dependent from 20 through 60 lM. At the maximum concen-
tration tested (60 lM) LDH efflux was observed in the cells just
after an hour, with total release of LDH within 3 h of
treatment.
However, when the hRPE cells were incubated with 10–
25 lM A2E, no significant changes in cell metabolic activity or
LDH efflux from the cells were observed even after 24 h of
treatment (data not shown).
recovered while those treated with the higher concentration of
the retinoid recovered to a level of 50%. A very slight increase
in metabolic activity was observed in the human RPE cells
preincubated with 50 lM atRal and exposed to visible light for
20 min.
Cytoplasmic membrane permeability, as monitored with the
LDH assay, showed changes corresponding to metabolic
activity in the hRPE cells preloaded with atRal and exposed
to visible light (Fig. 2b). The LDH efflux from cells pretreated
with 20 lM atRal increased from 40% detected in cells stored
in the dark to ꢀ50% in samples irradiated for 2 min and
decreased to 20% during prolonged irradiation. A similar
effect was observed in the cells pretreated with 30 lM
atRal—nearly 80% of LDH was released from the cells in
the dark, while the LDH level increased in cells exposed to
light for 2 min dropped to ꢀ50% in cells irradiated for a
longer time. Preincubation of human RPE cells with 50 lM
atRal in the dark induced significant cell membrane perme-
ability, which caused the release of almost all the LDH from
the cells. Samples of the cells pretreated with 50 lM atRal and
irradiated longer than 2 min showed a lower level of LDH
released from the cells to their surroundings.
Effect of light on cell metabolism and cytoplasmic membrane
integrity
We further investigated the effect of visible light exposure on
the metabolism and cytoplasmic membrane permeability of
hRPE cells with 0, 20, 30 and 50 lM all-trans-retinal. Incuba-
tion of hRPE cells for 5 h in the dark caused a decrease in their
metabolism by ꢀ20%, ꢀ90% and even more when the
medium contained 20, 30 and 50 lM all-trans-retinal, respec-
tively (Fig. 2a). A 2 min light (k > 400 nm) exposure of the
cells preincubated with 20 lM atRal diminished their metabolic
activity to ꢀ50% and almost stopped it in cells pretreated with
30 lM atRal. Interestingly, prolonged irradiation lessened the
harmful action of retinal. Metabolism of the cells preincubated
with 20 lM atRal and irradiated for up to 20 min totally
However, A2E incorporated into hRPE cells showed neither
detectable cytotoxicity in the dark nor when irradiated with
visible light at k > 400 nm (Fig. 2c,d), with either the MTS or
LDH assays.
Figure 2. Metabolic activity and cytoplasmic membrane permeability in human retinal pigment epithelial cells pretreated with 0–50 lM all-trans-
retinal (atRal) (a, b) or 0–25 lM A2E (c, d) and exposed to visible light. The cells cultured in 96-well plates were incubated in medium containing
atRal or A2E for 5 h. The medium was then replaced with PBS and the cells were exposed to light (k > 400 nm, 6.1 W m)2) for up to 20 min.
After irradiation, the cells were incubated overnight in medium containing 2% FBS. Metabolic activity of the cells and their cytoplasmic membrane
permeability were determined with the MTS assay (a) and lactate dehydrogenase (LDH) assay (b). In each series of different irradiation time
experiments, the metabolic activity of the cells was normalized to the sample containing no retinal. LDH release in each sample was normalized to
cells treated and irradiated in the same way but lysed before the assay.

Photochemistry and Photobiology, 2010, 86 785
Table 1. Spectrophotometric parameters of all-trans-retinal and its
products extracted from hRPE cells incubated with atRal for 5 h and
exposed to visible light or stored in the dark for 20 min.
catRal (lM) Treatment kp (nm)
Ap
kr (nm)
Ar
Ar ⁄ Ap
20
20
30
30
50
50
Dark
Light
Dark
Light
Dark
Light
325
325
325
325
334
332
0.481
0.466
0.664
0.610
1.140
0.945
382
382
382
382
382
382
0.142
0.122
0.280
0.247
0.940
0.757
0.30
0.26
0.42
0.40
0.82
0.80
catRal = atRal concentration used for incubation with cells; kr, kp =
wavelengths of the broad band maximal absorbances (Ar and Ap) of
atRal and its product absorption spectra, respectively.
fraction. The absorption spectrum of the extract from the
cells incubated with 30 lM atRal was generally similar to that
obtained after the 20 lM retinal incubation. However, the
extract of cells treated with 50 lM atRal was characterized by a
short-wavelength band with the maximum at 332–334 nm. The
absorbance of the long-wavelength one was relatively much
higher (A382 ⁄ A334 = 0.82) in comparison with the lower atRal
concentration extracts—A382 ⁄ A325 was 0.30 and 0.42 for the 20
and 30 lM atRal extracts, respectively (Table 1).
Light exposure of atRal-treated cells caused a decrease in
the extract absorbance normalized to protein content (Fig. 3a).
The strongest drop in the absorbance was observed in extracts
of cells incubated with 50 lL atRal. Simultaneously, the ratio
of absorbance at 382 nm to the short-wavelength band one
(Ar ⁄ Ap) decreased in each concentration in samples of cells
exposed to light compared with those stored in the dark
(Table 1).
The absorption spectrum of the extract of the cells
incubated with 25 lM A2E and stored in the dark was virtually
the same as the A2E standard (Fig. 3b). Light exposure caused
a slight absorbance decrease in both extract bands and a 5 nm
hypsometric (blue) shift.
Figure 3. Absorption spectra (normalized to protein content) of
extracts of human retinal pigment epithelial cells treated with 20
(dashed-dotted-dashed lines), 30 (dashed lines) or 50 lM (solid lines)
all-trans-retinal (atRal) (a) or 25 lM A2E (b). After treatment with a
retinoid, the cells were exposed to visible light (red lines) or stored in
the dark (black lines), collected and divided into two equal samples.
One of them was used to extract the retinoids in CHCl3 ⁄ CH3OH (2:1)
mixture. The lower (organic) layer was evaporated in gaseous nitrogen,
reconstituted in methanol and absorption spectra were recorded in a
spectrophotometric cell with 0.5 cm pathlength. The remaining cell
samples were used for determination of protein content with the
bicinchoninic acid assay. Standard solutions (green lines) of 20
(dashed-dotted-dashed) and 50 lM (solid) atRal and 25 lM A2E were
prepared in methanol and absorption spectra were recorded in the
same spectrophotometric cell.
Intracellular glutathione
Glutathione acts as an important molecule in the cellular
antioxidant system, with the ratio of its reduced to oxidized
form used as a factor in describing the cellular redox state.
A considerable increase in the concentration of the oxidized
form of glutathione inside the cells or an efflux from them can
serve to indicate a harmful effect of toxins or other environ-
mental factors on the cells.
When hRPE cells were incubated for 5 h in the presence of
25 lM atRal, the total glutathione concentration [ꢀ37nmol (mg
protein))1] in the cells was not affected in comparison with
Absorption spectra
The absorption spectrum of 20 lM all-trans-retinal in metha-
nol is characterized by one broad band with a maximal
absorbance at 382 nm (Fig. 3a). The chromophore absorbs
mostly in the UVA and in the short-wavelength range of
visible light. No significant shift in the maximal absorbance
wavelength was observed in a 50 lM atRal spectrum that could
indicate aggregation of the retinoid. However, after 5 h
incubation with 20 lM atRal, the absorption spectrum of the
chloroform ⁄ methanol extract from hRPE cells differed signif-
icantly from that of the initial chromophore. The main
absorption spectrum band was shifted toward short wave-
lengths and the maximal absorbance was at 325 nm, but the
original all-trans-retinal component remained as a small
control cells (Fig. 4a).
A
lowered level [ꢀ15 nmol (mg
protein))1] of intracellular GSx was observed only when the
cells were preincubated in medium containing 50 lM atRal and
not exposed to light. Cells irradiated with visible light did not
show changes in the cellular GSx level. When they were
preincubated with 25 lM atRal the GSx level was comparable
to that in hRPE cells untreated with all-trans-retinal. In cells
preincubated with 50 lM atRal a gradual increase in the GSx
level was observed in samples exposed for a longer time.
There was no significant effect of A2E at concentrations of
up to 25 lM on intracellular total glutathione concentration in

786 Albert R. Wielgus et al.
Figure 4. Concentration of intracellular total glutathione (GSx) (a), its reduced (GSH) (b) and oxidized (GSSG) (c) forms and redox state expressed
as a ratio of GSH ⁄ GSSG (d) in human retinal pigment epithelial cells preincubated with all-trans-retinal (atRal) or A2E and exposed to visible light.
Cells were incubated in the dark in medium containing 25 and 50 lM atRal or 12.5 and 25 lM A2E for 5 h. The medium was then replaced with PBS
and the cells were exposed to light (k > 400 nm) or stored in the dark (control) for 5 or 20 min. Cells were collected for internal total and oxidized
glutathione assays immediately after irradiation. The reduced glutathione concentration was calculated as GSH = GSx ) 2 GSSG. The
determined concentration of each form of glutathione was normalized to protein concentration in the same cell sample.
hRPE cells with either a 5 or 20 min light exposure or in those
cells retained in the dark (Fig. 4a). The changes in the level of
the reduced glutathione form inside the cells were similar to
those observed intracellular total glutathione (Fig. 4b).
In contrast, the presence of all-trans-retinal in hRPE cells
caused a substantial increase in the oxidized form of glutathi-
one (Fig. 4c). Incubation of the cells in a medium containing
25 lM atRal raised the intracellular GSSG level by about a
factor of 4 compared with the cells not fed with retinal. When
the cells were exposed to light for 2 min, the concentration of
the glutathione oxidized form decreased by ꢀ20%. Under
prolonged irradiation, the GSSG concentration in the cells
increased to its level in unexposed cells. Cells incubated in
medium containing 50 lM atRal and stored in the dark
exhibited a slight decrease in the GSSG level compared with
cells pretreated with 25 lM atRal. When cells preincubated
with 50 lM retinoid were exposed to light, the intracellular
GSSG concentration increased slightly.
incubation medium caused an almost three-fold increase in
intracellular GSSG concentration. Then, with a 20 min light
exposure, the level of oxidized glutathione decreased to one
comparable to that in cells not treated with A2E.
The cellular redox state
The cellular redox state of glutathione depended on the
concentration of a chromophore incorporated into hRPE cells
(Fig. 4d). Incubation of the cells in the presence of 25 and 50 lM
atRal led to a fall in the GSH ⁄ GSSG ratio by almost 70% and
more than 80%, respectively. During irradiation of the hRPE
cells, significant changes in the cellular redox state were not
observed in either the retinal-laden cells or the control ones.
In hRPE cells preincubated in medium containing 12.5 lM
A2E, the GSH ⁄ GSSG ratio did not change significantly
compared with the controls. However, the presence of 25 lM
A2E in the nutrient mixture decreased the cellular redox state
of glutathione by ꢀ70% in cells stored in the dark. Light
exposure caused an increase in the GSH ⁄ GSSG ratio in cells
pretreated with 25 lM A2E. Cells irradiated for 20 min had a
GSH ⁄ GSSG ratio ꢀ2.8 times higher than cells kept in the
dark.
In the cells fed A2E, the intracellular GSSG concentration
did not reach such a high level (Fig. 4c) as in the presence of
atRal. Cells incubated in medium containing 12.5 lM A2E
showed a level of oxidized glutathione similar to that of the
control cells. However, the presence of 25 lM A2E in the

Photochemistry and Photobiology, 2010, 86 787
Figure 5. Changes in extracellular glutathione determined in the supernatant of human retinal pigment epithelial (hRPE) cells pretreated with all-
trans-retinal (atRal) or A2E and exposed to visible light. Cells were incubated in the dark in medium containing 25 and 50 lM atRal or 12.5 and
25 lM A2E for 5 h. The medium was replaced with PBS and the cells were exposed to light (k > 400 nm) for 5 or 20 min or stored in the dark
(control). Directly after irradiation the cell supernatant was collected, treated appropriately as described in Materials and Methods and frozen.
Total glutathione (GSx) (a) and its oxidized form GSSG (c) were determined with the DTNB assay. Concentration of the glutathione reduced form
(GSH) (b) was calculated in accordance with the equation: GSH = GSx ) 2 GSSG. The concentration of each glutathione form was normalized
to protein concentration in the same sample of hRPE cells.
The presence of A2E in hRPE cells did not cause a
Extracellular glutathione
glutathione efflux from the cells even when they were irradi-
ated with visible light for 20 min (Fig. 5).
In samples containing hRPE cells incubated in the dark the
levels of the extracellular total, reduced and oxidized forms of
glutathione rose rapidly at the higher concentrations of atRal
(Fig. 5). For 50 lM atRal the levels of both the total and
oxidized glutathione forms were ꢀ30 times higher (Fig. 5a,c),
while GSH increased by a factor of ꢀ14 (Fig. 5b) compared
with controls. When cells previously incubated in medium
containing 25 lM atRal were irradiated with visible light, an
initial increase was generated in extracellular glutathione
concentration (total glutathione—more than 50% and twofold
in GSSG) (Fig. 5a,c). The level of GSH rose only slightly
(Fig. 5b). Prolonged exposure decreased the level of all forms
of glutathione and after 20 min of illumination, GSx and
GSSG concentrations returned to values comparable with the
initial ones (Fig. 5a,c) while the GSH level decreased by half
(Fig. 5b). The extracellular amounts of all forms of glutathi-
one decreased monotonically when the cells preincubated with
50 lM atRal were irradiated for a longer time period. Never-
theless, after 20 min irradiation, the concentrations of all
extracellular glutathione forms remained at a level about seven
times higher than in the control cells.
Generation of hydroperoxides
Under oxidative stress conditions, a significant amount of
hydroperoxides can be produced in cells. We observed that
similar to extracellular glutathione, in hRPE cells incubated
in the dark the total concentration of hydroperoxides and
the amounts of LOOH rose dramatically (seven to eightfold)
in the presence of 50 lM all-trans-retinal (Fig. 6a). The cells
exposed to light, both pretreated with 25 lM atRal and
controls, showed a slight decrease in the hydroperoxide level
with increasing irradiation time.
A similar and much
stronger effect was observed in cells preincubated in the
medium containing 50 lM all-trans-retinal. Nevertheless,
total and lipid hydroperoxide concentration after 20 min
light exposure was still several times higher than the control
samples or cells containing a lower concentration of all-
trans-retinal and illuminated for the same time periods.
The amount of hydroperoxides formed during irradiation of
hRPE cells with incorporated A2E was negligible compared

788 Albert R. Wielgus et al.
free form and the cell metabolism monitored with the MTT
assay (5,23). Another study performed on isolated mitochon-
dria determined that retinal as a product of b-carotene
cleavage inhibited oxidative phosphorylation by 12% in
isolated mitochondria at concentrations as low as 1 lM
(36). Additionally, all-trans-retinal inhibited mitochondrial
oxidation and uncoupled oxidative phosphorylation at con-
centrations >50 lM in mitochondria isolated from rat hearts,
whereas A2E did not affect these metabolic functions even at
a concentration of 600 lM (5).
All-trans-retinal in its free form has a high affinity with
membranes, including cytoplasmic membranes, and may alter
their stability (5). It can react via a Schiff base linkage with
primary amines (23) present in membrane phospholipids and
proteins. All-trans-retinal binding to those basic components
of the cellular membrane can modify the structure of the
membrane and make it more permeable. Simultaneously, the
all-trans-retinal structure can undergo modification during
incubation with hRPE cells, which manifests in a dramatic
change in its absorption spectrum, particularly at low concen-
trations of the retinoid (Fig. 3). However, the cells show a
limited potential for atRal modification, as more unchanged
atRal remained at higher concentrations of the retinoid used
for incubation with hRPE cells.
A2E also exhibits hydrophobic properties, but its structure
does contain groups as reactive as those of all-trans-retinal.
A2E also accumulates in membranes of human RPE cells
cultured in vitro, but mostly in the lysosomes and to a lesser
extent in mitochondrial membranes (37). The RPE cells
cultured in vitro showed inhibition of the ATP-driven proton
pump of lysosomes when incubated with 2 lM A2E (38),
perturbation of cholesterol metabolism when treated with
15 lM A2E (17), impairment of mitochondrial ATP synthesis
and phagocytotic properties of the cells at 50 lM A2E (39) and
inhibition of phagolysosomal degradation of photoreceptor
phospholipids by RPE cells at 100 lM A2E (40).
At lower concentrations of A2E of up to 25 lM using the
MTS and LDH assay, we did not observe either an effect of
cellular dehydrogenase activity or an increase in the cytoplas-
mic membrane permeability in hRPE cells, both in the dark
and when exposed to visible light.
Figure 6. Hydroperoxide formation in human retinal pigment epithe-
lial (hRPE) cells pretreated with all-trans-retinal (atRal) or A2E and
exposed to visible light. The cells were incubated in the dark in medium
containing 25 and 50 lM atRal or 12.5 and 25 lM A2E for 5 h. The
medium was then replaced with PBS and the cells were exposed to light
(k > 400 nm) for 5 and 20 min. After irradiation, the cells were
collected and frozen. Total hydroperoxide concentration (a) and the
level of LOOH (b) in the cells were determined with the ferric-xylenol
orange assay and normalized to protein concentration in hRPE cells
treated in the same way.
with that induced by retinal, which was intercepted during cell
incubation in medium containing a 50 lM solution of all-trans-
retinal (Fig. 6a,b). However, the concentration of hydroper-
oxides generated in A2E-laden cells, both irradiated and kept
in the dark, was similar to that in the cells preincubated with
25 lM atRal.
Phototoxicity
All-trans-retinal, in addition to its normal physiological
function, is also an efficient photosensitizer (25,26). A photo-
toxic effect of all-trans-retinal was also observed in previous
studies when ARPE-19 cells were preincubated with a higher
(0.1 mM) retinal concentration in phosphatidylcholine lipo-
somes and exposed to visible light (41). At much lower
concentrations of free all-trans-retinal, irradiation with blue
light caused photodamage in the same type of RPE cells and
also in HEK293 cells derived from human kidneys (5).
All-trans-retinal-mediated photodamage to lipids and pro-
teins is oxygen-dependent (3,42,43). In vitro experiments
showed that all-trans-retinal exposed to UV or visible light
produced the retinoid free radicals in the presence of hydrogen
atoms or electron donors (reduced form of nicotinamide
adenine dinucleotide), superoxide anions and singlet oxygen
(20,25,26,28,44). The semioxidized retinoid can further interact
with tyrosine and cysteine, leading to protein oxidation (44).
DISCUSSION
Dark cytotoxicity
All-trans-retinal is synthesized and accumulates in the pho-
toreceptor outer segments and RPE cells (3). It was shown
that the concentration of all-trans-retinal released from
mouse opsin in vivo (5) is much higher than the range of
retinal concentrations used in this study. We found that there
was
a significant decline of metabolic activity and a
substantial increase in cytoplasmic membrane permeability
in hRPE cells in the presence of all-trans-retinal. Similar
cytotoxic effects of all-trans-retinal were observed when
ARPE-19 cells were incubated in the dark with atRal in its

Photochemistry and Photobiology, 2010, 86 789
All-trans-retinal or other retinoid chromophores can play
an important role in light-induced damage of photoreceptors
and RPE cells (3). In vitro studies showed that all-trans-retinal
exposed to visible light impairs the ABCR transporter present
in photoreceptor outer segments (42). Other studies performed
on rpe65) ⁄ ) mice indirectly proved that all-trans-retinal can
induce light-dependent degeneration in the retina. The absence
of 11-cis-retinal and all-trans-retinal in the photoreceptors
made the retina resistant to light toxicity (3).
with 407 nm light produced superoxide anion ꢀ4 times faster
than all-trans-retinal. However, the quantum efficiency for
O2•) generation by A2E determined in the study was very low
(0.0003), and was more than 30 times lower than for riboflavin
(26). Therefore, although irradiation of A2E also releases
reactive oxygen species (ROS), the quantum yield is insuffi-
cient to account for the amount of ROS produced by
lipofuscin.
Photochemical reactions of retinoids proceed through
several possible routes—photoisomerization, photodegrada-
tion, photo-oxidation and photopolymerization (20,45). The
absorption spectra of the extracts of hRPE cells incubated with
atRal slightly decreased and the shape of their absorption
spectra did not change significantly (Fig. 3). This result
excludes a retinal photaggregation process in the cells. On
the other hand, retinoids including all-trans-retinol and all-
trans-retinal undergo isomerization into a mixture of trans-
and cis- isomers (20). When retinoids are exposed to light they
oxidize, giving products such as retinol, 5,6-epoxide- and
4-keto-retinol, and 5,6-endoperoxide (20). The all-trans-retinal
absorption spectrum consists of a broad band in the range
280–480 nm with a maximal absorbance at 382 nm. However,
continuous irradiation of atRal in argon-saturated methanol
caused almost total bleaching of the chromophore and
formation of 9,10-dihydroretinal, which has an absorption
spectrum shifted to below 300 nm (25). It seems that bleaching
of all-trans-retinal accompanied by the formation of less
photoreactive products when irradiated with visible light is the
main mechanism of its decreasing phototoxic effects in RPE
cells.
Studies performed previously on the cytotoxic effect of all-
trans-retinal on ARPE-19 cells showed that atRal toxicity
was significantly inhibited when PE was present in liposomes
containing the chromophore (41). Moreover, the semioxi-
dized atRal-PE adducts are less damaging to proteins than
semioxidized atRal (44). During the processes leading to A2E
formation, which take place in the photoreceptor outer
segments, one atRal molecule reacts with PE and the atRal-
PE adduct is produced. Then, another atRal molecule can
bind to the adduct to form A2E-PE—a direct precursor of
A2E. Thus, one effect of the formation of A2E and its
precursor A2E-PE is to lower the photoreactivity of all-trans-
retinal. Pawlak et al. (26) determined with in vitro experi-
ments that A2E is really much less photoreactive than its
biosynthetic precursor. It was demonstrated that liposome
samples containing atRal and irradiated with 407 nm light
consumed 20.6 times more oxygen than did those containing
A2E.
Glutathione efflux
The RPE cells are permanently exposed to factors that
potentially induce oxidative stress. Those conditions assure
that RPE cells contain high amounts of superoxide dismutase
and catalase (11). The presence of aldehyde dehydrogenase
was affirmed in the RPE cells (47); this cytosolic enzyme
participates in the oxidation of various aldehydes, including
all-trans-retinal. Interestingly, it was shown that the expression
of aldehyde dehydrogenase 6 was reduced in RPE cells isolated
from older human eyes (48). The second line of defense against
the negative effects of oxidative stress in the RPE cells is the
presence of low-molecular-weight nonenzymatic antioxidants.
Phototoxicity of all-trans-retinal toward RPE cells is signifi-
cantly reduced by a-tocopherol and zeaxanthin (41).
The defense system in the RPE cells is supported by
glutathione (11), particularly by its reduced form, which is the
most abundant cellular nonprotein thiol (3). Synthesis and
maintenance of GSH are important for appropriate protection
against reactive electrophiles and oxidants (3). In this study
the measurements of extracellular glutathione showed that the
concentration of both its forms dramatically increased. The
efflux of the thiol was caused by increased membrane
permeability in cells preincubated with all-trans-retinal. Sim-
ilar effects were observed in other studies, where b-carotene
cleavage products containing retinal induced depletion of
glutathione and protein-SH in isolated mitochondria (36). The
level of GSH dramatically decreased in mitochondria in the
presence of retinal and other cleavage products; however, a
parallel increase in GSSG content was observed.
Irradiation of RPE cells preincubated with all-trans-retinal
had diverse effects on both intracellular and extracellular GSH
and GSSG concentrations depending on atRal concentration
and time of light exposure. Cells partially impaired by the dark
cytotoxic effect of all-trans-retinal were exposed to additional
stress in the form of reactive intermediates of retinal and
oxygen. The cells produced more GSH, which effluxed from
them and was used to counteract the negative effects of
oxidative stress. Prolonged irradiation induced bleaching,
isomerization and auto-oxidation of the retinal, thus dimin-
ishing the effects of oxidative stress: cell metabolism and
cytoplasmic membrane integrity recovered, and glutathione
efflux from the cells decreased significantly.
The observed decrease in GSH content inside the cells
treated with atRal was caused by greater efflux and oxidation,
which was manifested by a dramatic increase in the extracel-
lular GSSG level. Light exposure induced positive effects in the
cells. The higher concentration of atRal in cell membranes
caused ROS photoproduced by retinal to react directly with its
molecules, leading to their faster auto-oxidation. Modification
of retinal structure allowed the cell to recover: cell metabolism,
cytoplasmic membrane integrity and intracellular GSH con-
The oxygen uptake in all-trans-retinal samples was used
1
mostly for O2 photogeneration, which was confirmed by the
formation of 5a-hydroperoxycholesterol, a specific indicator of
Type II photosensitized oxidation of cholesterol. Singlet
oxygen was photogenerated in samples containing atRal at a
rate ꢀ70 times faster than those with A2E, where only a
vestigial concentration of the specific 1O2-derived oxidation
product of cholesterol was detected. A2E is a relatively
efficient quencher of singlet oxygen, with a constant quenching
rate of 108
)1
M
s)1. On the other hand, when A2E is exposed
to light, it generates superoxide anion (46). Electron para-
magnetic resonance spin trapping showed that A2E irradiated

790 Albert R. Wielgus et al.
tent increased significantly, while the levels of both glutathione
forms fell dramatically during visible light exposure.
REFERENCES
1. McBee, J. K., K. Palczewski, W. Baehr and D. R. Pepperberg
(2001) Confronting complexity: The interlink of phototransduc-
tion and retinoid metabolism in vertebrate retina. Prog. Retin. Eye
Res. 20, 469–529.
Hydroperoxide generation
2. Imanishi, Y., M. L. Batten, D. W. Piston, W. Baehr and K.
Palczewski (2004) Noninvasive two-photon imaging reveals retinyl
ester storage structures in the eye. J. Cell Biol. 164(3), 373–383.
3. Gao, X. and P. Talalay (2004) Induction of phase 2 genes by
sulforaphane protects retinal pigment epithelial cells against
photooxidative damage. Proc. Natl. Acad. Sci. U.S.A 101, 10446–
10451.
4. Travis, G. H., M. Golczak, A. R. Moise and K. Palczewski (2007)
Diseases caused by defects in the visual cycle: Retinoids as
potential therapeutic agents. Annu. Rev. Pharmacol. Toxicol. 47,
469–512.
5. Maeda, A., T. Maeda, M. Golczak, S. Chou, A. Desai, C. L.
Hoppel, S. Matsuyama and K. Palczewski (2009) Involvement of
all-trans-retinal in acute light-induced retinopathy of mice. J. Biol.
Chem. 284, 15173–15183.
6. Lai, Y. L., A. T. Tsin, K. W. Lam and J. J. Garcia (1985)
Distribution of retinoids in different compartments of the
posterior segment of the rabbit eye. Brain Res. Bull. 15(2), 143–
147.
7. Maeda, A., T. Maeda, M. Golczak and K. Palczewski (2008)
Retinopathy in mice induced by disrupted all-trans-retinal clear-
ance. J. Biol. Chem. 283, 26684–26693.
8. Sun, H., R. S. Molday and J. Nathans (1999) Retinal stimulates
ATP hydrolysis by purified and reconstituted ABCR, the photo-
receptor-specific ATP-binding cassette transporter responsible for
Stargardt disease. J. Biol. Chem. 274, 8269–8281.
9. Sun, H. and J. Nathans (1997) Stargardt’s ABCR is localized to
the disc membrane of retinal rod outer segments. Nat. Genet. 17,
15–16.
10. Chrispell, J. D., K. L. Feathers, M. A. Kane, C. Y. Kim, M.
Brooks, R. Khanna, I. Kurth, C. A. Hubner, A. Gal, A. J. Mears,
A. Swaroop, J. L. Napoli, J. R. Sparrow and D. A. Thompson
(2009) Rdh12 activity and effects on retinoid processing in the
murine retina. J. Biol. Chem. 284, 21468–21477.
The antioxidant defense in hRPE cells was ineffective when
the cells were incubated in the dark with 50 lM all-trans-
retinal. A dramatic increase in LOOH as well as those formed
in the hydrophilic environment of the cytoplasm was
observed. All-trans-retinal can react with primary amines
(23), which then bind to proteins and deactivate enzymes
including the antioxidant enzymes. All-trans-retinal can
enhance superoxide formation, increase hydrogen peroxide
levels and induce lipid peroxidation. The effect of all-trans-
retinal on H2O2 formation in the dark was indirectly proved
by noting the reduction in dark toxicity in the presence of
catalase added to RPE cells preincubated with atRal (41).
Additionally, the formation of malonic dialdehyde, a product
of lipid peroxidation, was detected in mitochondria incubated
with b-carotene cleavage products containing retinal (36).
The thiobarbituric acid reactive substances (TBARS) assay
showed a direct effect of all-trans-retinal on lipid peroxidation
in ARPE-19 cells (3). The same study also showed that the
TBARS concentration rose during irradiation of cells, while
here the LOOH level decreased with prolonged exposure to
visible light. That apparent contradiction may be caused by the
fact that the cells in the studies of Gao and Talalay (3) were
irradiated with UVA, while we exposed them to visible light.
When cells preloaded with atRal are irradiated with UVA,
potential photosensitizing products of retinal photo-bleaching,
-isomerization, and -oxidation that do not absorb visible light
can produce further ROS when they are exposed to UVA.
Hydroperoxides are the intermediate products of lipid
peroxidation. Their reduced formation during visible light
exposure shows that the rate of photoproduction of lipid
peroxidation end products also decreases. The lessening rate of
LOOH photogeneration in hRPE cells preincubated with
50 lM atRal was accompanied by their light-induced recovery.
Increase in metabolic activity and formation of GSH, the
reduced form of glutathione, facilitated LOOH scavenging by
antioxidant enzymes.
11. Strauss, O. (2005) The retinal pigment epithelium in visual func-
tion. Physiol. Rev. 85, 845–881.
12. LaVail, M. M. (1973) Kinetics of rod outer segment renewal in the
developing mouse retina. J. Cell Biol. 58, 650–661.
13. Sakai, N., J. Decatur, K. Nakanishi and G. E. Eldred (1996)
Ocular age pigment ‘‘A2-E’’: An unprecedented pyridinium bis-
retinoid. J. Am. Chem. Soc. 118, 1559–1560.
14. Roberts, J. E., B. M. Kukielczak, D. N. Hu, D. S. Miller, P. Bilski,
R. H. Sik, A. G. Motten and C. F. Chignell (2002)The role of A2E in
prevention or enhancement of light damage in human retinal pig-
ment epithelial cells. Photochem. Photobiol. 75, 184–190.
15. Kim, S. R., Y. P. Jang, S. Jockusch, N. E. Fishkin, N. J. Turro and J.
R. Sparrow (2007) The all-trans-retinal dimer series of lipofuscin
pigments in retinal pigment epithelial cells in a recessive Stargardt
disease model. Proc. Natl. Acad. Sci. U.S.A. 104, 19273–19278.
16. Sparrow, J. R., C. A. Parish, M. Hashimoto and K. Nakanishi
(1999) A2E, a lipofuscin fluorophore, in human retinal pigmented
epithelial cells in culture. Invest. Ophthalmol. Vis. Sci. 40, 2988–
2995.
17. Lakkaraju, A., S. C. Finnemann and E. Rodriguez-Boulan (2007)
The lipofuscin fluorophore A2E perturbs cholesterol metabolism
in retinal pigment epithelial cells. Proc. Natl. Acad. Sci. U.S.A.
104, 11026–11031.
18. Friedman, D. S., B. J. O’Colmain, B. Munoz, S. C. Tomany, C.
McCarty, P. T. de Jong, B. Nemesure, P. Mitchell and J. Kempen
(2004) Prevalence of age-related macular degeneration in the
United States. Arch. Ophthalmol. 122, 564–572.
CONCLUSION
These studies support our previous conclusions (14,26) that
although A2E may be harmful at high concentrations,
especially when oxidized, its cytotoxic and phototoxic effects
even at the highest physiological level (20–25 lM) are much
lower compared with the free form of its precursor all-trans-
retinal at the same concentration and entirely negligible
compared with all-trans-retinal at twice the concentration.
Thus the endogenous production of A2E, which is formed of
two molecules of unbound all-trans-retinal, may serve as a
protective mechanism to prevent damage to the retina by the
free form of retinal.
19. Saari, J. C., G. G. Garwin, J. P. Van Hooser and K. Palczewski
(1998) Reduction of all-trans-retinal limits regeneration of visual
pigment in mice. Vision Res. 38, 1325–1333.
20. Tolleson, W. H., S. H. Cherng, Q. Xia, M. Boudreau, J. J. Yin, W.
G. Wamer, P. C. Howard, H. Yu and P. P. Fu (2005) Photo-
decomposition and phototoxicity of natural retinoids. Int.
J. Environ. Res. Public Health 2, 147–155.
Acknowledgements—This research was supported by the Intramural
Research Program of the NIH, National Institute of Environmental
Health Sciences. We thank Dr. Dan-Ning Hu for providing us with the
human RPE cell line and Dr. Ann Motten for help in preparation of
this manuscript.

Photochemistry and Photobiology, 2010, 86 791
21. Yang, M. and H. K. Fong (2002) Synthesis of the all-trans-retinal
chromophore of retinal G protein-coupled receptor opsin in cul-
tured pigment epithelial cells. J. Biol. Chem. 277, 3318–3324.
22. Boulton, M. and P. Dayhaw-Barker (2001) The role of the retinal
pigment epithelium: Topographical variation and ageing changes.
Eye (London) 15, 384–389.
23. Sparrow, J. R., Y. Wu, C. Y. Kim and J. Zhou (2009) Phospho-
lipid meets all-trans-retinal: The making of RPE bisretinoids.
J. Lipid Res. 51, 247–261.
37. Schutt, F., M. Bergmann, F. G. Holz, S. Dithmar, H. E. Volcker
and J. Kopitz (2007) Accumulation of A2-E in mitochondrial
membranes of cultured RPE cells. Graefes Arch. Clin. Exp.
Ophthalmol. 245, 391–398.
38. Bergmann, M., F. Schutt, F. G. Holz and J. Kopitz (2004) Inhi-
bition of the ATP-driven proton pump in RPE lysosomes by the
major lipofuscin fluorophore A2-E may contribute to the patho-
genesis of age-related macular degeneration. FASEB J. 18,
562–564.
24. Boulton, M., M. Ro
´
_zanowska, B. Ro
´
_zanowski and T. Wess (2004)
39. Vives-Bauza, C., M. Anand, A. K. Shirazi, J. Magrane, J. Gao, H.
R. Vollmer-Snarr, G. Manfredi and S. C. Finnemann (2008) The
age lipid A2E and mitochondrial dysfunction synergistically
impair phagocytosis by retinal pigment epithelial cells. J. Biol.
Chem. 283, 24770–24780.
The photoreactivity of ocular lipofuscin. Photochem. Photobiol.
Sci. 3, 759–764.
25. Harper, W. S. and E. R. Gaillard (2001) Studies of all-trans-retinal
as a photooxidizing agent. Photochem. Photobiol. 73, 71–76.
26. Pawlak, A., M. Wrona, M. Ro
´
_zanowska, M. Zare˛ba, L. E. Lamb,
40. Finnemann, S. C., L. W. Leung and E. Rodriguez-Boulan
(2002) The lipofuscin component A2E selectively inhibits
phagolysosomal degradation of photoreceptor phospholipid by
the retinal pigment epithelium. Proc. Natl. Acad. Sci. U.S.A. 99,
3842–3847.
J. E. Roberts, J. D. Simon and T. Sarna (2003) Comparison of the
aerobic photoreactivity of A2E with its precursor retinal. Photo-
chem. Photobiol. 77, 253–258.
27. Parish, C. A., M. Hashimoto, K. Nakanishi, J. Dillon and J.
Sparrow (1998) Isolation and one-step preparation of A2E and
iso-A2E, fluorophores from human retinal pigment epithelium.
Proc. Natl. Acad. Sci. U.S.A. 95, 14609–14613.
41. Ro_zanowski, B., M. B. Ro_zanowska and M. E. Boulton (2003)
´ ´
Toxicity of all-trans retinal to the retinal pigment epithelium. In-
vest. Ophthalmol. Vis. Sci. 44. E-Abstract 1643.
28. Dillon, J., E. R. Gaillard, P. Bilski, C. F. Chignell and K. J.
Reszka (1996) The photochemistry of the retinoids as studied by
steady-state and pulsed methods. Photochem. Photobiol. 63(5),
680–685.
29. Kanofsky, J. R., P. D. Sima and C. Richter (2003) Singlet oxygen
generation from A2E. Photochem. Photobiol. 77, 235–242.
30. Hu, D.-N., M. A. Del Monte, S. Liu and I. H. Maumenee (1982)
Morphology, phagocytosis, and vitamin A metabolism of cultured
human retinal pigment epithelium. Birth Defects Orig. Artic. Ser.
18, 67–79.
42. Ostrovskii, M. A. and I. B. Fedorovich (1994) Retinal as a photo-
damage sensitizer of retinal-containing retina proteins. Biofizika
39, 13–25.
43. Sun, H. and J. Nathans (2001) ABCR, the ATP-binding cassette
transporter responsible for Stargardt macular dystrophy, is an
efficient target of all-trans-retinal-mediated photooxidative dam-
age in vitro. Implications for retinal disease. J. Biol. Chem. 276,
11766–11774.
44. Ro
´
_zanowska, M. B., R. Edge, E. T. Land, S. Navaratnam, T. G.
_zanowski (2007) Free radical interactions of
Truscott and B. Ro
´
31. Hu, D.-N., S. A. McCormick, R. Ritch and K. Pelton-Henrion
(1993) Studies of human uveal melanocytes in vitro: Isolation,
purification and cultivation of human uveal melanocytes. Invest.
Ophthalmol. Vis. Sci. 34, 2210–2219.
32. Inbaraj, J. J. and C. F. Chignell (2004) Cytotoxic action of juglone
and plumbagin: A mechanistic study using HaCaT keratinocytes.
Chem. Res. Toxicol. 17, 55–62.
33. Gay, C. A. and J. M. Gebicki (2002) Perchloric acid enhances
sensitivity and reproducibility of the ferric-xylenol orange perox-
ide assay. Anal. Biochem. 304, 42–46.
34. DeLong, J. M., R. K. Prange, D. M. Hodges, C. F. Forney, M. C.
Bishop and M. Quilliam (2002) Using a modified ferrous oxida-
tion-xylenol orange (FOX) assay for detection of lipid hydroper-
oxides in plant tissue. J. Agric. Food. Chem. 50, 248–254.
35. Sodegren, E., J. Nourooz-Zadeh, L. Berglund and B. Vessby
(1998) Re-evaluation of the ferrous oxidation in xylenol assay for
measurement of plasma lipid hydroperoxides. J. Biochem.
Biophys. Methods 37, 137–146.
36. Siems, W., O. Sommerburg, L. Schild, W. Augustin, C. D. Lan-
ghans and I. Wiswedel (2002) Beta-carotene cleavage products
induce oxidative stress in vitro by impairing mitochondrial respi-
ration. FASEB J. 16, 1289–1291.
retinal constituents—Retinaldehyde and its adducts with phos-
phatidylethanolamine. Invest. Ophthalmol. Vis. Sci. 48. E-Abstract
3258.
45. Baron, M. H., M. J. Coulange, C. Coupry, D. Baron, J. Favrot
and M. M. Abo Aly (1989) All-trans retinal photoisomerization
and photooxidation from UV laser radiation. Vibrational assign-
ment of all-trans 5,8-peroxyretinal. Photochem. Photobiol. 49,
739–751.
46. Broniec, A., A. Pawlak, T. Sarna, A. Wielgus, J. E. Roberts, E. J.
Land, T. G. Truscott, R. Edge and S. Navaratnam (2005) Spec-
troscopic properties and reactivity of free radical forms of A2E.
Free Radic. Biol. Med. 38, 1037–1046.
47. Horton, J. D., N. A. Shah, J. A. Warrington, N. N. Anderson, S.
W. Park, M. S. Brown and J. L. Goldstein (2003) Combined
analysis of oligonucleotide microarray data from transgenic and
knockout mice identifies direct SREBP target genes. Proc. Natl.
Acad. Sci. U.S.A. 100, 12027–12032.
48. Ishibashi, K., J. Tian and J. T. Handa (2004) Similarity of mRNA
phenotypes of morphologically normal macular and peripheral
retinal pigment epithelial cells in older human eyes. Invest.
Ophthalmol. Vis. Sci. 45, 3291–3301.