Characterization of Photodynamic Actions
of Rose Bengal on Cultured Cells
Scheffer C. G. Tseng,* Robert P. G. Feenstra* and Brant D. Watsonf
Purpose. The authors have previously reported successful photodynamic occlusion of corneal
blood vessels using intravenous rose bengal and argon green laser irradiation. To explore the
action mechanism of this novel technique of photothrombosis, they examined the photodynamic effect of rose bengal on cultured fibroblasts, smooth muscle cells, and vascular endothelium—the cellular components of blood vessels.
Methods. Five types of cells were exposed to different concentrations of rose bengal and argon
green laser irradiation. The irradiated cell areas were analyzed by fluorescence microscopy and
fluorometry. Various potential quenchers and proteins were tested for their modulation of the
photodynamic action.
Results. Upon irradiation with 16 W/cm2 of argon green laser light in conjunction with rose
bengal concentrations extending above 1 X 10~4 M, all cultured cell types showed a dose-dependent photobiologic effect characterized by constriction and detachment of the laser-irradiated cell region from the rest of the cell monolayer. In addition, there was dye photobleaching and development of a blue shift of the fluorescence excitation and emission maxima in the
irradiated cell areas. Binding of rose bengal to intracellular components was demonstrated by
fluorescence microscopy and by fluorometry showing a red shift of the excitation maximum
compared to the maximum in solution. This binding was a prerequisite for expression of the
described photobiologic effect, because polymer-conjugated rose bengal (Sensitox II) failed to
reproduce it. The addition of native or heat-inactivated bovine serum albumin or catalase
decreased this photobiologic effect also owing to dye binding, as indicated by G-75 Sephadex
gel nitration chromatography.
Conclusion. These results indicate that the specific photobiologic effect of monolayer contraction, which simulates the vasoconstriction seen during photothrombosis under argon green
irradiation, appears to be caused by the photochemical interaction of rose bengal bound with
intracellular components. Invest Ophthalmol Vis Sci. 1994;35:3295-3307.
Jfcvose bengal, a halide derivative of fluorescein, has
been used in ophthalmology as a dye for the diagnosis
of various external eye diseases.1 As a photosensitizer,
rose bengal can kill microorganisms such as viruses,2
bacteria,3 and protozoa,4 and can induce photodyna-
From the * Department of Ophthalmology, Bascom Palmer Eye Institute, and the
•\Cerebral Vascular Disease Research Center, Department of Neurology, National
Parkinson Foundation, University of Miami School of Medicine, Miami, Florida.
Presented in part at the annual meeting of the Association for Research in Vision
and Ophthalmology, May 1990, Sarasota, Florida.
Supported in part by Public Health Service research grant EY06819 (SCOT), NS
23244 (BDW), and core grant FY02180 from the National Eye Institute,
Department of Health and Human Services, National Institutes of Health,
Bethesda, Maryland.
Submitted for publication January 12, 1994; accepted February 23, 1994.
Proprietary interest category: N (SCGT, RPGF); P (BDW owns U.S. patent no.
5,056,006 on the method of laser-driven photolhrombotic vascular occlusion
mediated by rose bengal or erythrosin B).
Reprint requests: Scheffer C. G. Tseng, MD, PhD, Bascom Palmer Eye Institute,
P.O. Box 016880, Miami, FL 33101.
mic effects in vitro on red blood cells,5 cardiomyocytes,6 and retinal pigment epithelial cells.7 The photodynamic effects of rose bengal ex vivo have also been
reported in nerve axon,8 corneal endothelium,9
heart,10 and pancreatic acini.11 Intravenously delivered rose bengal in vivo, in conjunction with optically
filtered xenon arc lamp or argon green laser irradiation, can induce occlusion of blood vessels in a procedure called photothrombosis.1213 This technique of
photothrombosis has been used to create a rat model
of reproducible brain infarction12 and to occlude rabbit corneal neovascularization14"16 and retinal—choroidal vessels.17"19 Arterial irradiation in the presence of
rose bengal induces constriction,13"19 for which the
mechanism remains unexplained.
In an in vitro aqueous environment containing
nucleotides, the photodynamic mechanism of rose
bengal under laser irradiation at 532 nm is thought to
Investigative Ophthalmology & Visual Science, July 1994, Vol. 35, No. 8
Copyright © Association for Research in Vision and Ophthalmology
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
3295
3296
Investigative Ophthalmology 8c Visual Science, July 1994, Vol. 35, No. 8
be predominantly type II, generating 80% singlet oxygen with the remainder being 20% superoxide anion.20
In a biologic system containing cell membranes, the
primary photobiologic effects of rose bengal are
thought to be mediated by the photomodification of
cell membranes.21 Recently, in a quest to determine
what is actually stained by an ophthalmic rose bengal
solution on the ocular surface, we reported that rose
bengal dose dependently stains several types of cultured cell.22 Even in the absence of light exposure, the
stained cells lose their vitality, indicating that rose bengal is intrinsically toxic; in the presence of light, there
is an additional photodynamic damaging effect.22 It
remains unclear whether there is a relationship between rose bengal's intrinsic toxicity and its photodynamic activity. The spectrum of photodynamic actions
will likely vary, however, according to the rose bengal
concentration because rose bengal stains plasma
membranes at low, but intracellular components at
high, concentrations.22 As a first step to explore the
mechanism of photothrombosis, we chose to study the
photodynamic effect of rose bengal on cultured vascular endothelial cells, smooth muscle cells, and fibroblasts, i.e., cell components normally present in blood
vessel structure. We characterized the photodynamic
effect of rose bengal on these cultures when the rose
bengal concentration and the argon green laser intensity and exposure time were comparable to those used
for in vivo photothrombosis.
MATERIALS
Two types of smooth muscle cell, one from rat embryonic thoracic aorta (A7r5) and the other from human
intestine (HISM), were obtained from the American
Type Culture Collection (Rockville, MD). These cells
and others to be described below were routinely
grown in tissue flasks (75 cm2) obtained from Corning
Glass Works (Corning, NY). For photodynamic treatment, all cells were grown on glass culture chamber
slides, and for subculture studies, on plastic culture
chamber slides obtained from Nunc, Inc. (Naperville,
IL). Fetal bovine serum (FBS) and Dulbecco's modified Eagle's medium (DMEM) were from Gibco, Inc.
(Grand Island, NY). Ethylene diamine tetraacetic acid
(EDTA), trypsin and fatty acid- and globulin-free bovine serum albumin (BSA) were from Sigma (St. Louis,
MO). Rose bengal, 97% pure, was obtained from Aldrich (Milwaukee, WI); polymer-conjugated rose bengal (Sensitox II) was obtained from Chemalog (South
Plainfield, NJ). Quenchers such as sodium azide, dimethyl sulfoxide (DMSO), Cu, Zn- superoxide dismutase (SOD) with 3570 U/mg protein and catalase with
15700 U/mg protein were obtained from Sigma (St.
Louis, MO), and bixin was from Pfaltz and Bauer (Wa-
terbury, CT). All experiments adhered to the ARVO
Statement for the Use of Animals in Ophthalmic and
Vision Research.
METHODS
Cell Cultures
Rabbit Tenon fibroblasts and human conjunctival fibroblasts were derived from explant cultures of either
rabbit Tenon fibrous tissue or human subconjunctival
soft connective tissue. The fibroblasts that migrated
from the explants were trypsinized and passaged on
plastic plates. For this study, fibroblasts derived from
passages 9 to 17 were used. Vascular endothelial
(CPA) cells from passages 33 to 36, derived from a
tissue culture of rat thoracic aorta, were kindly provided by Dr. Una Ryan (T-Cell Sciences, Boston, MA).
Two types of smooth muscle cell, A7r5 and HISM, at
passages 22 to 32 were used.
All cultures were grown in tissue culture flasks
containing DMEM and 10% FBS and incubated in humidified 5% CO2 at 37°C. Media were changed every 3
days. Rabbit Tenon fibroblasts and human conjunctivalfibroblastswere passaged by 0.25% trypsin in phosphate-buffered saline (PBS); HISM were passaged
with an additional 0.2% EDTA; CPA cells were passaged by scraping with a policeman without enzymatic
digestion.
Photodynamic Effect of Rose Bengal Under
High-Intensity Laser Irradiation
To determine the photodynamic effect that could be
created at rose bengal concentrations above 1 X 10~5
M, which allows rose bengal to stain intracellular components and quickly kills stained cells,19 photodynamic
treatments were conducted using confluent cultures
with rose bengal concentrations from 1.0 X 10~4 M to
5.0 X 10~3 M, a range that putatively spans that used
for in vivo photothrombosis.1214 The cell layers were
irradiated with an argon green laser (Lexel [Palo Alto,
CA] model 95-4) set at 514.5 nm and a beam diameter
of 1 mm (as determined by a stainless steel aperture).
The exposure time was varied from 0.5, 2.5, 5.0, 7.5,
and 15 minutes. Preliminary work showed that a minimum of 16 W/cm2 was needed to create a photodynamic effect observable by fluorescence microscopy (see
Fig. 1). This intensity level, which isfiveorders of magnitude higher than the ambient light intensity, is sufficient to induce photothrombotic occlusion of small
arteries in vivo.13 Therefore, 16 W/cm2 was used for
the rest of the experiments in photodynamic treatments on the cultured cells, which were placed at room
temperature under atmospheric oxygen pressure.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
3297
Photodynamic Action of Rose Bengal In Vitro
Photodynamic Treatment With Bound Rose
Bengal Under High-Intensity Irradiation
The cultures were also incubated with a suspension of
polymer-conjugated, immobilized rose bengal beads,
Sensitox H23>24 prepared at a concentration of 0.4 g/
ml in PBS. This concentration yielded the same absorbance as that of a 5.0 X 10~3 M rose bengal solution at
A maximum of 562 nm, based on a previous study.25 In
a parallel experiment, the photodynamic treatment
was conducted in this solution containing suspended
Sensitox II beads. Other experimental conditions of
laser treatment were identical to those described
above. These experiments were repeated by using a
Sensitox II suspension not washed with PBS before
the laser irradiation so as to ensure against loss of
beads during washout.
That BSA binds strongly with rose bengal26 suggests that the photodynamic effect of rose bengal can
be inhibited if the rose bengal is in the bound, rather
than the free, form in the incubation solution. To examine if this was also the case, photodynamic treatment was performed in 5.0 X 10~3 M rose bengal solutions containing BSA at concentrations from 4.0 X
10"5 to 5.0 X 10~2 g/ml. After incubation and washing
with protein-containing PBS, laser irradiation at 16
W/cm2 was performed for 7.5 min in BSA-containing
PBS without rose bengal. The above experiment was
also compared with BSA, catalase, or SOD inactivated
by heating at 100°C for 30 minutes. Other experimental conditions were identical to those described above.
Photodynamic Treatment of Rose Bengal
Under High-Intensity Irradiation in the
Presence of Various Quenchers
The effect of quenchers was studied by adding solution of 0.05 to 500 mM sodium azide, 10"7 to 10"4 M
bixin, 0.5 to 5% DMSO, 10~4 to 0.1 g/ml catalase, or 1
X 10"4 to 1 X 10"3 g/ml SOD to a 5 X 10"3 M rose
bengal solution. After incubating in each of the above
solutions for 5 minutes, the cells were washed twice
and irradiated for 7.5 minutes in quencher-containing
PBS without rose bengal. Other conditions of laser
treatment were identical to those described above.
Fluorescence and Phase-Contrast Microscopy
After photodynamic treatment, the chamber slides
were covered with a sheet of aluminum foil to avoid
any further light exposure. The treated as well as untreated areas were examined first with a Zeiss (Oberkochen, Germany) Axiophot Photomicroscope with
epifluorescence capability. Pictures were taken using
Kodak Ektachrome 160 ASA film (Eastman Kodak,
Rochester, NY) with the "blue" filter combination
with excitation (Exc) at 485 to 510 nm and emission
(Em) at 515 to 565 nm, as well as with the "green"
filter combination with Exc at 546 to 580 nm and Em
at >590 nm. Exposure time was set at 45 seconds if
pictures were taken using the blue filter combination
and on automatic if using the green filter combination. Afterward, phase-contrast photographs were
taken with a Nikon (Tokyo, Japan) phase-contrast inverted microscope using Kodak Ektachrome 50 ASA
film.
Excitation and Emission Spectra of Rose
Bengal Before and After Photodynamic
Treatment
The effect of photooxidation on the spectral properties of rose bengal was investigated using cultured
CPA cells that were grown to con fluency on 8 mm X
35 mm fused silica plates and incubated for 5 minutes
with 5 X 10~4 M rose bengal in PBS after three PBS
washes. The rose bengal solution was replaced with
two washes of PBS, and the stained cell preparation,
together with the fused silica plate, was then placed
vertically into a standard 1 cm X 1 cm fused silica
cuvette containing PBS. This cell preparation was irradiated for 2 minutes (until cells appeared to be
bleached) using the same argon laser set at an intensity
of approximately 0.5 W/cm2. The cell plates were then
introduced into a Perkin-Elmer (Norwalk, CT) Model
LS-5B fluorometer with a front surface accessory used
for light-scattering media and slits set at 5 nm, and Exc
and Em spectra of the photodynamically treated and
untreated areas were recorded.
Gel Filtration Chromatography
To determine whether the inhibitory effect of BSA,
catalase, or SOD might come from the binding of
these proteins to rose bengal, we performed gel filtration chromatography using a Sephadex G-75 column
(1.5 X 50 cm), which was equilibrated with 20 mM
Tris/HCl, pH 7.4, containing 150 mM NaCl and 2 mM
sodium azide. A 0.5-ml sample was loaded containing
1 X 10~3 M rose bengal and 8 X 10"5 M BSA, 9 X 10"5
M catalase, or 8 X 10~5 M SOD in the same buffer
solution as described above, and 2-ml fractions were
collected at a flow rate of 8 ml/hr. Each fraction was
monitored with a spectrophotometer at 548 nm and
562 nm for free and bound rose bengal, respectively,
as well as at 280 nm for the added proteins.
RESULTS
Photobiologic Effect of Rose Bengal With and
Without Various Quenchers
Under laser irradiation at 16 W/cm2, all five types of
cultured cell showed a similar dose-response relationship with respect to the laser exposure time, rose bengal concentration, and photobiologic effects. An example of CPA cells can be seen in Figure 1. The photo-
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
3298
Investigative Ophthalmology 8c Visual Science, July 1994, Vol, 35, No. 8
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
3299
Photodynamic Action of Rose Bengal In Vitro
FIGURE l. Dose-response relationship with respect to laser exposure time, rose bengal concentration, and photobiologic effects under argon green laser irradiation. The photodynamic treatment was performed on CPA cells, as described in Methods, using 1 X 10"4 M, 5 X
10"4 M, 2.5 X 10~3 M, or 5 X 10~3 M rose bengal (from left to right vertical rows, respectively) and exposure time of 0.5, 2.5, 5.0, 7.5, or 15 minutes (from top to bottom horizontal
rows, respectively). The photobiologic effects were observed by a fluorescence microscope
using the green filter combination (Exc 546 to 580 nm, Em > 590 nm) (a) and the blue filter
combination (Exc 485 to 510 nm, Em 515 to 565 nm) (b). The fluorescence change under the
blue filter combination was obvious at 5 X 10~4 M rose bengal (b), and the monolayer
separation and cell detachment was obvious at 2.5 X 10~3 M or 5 X 10~3 M rose bengal for
longer exposure times.
biologic effects were characterized by a change in rose
bengal fluorescence properties and retractive separation of the cell monolayer in the photodynamically
treated area from the surrounding unirradiated cell
area.
Under the green filter combination (Exc 546 to
580 nm, Em > 590 nm), both irradiated and unirradiated cells exposed to 5 X 10~4 M rose bengal exhibited orange fluorescence, decorating the cell membranes as well as intracellular cytoplasmic structures,
among which the nucleus was strongly stained (Fig.
2a). This finding was consistent with our previous report in which similar fluorescence was noted at 9.8 X
10~5 M or 0.01% rose bengal. 22 In comparison, the
laser-irradiated cell region at higher rose bengal concentrations showed an increase of yellow-orange fluorescence (Figs, la, 2b, and 2c). Separation of cell
monolayers at the edge of the treated area was observed at the rose bengal concentrations of 2.5 X 10~3
M and 5.0X10~ 3 M and for exposure times equal to or
greater than 7.5 and 5 minutes, respectively (Figs, l a
and 2c). Under high magnification, separation of the
monolayer appeared to be generated by waves of a
contraction force that acted only on the irradiated
cells (Fig. 2d). When the exposure time was increased
to 15 minutes at either of these two concentrations,
cell detachment occurred at the edge of the irradiated
area (Figs, l a and 2c). Under the blue filter combination (Exc 485 to 510 nm, Em 515 to 565 nm), the
untreated cells exhibited no fluorescence (Fig. lb). In
5.0 X 10~4 M rose bengal, the irradiated cells yielded a
bright green fluorescence, the intensity of which increased as the exposure time increased from 2.5 to 15
minutes (Fig. l b , second column, and Fig. 2e). However, the green fluorescence of the irradiated cells decreased when the rose bengal concentration was increased from 5 X 10" 4 M to 2.5 X 10" 3 M or 5 X 10" 3
M (Fig. l b [see second column versus third and fourth
columns] and see Fig. 2e versus Fig. 2f). The monolayer separation follows a reciprocal dose-response
relationship between energy fluence (intensity X exposure time) and rose bengal concentration in a quali-
tative manner (Fig. 2). This dose-response relationship was similar for all five types of cultured cell except
for the two smooth muscle cell cultures in which separation of the cell monolayer occurred at a slightly
lower laser fluence. That is, for 5.0 X 10" 4 M, 2.5 X
10~3 M, and 5.0 X 10~3 M rose bengal, such a change
could be noted for the exposure times of 7.5, 5, and 2
minutes, respectively (data not shown). The controls
with laser irradiation alone did not show any visible
change under either filter combination (data not
shown).
Under this high-intensity irradiation, the observed
photobiologic effects could not be inhibited by the
addition of 10~4 M bixin, 0.5 M sodium azide, 1 X 10~3
g/ml native or heat-inactivated SOD, or 5% DMSO,
but they could be inhibited by 0.1 g/ml native or heatinactivated catalase (data not shown).
Migration of Photobiologic Effects Using
Polymer-Conjugated Rose Bengal or BSA
It should be noted that the above photobiologic effects occurred at rose bengal concentrations, which, as
shown previously,21 permit rose bengal to enter the
cell and cause intrinsic toxicity. Photodynamic treatment was thus performed on HISM cells using polymer-conjugated rose bengal beads (Sensitox II), which
prevents rose bengal from entering the cells and, accordingly, limits singlet oxygen production to the extracellular environment. After 7.5 minutes of irradiation, monolayer cell separation occurred with the free
rose bengal solution (Fig. 1) but did not occur with the
use of Sensitox II (Figs. 3a and 3b). In fact, the fluorescence changes under the green or blue filter combinations were limited to the polymer beads, and the intracellular fluorescence under either filter combination
was notably decreased, confirming that the intracellular uptake of rose bengal was not occurring (Figs. 3c
and 3d). Similar results were obtained even when the
treatment was performed in the Sensitox II-containing solution without prior PBS washes (Fig. 3e), and
the photobleaching effect could only be observed on
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
FIGURE 2. Color fluorescence photographs of CPA cells under photodynamic treatment using argon green laser irradiation (also see Fig. 7). Under the green filter combination, all cells
of the laser-irradiated and unirradiated area fluoresced orange-red. At 5 X 10~4 M rose
bengal, the low magnification showed a relatively low-intensity fluorescence in the nuclei due
to an autoquenching effect, consistent with our previous finding.22 Under high magnification, strong nuclear fluorescence could be appreciated as well as other intracellular staining
(b). At 5 X 10~;> M rose bengal, there was an increase of orange-yellow fluorescence in the
irradiated area (c) in addition to the striking finding of monolayer separation, which, under
high magnification, disclosed waves of contractions in irradiated cells (arrows, d). Under the
blue filter combination, the same region of the photograph (a) revealed a strong green
fluorescence (e). However, there was a decrease in fluorescence in (f) for the same area
shown in photograph (c).
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
Photodynamic Action of Rose Bengal In Vitro
FIGURE 3. Photodynamic treatment using polymer-conjugated rose bengal beads (Sensitox
II); details of the treatment condition are described in Methods. When the Sensitox II suspension was washed before laser irradiation in the same manner as that described for free rose
bengal solutions, there was no monolayer separation as shown in the phase-contrast micrograph (a). To verify this, the photograph was retaken on the same area after additional PBS
washes (b). The fluorescence change of rose bengal under the green (c) and blue (d) filter
combinations following photodynamic treatment can be seen only on the polymer beads but
not in the cells. No monolayer separation was seen even when Sensitox II polymer beads were
not prewashed before treatment (e). To take the photograph, these polymer beads were
washed after the treatment. The photobleaching effect, however, was noted also on the
polymers as the treated beads become colorless (at center), which showed an increased yellow
fluorescence under the green filter combination (£).
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
3301
3302
Investigative Ophthalmology & Visual Science, July 1994, Vol. 35, No, 8
FIGURE 4. Effect of BSA on photodynamic treatment during laser irradiation. As described
under Methods for HISM cells, the addition of 2% BSA inhibited the photobiologic effects
that otherwise would be noted by 5 X 10"3 M rose bengal (a, b) using the green (left column) or
blue (right column)filtercombination. This inhibitory effect was less at 1% or 0.5% BSA,
where there was no monolayer separation (c and e, respectively, under the green filter combination), but a clear green fluorescence was noted on the treated area (d and f, respectively,
under the blue filter combination) that otherwise would occur at 5 X 10~4 M rose bengal if no
BSA was added. When BSA was lowered to 0.02%, the photobiologic effect of monolayer
separation recurred (g and h).
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
Photodynamic Action of Rose Bengal In Vitro
the beads of the irradiated area (Fig. 3f). These data
indicate that the observed monolayer separation was
caused by photodynamic activity of intracellularly absorbed rose bengal.
To substantiate this concept, we performed photodynamic treatment on HISM cells with an exposure
time of 7.5 minutes using the 5.0X10~3 M rose bengal
solution containing various BSA concentrations from
0.004% to 5%, based on the fact that BSA binds rose
bengal strongly.26 Under the green filter combination,
we observed that BSA concentrations between 2% and
5% could effectively prevent intracellular uptake of
rose bengal, as evidenced by the decreased intracellular rose bengal fluorescence, and thus inhibited the
occurrence of monolayer cell separation (Figs. 4a and
4b). At 1% and 0.5% BSA, the treated cells exhibited
increased bright green fluorescence under the blue
filter combination, indicating that the effective concentration of rose bengal solution was actually reduced from 5.0 X 10"3 M to the level of 5.0 X 10"4 M
BSA-free rose bengal solution, and monolayer separation could still be prevented (see Figs, lb and 4c to 4f).
When BSA was lowered to 0.02% or below, monolayer
separation again occurred (Figs. 4g and 4h), indicating
that 0.02% BSA was not sufficient to inhibit the rose
bengal-induced photobiologic effect.
Changes of Fluorescence Spectra Within the
Irradiated Area
As shown in Figure 5 (upper graph), the unirradiated
cell areas revealed a biphasic Exc spectrum with max at
562 nm when Em was set at 650 nm. When Exc was set
at 520 nm (to prevent emission spectral overlap), the
Em spectrum was sharp and peaked at 580 nm. This
spectrum explains why the untreated cell layer fluorescence was readily detected under the green filter combination but poorly under the blue filter combination
(Figs. 1 and 2). In contrast, the photodynamically
treated cell layer showed rapid bleaching with a dramatic decrease of fluorescence under either filter combination. There was also a shift of the Em spectrum
max to 570 nm when Exc was set at 510 nm (Fig. 5,
lower graph). This blue shift explains why there was
more fluorescence detected in the emission range of
the blue filter (Fig. 1). The controls with laser irradiation alone or rose bengal alone did not show any such
fluorescence spectral changes. These results indicate
that the observed spectral changes are specific and are
mediated by the photosensitizer dye rose bengal in the
bound state, as indicated first by the red shift of its
excitation spectrum compared to its solution spectrum. Photodynamic activity of this bound rose bengal
upon irradiation is further indicated by the subsequent blue shift of its excitation and emission maxima,
as well as by diminution of fluorescence intensity.
Bleaching of rose bengal was observed previously to
3303
«
2
TREATED AREA
Wavelength |nm|
FIGURE 5. Photobleaching effect as evidenced by the changes
of fluorescence spectra on the untreated {top graph) and the
treated {bottom graph) cell area. For the upper graph, Exc was
obtained with Em set at 650 nm and Em with Exc set at 520
nm. For the bottom graph, Exc was obtained with Em set at
570 nm and Em with Exc set at 510 nm.
occur in the milieu of a platelet suspension over cultured endothelium and irradiated with green light
(Watson, unpublished observation, 1984).
Binding of Rose Bengal With BSA, Catalase,
and SOD
Because the inhibitory effect of catalase on monolayer
separation could still be produced by heat-inactivated
catalase and thus could not be attributed to its enzymatic action, we suspected that catalase might bind
with rose bengal in a manner similar to BSA. To verify
this, G-75 sephadex gel filtration chromatography was
performed. Without adding any protein, free rose
bengal alone was eluted as a single peak between fractions # 85 to #110. The higher ratio of absorbance
measured at X548 nm to that measured at X562 nm
indicated that this peak was due to the free form of
rose bengal. In this peak, the absorbance measured at
X280 nm was due to free rose bengal when measured
at high concentrations and not due to protein because
protein alone without rose bengal was exclusively
eluted in the void volume at the first peak (data not
shown). The elution of rose bengal was three times
delayed from the total volume of this column calibrated by fluorescein alone (data not shown), indicating that this anomalous elution behavior is attributed
to the added halide derivatives. Although the mechanism of this anomalous delay is not fully understood,
the addition of BSA clearly demonstrated its binding
with rose bengal, yielding at the void volume a separate peak that had a relatively high absorbance at X562
nm and X280 nm (Fig. 6). A similar result was also
obtained with catalase, except that the absorbance at
X562 nm of the bound rose bengal was less than that of
BSA (Fig. 6). Under the same condition, SOD exhibited negligible binding with rose bengal (Fig. 6).
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
3304
Investigative Ophthalmology & Visual Science, July 1994, Vol. 35, No. 8
Catalase
3
3 •
2
2 •
562nm
548nm
0
0
30
60
90
120
150
0-
o
30
$0
90
120
0
150
1.5
0
0
30
60
90
.
120
150
3.
562nm
280nm
30
60
90
120
150
30
60
90
120
150
FIGURE 6. Sephadex G-75 gel chromatography. Conditions of column chromatography were
described in Methods. Addition of 8 X 10~5 M BSA clearly demonstrated two peaks. The first
sharp peak was eluted at void volume and had a relatively higher absorbance for X 562 nm
and X 280 nm, indicating the presence of bound rose bengal with BSA. The second broad
peak eluted at the fractions # 85 to # 95 had a relatively higher absorbance for X 548 nm than
for X 562 nm, representing the free rose bengal. A similar elution profile was obtained for 9
X 10~5 M catalase, except that the absorbance for X 562 nm of the first peak was less. For
SOD, 8 X 10~5 M yielded negligible amounts of bound rose bengal. The total volume calibrated by fluorescein is at fraction # 30. The elution of free rose bengal was three times
delayed.
DISCUSSION
When higher concentrations of rose bengal were used
beyond the intrinsic toxicity level of 1 X 10~5 M, a
unique photobiologic effect, i.e., retractive monolayer
separation, was observed under argon green laser irradiation at the intensity of 16 W/cm 2 . This photobiologic effect followed a qualitative dose-response relationship with respect to rose bengal concentration and
laser energy fluence (Fig. 1). Because cultured cells are
devitalized and intracellular components are stained
at this range of rose bengal concentrations, 22 the observed photobiologic effect resulted from photodynamic action generated by rose bengal bound to intracellular components. The bound status of rose bengal
was confirmed by the failure to reproduce the fluorescence change and monolayer separation with polymerconjugated rose bengal, which precludes rose bengal
from entering the cells (Fig. 3). Furthermore, competition with this bound status by BSA added to rose
bengal solutions reduced its photobiologic effect (Fig.
4). It is particularly interesting to note that bixin,
DMSO, or sodium azide were ineffective, suggesting
that the bound status of rose bengal with intracellular
components might render these quenchers ineffective. This possibility is supported by the studies of Oxford et al,27 which showed that /3-carotene cannot inhibit the photodynamic effect of rose bengal when
rose bengal was internally applied to the axons. An-
other possibility is that the bound status of rose bengal
might render the targeted cellular components more
susceptible to the photodynamic action of rose bengal,
a concept elucidated in a review by Valenzeno and
Pooler, 5 or that this photobiologic effect is mediated
by a photochemical action not involving singlet oxygen
or oxygen free radicals (also see below).
The bound status of rose bengal was also revealed
by the red shift of the maximal absorption spectrum
from 548 nm for the free rose bengal in aqueous solution 28 to 562 nm noted in the irradiation cell area (Fig.
5, upper graph). A similar red shift of the absorption
spectrum has been reported when several halide derivatives of fluorescein, including rose bengal, interact
with cell membranes, 8 and when rose bengal binds to
several different purified proteins. 29 " 31 Based on the
studies by the laboratory of Neckers, 25 ' 28 the red shift
of rose bengal can be enhanced primarily when the
C-6 phenoxide is fully ionized and partly when the C-2'
carboxyl group is conjugated, for example, in esterification. Therefore, it is tempting to speculate that the
binding of rose bengal with cellular components might
reinforce the ionized state of C-6, the conjugated state
of C-2', or both. Previously, rose bengal has been reported to bind to several proteins, 29 " 34 and the binding
domain of some of these proteins is thought to be
hydrophobic. 2931
In this report, we demonstrated the binding capacity of BSA or catalase with rose bengal by gel filtration
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
3305
Photodynamic Action of Rose Bengal In Vitro
chromatography (Fig. 6). Because the binding of rose
bengal with cellular components appears to dictate the
severity of the observed photodynamic effects, one
can imagine that prevention of rose bengal entry into
cells by binding with exogenous proteins, such as albumin or catalase, can reduce the damaging effect arising from either intrinsic toxicity or photodynamic action (Fig. 4). A similar inhibitory effect with albumin
was previously reported.6 This albumin binding also
explains why intravenously delivered rose bengal is
principally cleared via the liver,26 thereby eliminating
systemic toxicity.14
Previously, Hull et al9 reported that 1 X 10~4 g/ml
SOD cannot, but 2 X 10"4 g/ml catalase can, eliminate
corneal swelling induced by the photodynamic action
of 5 X 10~6 M rose bengal. Ver Donck et al6 noted that
2500 U/ml SOD cannot, but 1500 U/ml catalase can,
completely inhibit the shape change of cardiomyocytes
induced by 5 X 10~8 M rose bengal. Interestingly, they
also found that heat-inactivated or aminotriazole-inhibited catalase also provides protection.6 Recently,
Bernier et al35 reported that 80 U/ml SOD does not
inhibit cardiac arrhythmias induced by 10~6 M rose
bengal. As shown by gel filtration chromatography
(Fig. 6), both catalase and BSA had a noticeable binding capacity with rose bengal that was much higher
than that with SOD. We thus speculate that their inhibitory effects are primarily derived from preventing
rose bengal from binding to cellular components owing to competitive binding by these extracellularly
added proteins. This notion not only raises a caution
on data interpretation when these two enzymes are
used to probe the action mechanism of an observed
photodynamic effect but also emphasizes the importance of the bound status of rose bengal with cellular
components in inducing the observed phototoxicity.
The photobiologic effect of retractive monolayer
separation created by rose bengal under high-intensity
laser irradiation was similar in five types of cultured
cell (Fig. 1), suggesting that the underlying action
mechanism might be the same in these different cell
components of the blood vessel. Compared to in vivo
photothrombosis, a similar irradiation fluence and
rose bengal concentration were used in this in vitro
study to characterize observable photobiologic effects.
These effects include fluorescence spectral changes
(Figs. 1, 2, and 5), and monolayer separation (Fig. 1) in
the irradiated cell area. After laser treatment, the fluorescence spectral changes noted predominantly at 5 X
10~4 M rose bengal were characterized by marked dye
photobleaching and a slight blue shift of the emission
spectrum (Fig. 1 and 5). Linden and Neckers25 reported that photobleaching of rose bengal is the result
of an electron transfer process in either oxidation or
reduction. Because there was a strong binding between rose bengal and various intracellular compo-
nents, the photobleaching of rose bengal in the irradiated cell area signifies the presence of accompanied
oxidation or reduction of the bound materials. This
view is consistent with the early report that some
amino acids, such as histidine, tryptophan, tyrosine,
methionine, and cysteine, are sensitive to photosensitized oxidation.36 For rose bengal, histidine30'32"34 and
tryptophan3137 have been reported to be preferentially photooxidized. Future studies are needed to determine which cellular proteins have a higher binding
affinity with rose bengal and whether the photooxidation of these proteins can be linked to the observed
intrinsic toxicity and/or phototoxicity, as well as photobiologic effects.
It is intriguing to observe the monolayer separation in the irradiation cell area at 2.5 X 10~3 M and 5 X
10~3 M rose bengal under high-intensity argon laser
irradiation. This monolayer separation was likely mediated by cellular contraction (Fig. 2) and, thus, mimicked the phenomenon of blood vessel constriction
observed during in vivo photothrombosis of arteries.13"19 Because this monolayer separation was not
caused by a thermal effect (Tang, unpublished observation, 1992) and occurred when rose bengal was
bound to intracellular organelles, it is tempting to
speculate that this phenomenon might result from the
activation of cellular contractile machinery. For nonmuscle cells, initiation of cellular contraction is controlled by the structural protein myosin, which has
Mg2+-ATPase activity, and activity is further controlled by the intracellular calcium concentration and
phosphorylation.38 This type of cellular contraction
can take place even when cells are permeated by glycerol 3 ' or detergent,40 suggesting that rose bengal-induced devitalization might permit such contraction to
occur in the cultured cells. Erythrosin B can inhibit
ATP-dependent calcium transport41-42 and increase
membrane permeability to calcium.43 It is noteworthy
that photoactivation of erythrosin B can cause a
marked calcium-dependent contraction of the smooth
muscle cells of guinea pig taenia coli in vivo.44 For rose
bengal, photosensitized oxidation of sarcoplasmic reticulum vesicle membrane also inhibits calcium uptake
and ATPase activity.37 Interestingly, like erythrosin B,
rose bengal can induce hypercontraction of cultured
cardiomyocytes, which can be counteracted by some
Ca2+-channel blockers.6'45 These Ca2+ antagonists can
also protect against rose bengal-induced cardiac arrhythmias35 and erythrosin B-induced smooth muscle
contraction.44 Future studies are needed to determine
whether disturbances of the Ca2+ flux and/or the cellular contractile machinery are involved in this retractive monolayer separation. If so, the binding of rose
bengal with such cellular components may then be
linked to these disturbances. Research in this direction
should clarify the action mechanism by which rose
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
3306
Investigative Ophthalmology & Visual Science, July 1994, Vol. 35, No. 8
bengal induces its intrinsic toxicity and photodynamic
actions in biologic systems and may provide a better
understanding of the phenomenon of photothrombo-
14.
sis.
Key Words
15.
cell culture, cell contraction, photodynamic effect, rose bengal
References
1. PassmoreJW, KingJHJr. Vital staining of conjunctiva
and cornea; review of literature and critical study of
certain dyes. Arch Ophlhalmol. 1955;53:568-574.
2. Roat MI, Romanowski E, Araullo-Cruz T, Gordon YJ.
The antiviral effects of rose bengal and fluorescein.
Arch Ophthalmol. 1987; 105:1415-1417.
3. Dahl TA, Midden WR, Neckers DC. Comparison of
photodynamic action by rose bengal in gram-positive
and gram-negative bacteria. Photochem Photobiol.
1988;48:607-612.
4. Cruz FS, Lopes LAV, DeSouza W, Moreno SNJ, Mason RP, Docampo R. The photodynamic action of
rose bengal on Trypanosoma cruzi. Ada Trop.
1984;41:99-108.
5. Valenzeno, DP, Pooler JP. Cell membrane photomodification: Relative effectiveness of halogenated
fluoresceins for photohemolysis. Photochem Photobiol.
1982;35:343-350.
6. Ver Donck L, Van Reempts J, Vandeplassche G,
Borgers M. A new method to study activated oxygen
species induced damage in cardiomyocytes and protection by Ca2+-antagonists. J Mol Cell Cardiol.
1988;20:811-823.
7. Menon IA, Basu PK, Persad SD, Rosatone S, Wiltshire
JD. A study on the sequence of phototoxic effects of
rose bengal using retinal pigment epithelial cells in
vitro. Exp Eye Res. 1989;49:67-73.
8. Pooler JP, Valenzeno DP. Kinetic factors governing
sensitized photooxidation of excitable cell membranes. Photochem Photobiol. 1978;28:219-226.
9. Hull DS, Strickland EC, Green K. Photodynamically
induced alteration of cornea endothelial cell function.
Invest Ophthalmol Vis Sci. 19 7 9; 18:12 2 6-12 31.
10. Hearse DJ, Kusama Y, Bernier M. Rapid electrophysiological changes leading to arrhythmias in the aerobic
rat heart: Photosensitization studies with rose bengalderived reactive oxygen intermediates. Circ Res.
1989;65:146-153.
11. Matthews EK, Cui ZJ. Photodynamic action of rose
bengal on isolated rat pancreatic acini: Stimulation of
amylase release. FEBS Lett. 1989;256:29-32.
12. Watson, BD, Dietrich WD, Busto R, Wachtel MS,
Ginsberg MD. Induction of reproducible brain infarction by photochemically initiated thrombosis. Ann
Neurol. 1985; 17:497-504.
13. Watson BD, Dietrich WD, Prado R, Ginsberg MD. Argon laser-induced arterial photothrombosis: Characterization and possible application to therapy of arte-
16.
17.
18.
19.
riovenous malformations. J Neurosurg. 1987;66:748754.
Mendelsohn AD, Watson BD, Alfonso EC, et al. Amelioration of experimental lipid keratopathy by photochemically induced thrombosis of feeder vessels. Arch
Ophthalmol 1987; 105:983-988.
Huang AJW, Watson BD, Hernandez E, Tseng SCG.
Photothrombosis of corneal neovascularization by intravenous rose bengal and argon laser irradiation.
Arch Ophthalmol. 1988; 106:680-685.
Corrent G, Roussel TJ, Tseng SCG, Watson BD. Promotion of graft survival by photothrombotic occlusion of corneal neovascularization. Arch Ophthalmol.
1989;107:1501-1506.
Nanda SK, Hatchell DL, Tiedeman JS, Dutton JJ,
Hatchell MC, McAdoo T. A new method for vascular
occlusion: Photochemical initiation of thrombosis.
Arch Ophthalmol. 1987;105:l 121-1124.
Royster AJ, Nanda SK, Hatchell DL, Tiedeman JS,
Dutton JJ, Hatchell MC. Photochemical initiation of
thrombosis: Fluorescein angiographic, histologic, and
ultrastructural alterations in the choroid, retinal pigment epithelium, and retina. Arch Ophthalmol.
1988;106:1608-1614.
Lewis ML, Dabbs C, Watson BD, Hernandez E. Photochemical thrombosis of retinal and choroidal vessels
using rose bengal. In: Joffe SN, Goldblatt NR, Atsumi
K, eds. Laser Surgery: Advanced Characterization, Thera-
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
peutics, and Systems. Bellingham, WA: Soc Photo-Optical Instrumentation Engineers; 1989:29-33.
Plato CC, Rodgers MAJ. Laser flash photokinetic studies of rose bengal sensitized photodynamic interactions of nucleotides and DNA. Photochem Photobiol.
1987;45:79-86.
Valenzeno DP. Photomodification of biological membranes with emphasis on singlet oxygen mechanisms.
Photochem Photobiol. 1987;46:147-160.
Feenstra RPG, Tseng SCG. What is actually stained by
rose bengal? Arch Ophthalmol. 1992;110:984-993.
Schaap AP, Blossey EC, Thayer AL, Neckers DC. Polymer-based sensitizers for photooxidations: II. J Am
ChemSoc. 1975;97:3741-3745.
Thompson A, Nigro J, Seliger HH. Efficient singlet
oxygen inactivation of firefly luciferase. Biochem
Biophys Res Comm. 1986; 140:888-894.
Linden SM, Neckers DC. Type I and type II sensitizers
based on rose bengal onium salts. Photochem Photobiol.
1988;47:543-550.
Meurman L. On the distribution and kinetics of injected I131 rose bengal: An experimental study with
special reference to the evaluation of liver function.
Acta Med Scand. 1960;354(suppl): 18-20.
Oxford GS, Pooler JP, Narahashi T. Internal and external application of photodynamic sensitizers on
squid giant axons. J Membr Biol. 1977;36:159-173.
Paczkowski J, Lamberts JJM, Paczkowska B, Neckers
DC. Photophysical properties of rose bengal and its
derivatives (XII). / Free Radic Biol Med. 1985;1:341351.
Brand L, Gohlke JR, Rao DS. Evidence for binding of
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017
Photodynamic Action of Rose Bengal In Vitro
30.
31.
32.
33.
34.
35.
36.
37.
38.
rose bengal and anilinonaphthalenesulfonates at the
active site regions of liver alcohol dehydrogenase. Biochemistry. 1967;6:3510-3518.
Bond JS, Francis SH, ParkJH. An essential histidine
in the catalytic activities of 3-phosphoglyceraldehyde
dehydrogenase. J Biol Chem. 1970;245:1041-1053.
Wu CW, Wu FYH. Rose bengal: A spectroscopic
probe for ribonucleic acid polymerase. Biochemistry.
1973;12:4349-4355.
Westhead EW. Photooxidation with rose bengal of a
critical histidine residue in yeast enolase. Biochemistry.
1984;10:2139-2144.
Groudinsky, O. Study of heme-protein linkage in cytochrome b2: Destruction of a crucial histidine residue
by photooxidation of 'apo' cytochrome b2 core in the
presence of rose bengal. EurJ Biochem. 1971;18:480484.
Seery VL, Morgan WT, Muller-Eberhard U. Interaction of rabbit hemopexin with rose bengal and photoxidation of the rose bengal-hemopexin complex. J
Biol Chem. 1975;250:6439-6444.
Bernier M, Kusama Y, Borgers M, et al. Pharmacological studies of arrhythmias induced by rose bengal photoactivation. Free Radical Biol Med. 1991 ;10:287-296.
Spikes JD, Glad BW. Photodynamic action. Photochem
Photobiol. 1964;3:471-487.
Kondo, M, Kasai M. Photodynamic inactivation of sarcoplasmic reticulum vesicle membranes by xanthene
dyes. Photochem Photobiol. 1974;19:35-41.
Adelstein RS, Klee CB. Purification and characteriza-
3307
39.
40.
41.
42.
43.
44.
45.
tion of smooth muscle myosin light chain kinase./ Biol
Chem. 1981;256:7501-7509.
Goldman RD, Schloss JA, Starger JM. Organizational
changes of actin-like microfilaments during animal
cell movement. In: Goldman R, Pollard T, Rosenbaum J, eds. Cell Motility. Cold Spring Harbor, NY:
Cold Spring Harbor Conferences on Cell Proliferation. Vol. 3. 1976:217-245.
Colombini M, Wu YC. A food dye, erythrosine B, increases membrane permeability to calcium and other
ions. Biochim Biophys Ada. 1981;648:49-54.
Morris SJ, EK Silbergeld, RR Brown, Haynes DH.
Erythrosin B (USFD&C red 3) inhibits calcium transport and ATPase activity of muscle sarcoplasmic reticulum. Biochem Biophys Res Comm. 1982; 104:13061311.
Heffron JJA, O'Callaghan AM, Duggan PF. Food dye,
erythrosin B, inhibits ATP-dependent calcium ion
transport by brain microsomes. Biochem Int.
1984;9:557-562.
Burnside B, Smith B, Nagata M, Porrello K. Reactivation of contraction in detergent-lysed teleost retinal
cones./ Cell Biol. 1982;92:199-206.
Matthews EK, Mesler DE. Photodynamic effects of
erythrosine on the smooth muscle cells of guinea-pig
taenia coli. Br J Pharmacol. 1984;83:555-566.
Ver Donck, L, Verellen G, Vandewiele R, De Clerck F,
Borgers M. Class IV Ca2+ antagonists do not affect
lipid peroxidation in singlet oxygen challenged cardiomyocytes. Cell Biol hit Rep. 1990; 14:399-406.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933406/ on 06/15/2017