Journal of Photochemistry and Photobiology B: Biology 79 (2005) 25–34
www.elsevier.com/locate/jphotobiol
Evaluation of phototoxic and photoallergic potentials of
13 compounds by different in vitro and in vivo methods
Norbert J. Neumann a,*, Andrea Blotz a,b, Grazyna Wasinska-Kempka b,
Martin Rosenbruch b, Percy Lehmann a, Hans Jürgen Ahr b, Hans-Werner Vohr
a
b
Hautklinik, Heinrich-Heine-Universität, Moorenstrasse 5, 40225 Düsseldorf, Germany
b
Institut für Toxikologie der Bayer HealthCare AG, Wuppertal, Germany
Received 1 September 2004; received in revised form 9 November 2004; accepted 10 November 2004
Available online 12 January 2005
Abstract
Phototoxic side effects of pharmaceutical and cosmetic products are of increasing concern for patients, dermatologists and the
chemical industry. Moreover, the need of new chemicals and drugs puts pressure on pre-clinical test methods for side effects, especially interactive adverse-effects with UV-light. So, the predictive potential of different established test methods, which are used regularly in our departments in order to detect the phototoxic potential of chemicals, were analyzed. Namely the fibroblast 3T3 test, the
photo henÕs egg test, a guinea pig test for measuring acute photoreactions, and a modified Local Lymph Node Assay, the Integrated
Model for the Differentiation of Skin Reactions. Various agents with different photoreactive potential were tested: quinolones like
Bay y 3118, ciprofloxacin, enoxacin, lomefloxacin, moxifloxacin, ofloxacin, sparfloxacin, as well as promethazine, chlorpromazine,
8-methoxypsoralen and olaquindox serving as control. Special emphasis was taken to evaluate the capability of the employed test
procedures to predict phototoxic side effects in patients. Following our results, both in vitro assays were useful tools to detect photoirritancy while the photoallergic potentials of tested compounds were exclusively detected by an in vivo assay.
As long as no in vitro model for photoallergy is available, the UV-IMDS should be considered to evaluate photoallergic properties of a supposed photoreactive agent.
2004 Elsevier B.V. All rights reserved.
Keywords: Phototoxic; Photoallergic; Photosensitizer; Quinolone; Photo henÔs egg test; In vitro phototoxicity assay; Acute phototoxicity assay;
Integrated model for the differentiation of skin reactions
1. Introduction
Phototoxic side effects of pharmaceutical and cosmetic products are of increasing concern for patients,
dermatologists and the chemical industry. The growing
use of toiletries and cosmetics in combination with relative high UV-light exposure potentiate this problem.
Photoreactions to exogenous agents occur mostly in
response to ultraviolet A (UVA) light (315–400 nm)
*
Corresponding author. Tel.: +49 211 811 7620; fax: +49 211 811
7617.
E-mail address: neumannt@uni-duesseldorf.de (N.J. Neumann).
1011-1344/$ - see front matter 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jphotobiol.2004.11.014
rather than to shorter wavelengths out of the UVB
or UVC band. These photoreactions can be divided
into two phenomena: photoirritancy (phototoxicity)
and photoallergy. Photoirritancy can be elicited by a
wide range of pharmaceutical agents or their metabolites like chlorpromazine, quinine, NSAIDS, phenylbutazone and various anti-infective agents [1–3].
Anti-infective agents like griseofulvin, tetracyclines
and fluoroquinolones are well known to induce mainly
photoirritancy.
Acute phototoxic reactions have their maximum
intensity in the beginning followed by a decrescendo
reaction within 24–72 h after UV exposure. They are
26
N.J. Neumann et al. / Journal of Photochemistry and Photobiology B: Biology 79 (2005) 25–34
much more prevalent than photoallergic reactions and
they are very similar to sunburn (dermatitis solaris)
which is characterized by erythema, infiltration, oedema
or blisters followed by desquamation and hyperpigmentation. Chronically repeated phototoxic damages may
lead to fragility, blistering and milia formation, or even
actinic keratoses and skin cancers. The underlying photochemical mechanisms differ completely from that supposed in photosensitization (photoallergy) [2].
In contrast, photoallergic reactions are uncommon
and immunologically mediated [3–5]. Clinically, they
are very similar to plain contact allergic reactions characterized by a delayed onset and a crescendo reaction
pattern. The photoallergic skin lesions are composed
of erythema, infiltration, and papulovesicles erosions
or bullae.
UVA may cause structural and activation changes in
the UV-light absorbing molecules, leading to one or
more photochemically converted stable reactions products. In combination with carrier molecules (proteins)
these reactions products (haptens) represent complete
antigens. Once a hapten–protein complex has been
formed, epidermal LangerhansÕ cells can present the
antigen to immunocompetent cells, thereby inducing
sensitization [1–5]. This is comparable to plain contact
allergic reaction (type IV reaction), with the UVA induced photo-product carrier molecule combination acting as the sensitizing agent.
The potential of quinolones to induce phototoxicity
(photoirritancy) is a well-documented side effect in
man and has been demonstrated in a variety of experimental studies [6–13]. However, quinolones differ considerably in their extent of phototoxicity in animals as
well as in man. While Ciprofloxacin leads to low incidences, other quinolones like pefloxacin, lomefloxacin
and sparfloxacin appear to be associated with higher
incidences of phototoxic side effects observed in clinical
use [14–19]. During the last years several in vitro models
have been developed to screen photoreactivity of chemicals. Here, we compare different in vitro (3T3 and
photo henÕs egg test) and in vivo systems (guinea pig
and mice) for the detection of phototoxic potentials of
test compounds. One of these tests has reached an internationally accepted status, the so-called 3T3 neutral red
uptake (NRU) Assay [20–22].
2. Materials and methods
2.1. Animals
2.1.1. Guinea pigs
Four- to six-week old outbred female DHPW guinea
pigs (Winkelmann, Borchen, Germany) with an initial
weight range of 190–250 g were used after at least 5 days
of acclimatization.
2.1.2. Mice
This study used 8–12 weeks old, outbred female
NMRI mice with an initial weight range from 27 to 32
g after at least one week of acclimatization. The animals
were maintained on a freely available standard diet and
water ad libitum and were used throughout our study in
accordance to the German animal protection legislation.
2.1.3. HenÕs eggs
Fertile white Leghorn eggs (Shaver Starcross 288A,
Lohmann, Cuxhaven, Germany) were incubated in a
horizontal position using a commercial incubator at
37.5 C and 65% relative humidity. After three days of
incubation, all eggs were candled in order to discard
those that were defect.
2.2. Test compounds
For evaluation of the laboratory test systems the following agents were chosen: Well-known photo irritating
agents: 8-methoxypsoralen, Bay y 3118, promethazine
and chlorpromazine sparfloxacin, lomefloxacin, enoxacin and ciprofloxacin. As photoallergy inducing standard we used olaquindox and PBS as negative control
(non-photoxic agent).
All test substances used in the studies were obtained
from Sigma (Deisenhofen, Germany) with the exception
of ciprofloxacin, Bay y 3118 and olaquindox, which
were provided by Bayer HealthCare AG (Wuppertal,
Germany).
All chemicals were dissolved to the final concentrations in double distilled water just before the oral
application.
For dermal application, the test compounds were dissolved in a mixture of dimethylacetamide (40%), acetone
(30%) and ethanol (30%) (DAE 433), as described by
Maurer [23]. Vehicle compounds were obtained from
Merck (Darmstadt, Germany) and Sigma Chemicals
(Diesenhofen, Germany).
For application on the yolk sac blood vessel system,
the test compounds were dissolved in physiological salt
solution (PSS).
2.3. Treatment protocols
2.3.1. In vitro phototoxicity assay – mouse 3T3 fibroblasts
The BALB/c mouse fibroblast cell line 3T3.A31 was
obtained from the ATCC. The cells were cultivated in
DMEM (without phenol red) containing 10% fetal calf
serum, 1% glutamine and 1% penicillin/streptomycin.
The NRU assay was adapted to determine phototoxicity as follows. The test was performed in 96-well
microtitre plates. Twenty-four hours after seeding, the
mouse fibroblast cells were pre-incubated in the growth
medium for 1 h and then irradiated for 50 min in the
presence of different concentrations of the test
27
N.J. Neumann et al. / Journal of Photochemistry and Photobiology B: Biology 79 (2005) 25–34
compounds (0, 1, 3, 10, 30 or 100 mg/l). All plates, except the control plates, were irradiated with UVA using
a mercury halide lamp fitted with an H1 filter and an
intensity of 1.67 mW/cm2 (5 J/cm2).
After UV exposure, the existing growth medium was
replaced by fresh medium. Cytotoxic effects were determined 24 h later by measuring the NRU. Neutral red
solution (33 mg/ml) was added to the cell cultures (50
ll per well) and incubated for 2 h at 37 C. The cells
were then washed with PBS and the neutral red dye extracted from the lysosomes using a mixture of ethanol
and acetic acid in ice (50% ethanol, 1% acetic acid).
The number of cells remaining viable was measured
using a spectrophotometer (ICN-Flow, Meckenheim,
Germany) at 540 nm.
The phototoxic effect was evaluated according to the
OECD 432 guideline draft.
A Photo-Irritation-Factor (PIF) was calculated using
the following formula:
PIF ¼ IC50ðIrrÞ=IC50 ðþIrrÞ
where ‘‘Irr’’ means the determination of cytotoxicity,
i.e., the non-irradiated control plate, while ‘‘+Irr’’
stands for the determination of photocytotoxicity.
The significance of differences between study groups
in particular experiments was analysed using the twotailed non-parametric Mann–Whitney U-test.
2.3.2. The photo henÕs egg test
Based on the henÕs egg test (HET), originally introduced by toxicologists as a screening model for mucocutaneous toxicity [24–26] as an alternative to the rabbitÕs
eye test (Draize test), Neumann et al. employed the yolk
sac blood vessel system (YS) of the incubated henÕs egg
in combination with UV-irradiation as a new valid and
inexpensive screening model for phototoxicity [27–31].
Fertile white Leghorn eggs (Shaver Starcross 288A,
Lohmann, Cuxhaven, Germany) were incubated in a
horizontal position using a commercial incubator at
37.5 C and 65% relative humidity. After three days of
incubation, all eggs were candled in order to discard
those that are defective. Without damaging the shell
membrane, a hole had to be drilled into the shell,
through which 5 ml of egg white had to be sucked out
to lower the embryo and its surrounding YS. Afterwards
a 1.5 · 2.5 cm window had to be sawed out of the shell.
The eggs were covered with a wax sheet and had to be
placed back into the incubator. At day 4 of incubation,
only eggs with normally developed embryos and YS
were used for testing.
The PHET has a 2 · 2 factorial test design with the
factors ‘‘irradiation’’ and ‘‘substance application’’ and
the levels ‘‘yes’’ and ‘‘no’’ (Table 1).
Since the PHET was established to investigate phototoxic reactions, the yolk sac blood vessel system of the
incubated henÕs eggs were exposed only to nontoxic con-
Table 1
PHET 2 · 2 factorial test design (n = number of yolk sac blood vessel
systems)
Irradiation
Yes
No
Substance application
Yes
No
n = 12
n = 12
n = 12
n = 12
centrations of the test substances and to a nontoxic
UVA dose (5 J/cm2) concurrently.
At day 4 of the incubation period, test group I (each
group consisted of 12 eggs) was exposed to a test substance, immediately followed by an irradiation with 5
J/cm2 UVA (320–400 nm, Philips TL 09/40W, Hamburg, Germany). PSS was normally used as vehicle.
Serving as controls, three additional test groups were exposed only to PSS and 5 J/cm2 UVA or to PSS or to a
test substance alone. Readings were performed 24 h
after irradiation. During this observation period, the
morphological parameters such as membrane discoloration (MD), hemorrhage (HR) were monitored via a
macroscope (M 420, Leitz, Wetzlar, Germany) and
graded following a four point scale:
Level 0: No visible MD or HR
Level 1: Just visible MD or HR
Level 2: Visible MD or HR, structures are covered
partially
Level 3: Visible MD or HR, structures are covered
totally
Additionally, the embryo lethality was assessed. The
test parameters MD and HR as well as embryo lethality
can be summarized in a morphology and a lethality index (Table 2). Employing these indices, the relative phototoxic potential of an assumed photosensitizer
compared to other well-known photosensitizers could
be calculated.
2.3.2.1. Statistical analysis. In order to analyze the frequencies statistically, non-parametric tests (the contingency-table test for ordered categories and the FisherÕs
contingency-table test) were employed.
2.3.3. Acute phototoxicity assay in guinea pigs
A photoreactive assay was carried out in groups of
five guinea pigs after the administration of a single oral
dose, by gavage, of 0, 30 or 100 mg/kg of the test compounds, with or without additional UVA irradiation.
Animals were irradiated with 20 J/cm2 UVA (see below
for details) 30 min after dosing. Hairless (shaved) skin
on the back, as well as the ears, nose and eyelids were
evaluated for reddening, swelling, oedema and necrosis
1, 3, 24, 48, 72 and 96 h after irradiation using the following scoring system:
28
N.J. Neumann et al. / Journal of Photochemistry and Photobiology B: Biology 79 (2005) 25–34
Table 2
Photo henÕs egg test
Relative lethality:
Lethality rate of the interaction group (Li) minus the average lethality rate of controls LC = (lc1 + lc2 + lc3)/3
Relative lethality L = Li Lc
Relative lethality L (%) = L · 100
12
Relative hemorrhage I and II:
Relative
Relative
Relative
Relative
0
1
2
3
4
Sums of the hemorrhage(HR) levels of the interaction group HRG1i and HRG2i:
Frequencies of level 0 HR + level 1 HR = HRG1i
Frequencies of level 2 HR + level 3 HR = HRG2i
Sums of the hemorrhage(HR) levels of the controls:
Frequencies of (level 0 HR + level 1 HR = HRG1c
Frequencies of level 2 HR + level 3 HR = HRG2c
hemorrhage I (HR I) = HRG1i–[(HRG1c)/3]
hemorrhage II (HRII) = HRG2i–[(HRG2c)/3]
Sums of the membrane discoloration (MD) levels of the interaction group MDG1i and MDG2i:
Frequencies of level 0 MD + level1 MD = MDG1i
Frequencies of level 2 MD + level 3 MD = MDG2i
Sums of the membrane discoloration (MD) levels of the controls:
Frequencies of (level 0 MD + level1 MD = MDG1c
Frequencies of level 2 MD + level 3 MD = MDG2c
membrane discoloration I (MD I) = MDG1i–[(MDG1c)/3]
membrane discoloration II (MD II) = MDG2i–[(MDG2c)/3]
No reaction
Just detectable reddening
Reddening
Strong sunburn reaction
Extreme reaction/swelling of the skin
The daily scores for each animal were averaged and
the mean values calculated. The weight of the animals
was recorded each day. For irradiation of animals, a
Hönle Xenon UV lamp (sol3-h1) with a Hönle H1
UV-filter system (Hönle GmbH, Planegg, Germany)
was used. The emission spectrum of this light source included visible light and UVA (range 320–400 nm, with
an emission peak at 380 nm), and less than 1% UVB.
The emission was monitored routinely at the skin surface of the animals, about 40 cm from the light source,
with a Hönle UVA/B meter.
2.3.4. The integrated model for the differentiation of skin
reactions
On three consecutive days 5 mice per group were treated orally with the test compound or the vehicle. 30 min
after each treatment an irradiation step with an irradiation of 20 J/cm2 (UVA including about 1% UVB) followed. Irradiation was done in accordance to the draft
proposal of the OECD guideline for the acute dermal
photoirritation test with little modification. For the
dose-response tests we used a Dermalight vario 2 UVlamp equipped with a Dr. Hönle H1 UV-filter system
(Dr.Hönle GmbH, Planegg, Germany). The ears and
ear-draining auricular lymph nodes (LN) were excised
24 h after the last exposure.
Ear swelling was determined at the helical edge of the
ears via Oditest micrometer (Dyer, Lancester, USA).
Measurement was carried out before the first treatment
and immediate before sacrificing the mice.
The weight of the paired auricular LN of each animal
was obtained as described previously [32–34]. To determine the individual LN cell counts, single cell suspensions of the paired LN of each mouse were prepared
by mechanical tissue disaggregation and measured by
conductometry (Coulter cell counter Z 1, Coulter electronics, Krefeld, Germany) gating on a particle diameter
above 4.34 lm.
2.3.4.1. Calculation of the IMDS differentiation index.
The underlying principles of the Integrated Model for
the Differentiation of chemical-induced allergic and
irritant Skin reactions are described elsewhere [33].
The Differentiation Index (DI) describes the relation
between the activation of the local skin-draining lymph
nodes and the skin inflammation at the site of topical
treatment. A DI > 1 indicates an allergic reaction,
whereas a DI < 1 demonstrates an irritant potential
of the tested compounds at the concentration tested.
For skin inflammation, the maximum increase in ear
thickness was defined as 15 · 0.01 mm, i.e., doubling
of thickness. For the activation of the skin draining
lymph nodes the maximum lymph node cell counts index was 5. The defined maximal values were set as
100%.
2.3.4.2. Statistical analysis. The increase in ear thickness
is expressed as the mean difference between d0 and
d3 ± SD. The LN proliferation is expressed as the
LLN Index ± SD, calculated as described previously
[33,34]. Statistical significance of differences between
the groups was analyzed on the basis of individual data
29
N.J. Neumann et al. / Journal of Photochemistry and Photobiology B: Biology 79 (2005) 25–34
of each mouse using the two-tailed StudentÕs t test. Pvalues less than 0.05 were considered to be significant.
3. Results
3.1. In vitro mouse fibroblast phototoxicity assay (NRU)
Following the irradiation of fibroblast cells, phototoxic reactions were evaluated by calculating the photoirritation factor (PIF). After an UVA exposure of 5
J/cm2, there was a marked difference between the compounds tested (Table 3). Moxifloxacin, tetracycline, ciprofloxacin and olaquindox had either no or faint effects
on cell viability, but a marked phototoxic reaction was
observed in the presence of sparfloxacin (PIF 10 lg/ml)
and protoporphyrine IX (PIF 16.7 lg/ml). Olaquindox
turned out to have only faint impact on the cell viability
after UVA light exposure with 5 J/cm2 in this assay.
UV exposure of 5 J/cm2 resulted in pronounced phototoxic reactions induced by Bay y 3118, meladinin, 8methoxypsoralen (8-MOP), and acridin with far higher
PIF values (P20 lg/ml).
3.2. The photo hen’s egg test
Based on the PHET results, the ranking list of several
photosensitizers is depicted in Table 4. Bay y 3118 and
8-MOP were the leading phototoxic agents followed
by promethazine and sparfloxacin. On the other side,
ciprofloxacin revealed a weak response while moxifloxacin and olaquindox showed no detectable phototoxic effects in the PHET.
3.3. Photoreaction after single oral application in guinea
pigs
Throughout this study, animals treated with vehicle
alone and animals that received quinolones without
Table 3
Phototoxicity scaling in accordance to the PIF factor (100 lg/ml max)
(3T3 test)
Substance
PIF value
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
50.0
33.3
33.3
20.0
16.7
10.0
9.2
4.0
3.3
2.5
1.7
1.0
1.0
Acridin
8-MOP
Meladinin
Bay y 3118
Protoporphyrin
Sparfloxacin
Promethazin
Lomefloxacin
TCSA
Olaquindox
Ciprofloxacin
Tetracyclin
Moxifloxacin
Table 4
Ranking list photoxicity (Photo henÕs egg test)
Substance
Relative
lethality
Relative
HR II
Relative
MD II
Concentration
(mol)
8-Methoxypsoralen
BAY y 3118
Protoporphyrine IX
Promethazine
Sparfloxacin
Acridine
TCSA
Hematoporphyrine
Lomefloxacin
Ciprofloxacin
Tetracyclin
Moxifloxacin
Olaquindox
100
100
94.5
94.5
58.33
58.33
44.44
41.67
30.56
0
0
0
<0
12
9
7
10
6.33
5
2.33
6
4.67
2
0
0
0
9
11.67
12
8.67
7.33
6.67
4.33
8.33
6.67
5
0.33
0.33
0
105
104
104
103
105
105
106
105
105
6.035 · 103
103
104
103
additional irradiation did not exhibit any skin reaction
(data not shown).
The results in irradiated animals are shown in Fig.
1. After single oral doses of quinolones (30 or 100
mg/kg) plus UVA irradiation, different grades of sunburn-like reaction were found in treated animals as
compared to the control group. The gut flora of guinea
pigs is especially sensitive to the bactericidal effect of
some antiinfectives. Therefore, the body weights had
been recorded until the end of the observation period
to exclude toxic effects. The body weights of treated
and control animals did not differ statistically (data
not shown).
No photoreaction was detected after single oral doses
of up to and including 100 mg moxifloxacin/kg
bodyweight.
A single oral dose of 30 mg/kg ciprofloxacin plus
UVA irradiation resulted in a slight sunburn-like reaction in the guinea pigs, which decreased rapidly over
the 24 h following light exposure. At a dose of 100
mg/kg, ciprofloxacin produced a slightly greater response, which subsided after 48 h.
In contrast to the findings with moxifloxacin and
ciprofloxacin, guinea-pigs that received a single oral
application of sparfloxacin at both low (30 mg/kg)
and high dose (100 mg/kg) plus UVA irradiation developed a strong phototoxic reaction, with a progressive
(long-lasting) skin reaction. Following a dose of 100
mg/kg, the response remained elevated for over 72 h,
with some oedema observed in all hairless parts of
the body after 24–48 h. Following a dose of 30 mg/kg
of sparfloxacin, the response subsided after 24 h. Those
animals that received a dose of 100 mg/kg exhibited
swelling of the ears and a pronounced, but variable
reduction in weight gain during the days following irradiation. 24 h after application of 100 mg Bay y 3118/kg
bw. and UVA exposure animals had to be sacrificed in
moribund conditions due to severe phototoxic
reactions.
30
N.J. Neumann et al. / Journal of Photochemistry and Photobiology B: Biology 79 (2005) 25–34
Skin Reactions of Guinea Pigs After UV Irradiation
and Oral Application of Quinolones
Score (reddening)
(100 mg/kg)
2.5
2
1.5
BAY Y 3118
Sparfloxacin
Ciprofloxacin
1
0.5
0
Moxifloxacin
1
3
24
48
72
96
h after UV irradiation
Score (reddening)
Skin Reactions of Guinea Pigs After UV Irradiation
and Oral Application of Quinolones
(30 mg/kg)
2.5
2
1.5
1
0.5
0
8-MOP*
Sparfloxacin
Ciprofloxacin
Moxifloxacin
1
3
24
48
72
96
h after UV irradiation
*: Only 10mg 8-MOP per Kg bw have been applied in this experiment
Fig. 1. Acute phototoxic skin reaction after oral treatment of quinolones. 100 or 30 mg/kg of the quinolones indicated in the Figure were applied
orally to guinea pigs by gavage. 30 min after application the animals were UVA irradiated as described in Section 2. Skin reactions of the shaved back
are scored to the scheme mentioned under Section 2. The values shown in the Figure give the mean score of 5 animals each. Standard deviation were
extremely low and are therefore not given in the Figure.
In summary, the potency of different test compounds
to induce acute photoreactions in guinea pigs could be
ranked as follows (from no/weak ! strong/severe):
Moxifloxacin < ciprofloxacin < sparfloxacin < 8-MOP <
Bay y 3118.
3.4. UVA dependent IMDS test in mice
No photoreaction was observed in any of the animals
receiving moxifloxacin when applied dermally, whereas
erratic and mild responses were seen following the
administration of ciprofloxacin and sparfloxacin. When
compared with vehicle-treated animals, no difference in
the weight or cell counts of the draining lymph nodes
could be detected in these animals after dermal application (data not shown).
After oral application, however, the quinolones could
be ranked according to their ability to induce weak,
moderate or strong photoreactions as shown in Table
5 and in previous publications [7,10]. Table 5 shows
the result of one experiment as an example for such
ranking studies. For the quinolones concentrations applied were adjusted to comparable blood concentrations
Table 5
Ranking in accordance to the photoreactive potential of test substances (UV dependent IMDS ranking)
Index in ear thickness
LLN index (cell counts)
BAY y 3118
Olaquindox
2.90*
1.03
2.45*
3.18*
8-MOP
Sparfloxacin
Lomefloxacin
Ofloxacin
2.58*
2.98*
2.28*
1.48*
2.32*
2.19*
1.90*
1.37
Mice (5 animals per group) were treated orally with BAY y 3118 (50
mg/kg bodyweight), sparfloxacin (400 mg/kg bodyweight), lomefloxacin (400 mg/kg bodyweight), and ofloxacin (4000 mg/kg bodyweight).
Evaluation was done as described under Section 2.
As controls, mice were treated orally with the following photoreactive
compounds: 8-methoxypsoralen (8-MOP; 10 mg/kg bodyweight) or
olaquindox (50 mg/kg bodyweight).
In vehicle-treated groups no increase in ear thickness was detectable
(data not shown).
The increase in ear thickness is expressed in 0.01 mm units.
The standard deviation was in all cases less than 20% of the mean.
The calculation of the LLN index and the DI is described in the
methods section.
Significant differences (p 6 0.01) comparing vehicle- and related substance-treated groups are marked with *.
31
N.J. Neumann et al. / Journal of Photochemistry and Photobiology B: Biology 79 (2005) 25–34
dependent IMDS is based on several experiments like
that described above.
A summary of the ranking determined in all four tests
compared here is shown in Table 6.
The ability to discriminate the different endpoints of
photoreactions is demonstrated by investigating the effect of chlorpromazine after oral as well as epicutaneous
application. Chlorpromazine is known to cause different
photoreactions depending on the route of administration [35]. After oral administration of chlorpromazine
(100 mg/kg bodyweight) mice developed a strong UVdependent acute inflammatory reaction measured in
the ears by a mean increase in ear thickness of
6.80 · 0.01 mm. At the same time the ear-draining
lymph nodes showed a moderate proliferation response
with a cell count index of 1.51. The calculated DI (0.38)
clearly indicates a photo-irritating potential of orally
administered chlorpromazine (Fig. 3). All in all, as
while olaquindox was adjusted to the maximum tolerated dose in mice. The primary ranking is done in accordance to the cell count based stimulation index. This
ranking is then corrected to the amount of acute skin
reactions, i.e., photoirritating effects, measured by the
increase of ear thickness. Based on these parameters
Bay y 3118 for example was ranked higher than olaquindox for the overall photoreactive potential. A third
parameter, which may be taken into consideration, is
the dose of a test substance, which causes comparable
reaction pattern. An example is given in Fig. 2 where
nearly the same differentiation index is observed after
application of 25 mg Bay y 3118/kg or 400 mg sparfloxacin/kg. From that it could be concluded that the potential to induce photo irritancy is eight times higher with
Bay y 3118 than with sparfloxacin. However, such an
interpretation should be used with caution as discussed
below. The ranking indicated in Table 6 for the UV
Differentiation Index (DI)
photoirritating potential
0.01
0.1
400 mg Spa
10
100
0.19
0.27
200 mg Spa
100 mg Spa
25 mg Bay y 3118
photoallergic potential
1
0.52
0.17
32
50 mg Ola
Fig. 2. Differentiation indices (DI) calculated on the basis of LN cell counts and ear swelling were calculated as given in Section 2. DI after
application and UV-irradiation of the quinolones or olaquindox. Concentrations as indicated in the figure. The so-called Differentiation Indices: grey
bars indicate photo-irritating properties while black bars indicate photoallergic properties of the test substances at the concentration tested.
Table 6
Ranking of the photoreactions in dependency of the assay used
PHET
3T3
GP Acute test
UVA dep. IMDS
8-MOP
BAY y 3118
Protoporphyrine IX
Acridine
8-MOP
BAY y 3118
BAY y 3118
8-MOP
BAY y 3118
Olaquindox
8-MOP
Promethazine
Sparfloxacin
Acridine
TCSA
Hematoporphyrine
Lomefloxacin
Protoporphyrine IX
Sparfloxacin
Promethazine
Lomefloxacin
TCSA
Ciprofloxacin
Tetracyclin
Moxifloxacin
Olaquindox
Olaquindox
Ciprofloxacin
Tetracyclin
Moxifloxacin
Sparfloxacin
Lomefloxacin
Ciprofloxacin
Tetracyclin
Moxifloxacin
Chlorpromazine
TCSA
Promethazine
Sparfloxacin
Acridine
Lomefloxacin
Ofloxacin
Ciprofloxacin
Moxifloxacin
The compounds can be categorized into three classes indicated by the following parts of Table 5: Upper part: strong phototoxicity. Mid part:
moderate phototoxicity. Lower part: no or very weak phototoxicity.
32
N.J. Neumann et al. / Journal of Photochemistry and Photobiology B: Biology 79 (2005) 25–34
Differentiation Indices of Chlorpromazine
compared to 8-MOP
oral + UV-A
10 mg/kg 8-MOP
0.3
50 mg/kg
1.8
100 mg/kg
0.4
200 mg/kg
0.3
dermal + UV-A
0.3% 8-MOP
0.1
3%
27.3
10%
32
0.01
0.1
1
10
100
Fig. 3. LN weight and cell count indices after application and UV-irradiation of the reference substances 8-MOP or chlorpromazine. Concentrations
as indicated in the Figure. The so-called Differentiation Indices (DI) were calculated as given in Section 2. Grey bars indicate photo-irritating
properties while black bars indicate photoallergic properties of the test substances at the concentration tested.
expected from the literature UVA irradiation of chlorpromazine treated mice resulted in a DI > 1 after dermal
(or low dose oral) and a DI < 1 after oral application
(Fig. 3). This confirmed the potential of chlorpromazine
to induce photo irritating reactions after oral application of higher doses or photoallergic reactions after epicutaneous exposure as expected from the human data.
As control we used 8-MOP which in all cases induced
photo irritating reactions.
4. Discussion
In this study two in vitro assays were compared to
two in vivo models for their ability to rank the phototoxic potential of some selected chemicals.
The in vitro mouse fibroblast phototoxicity assay
(NRU) revealed a distinct differentiation between strong
(8-MOP, Bay y 3118) and moderate photoirritants, but
at low dosages of UVA light (5 J/cm2) the spectrum
among moderate or non-photo irritants (ciprofloxacin,
moxifloxacin) was narrowed.
As has been shown before, this photoirritancy due to
moderate phototoxic agents was more distinctly elucidated on increase of irradiation dosage [7]. In particular,
ciprofloxacin became more photoirritant than moxifloxacin. The latter molecule remained non-phototoxic
[7,10].
On the other hand, olaquindox well-known as a photoallergic substance did else elicit a faint phototoxic effect at low dose UVA.
This methodology provides a useful and rapid
method to screen phototoxic potential of new drugs or
suspected photo irritants. The ranking of potency values
of weak phototoxic agents might be an issue.
The photo henÕs egg test (PHET) showed a clear
ranking of strong, moderate as well as weak and nonphototoxic chemicals (Table 4).
8-MOP (a severe photo irritant without known photoallergic potential) and Bay y 3118 were the leading
phototoxic agents followed by promethazine and sparfloxacin. On the other hand, ciprofloxacin revealed only
a very weak and olaquindox (a photoallergen without
phototoxic potential) showed no detectable phototoxic
effects in the PHET. These results are in agreement with
clinical findings in human skin. Thus, the PHET also
represents an inexpensive and valid screening method
for photo irritants, especially in order to classify the different phototoxic potentials of moderate or weak
photoirritants.
No photoirritating or photoallergic reactions to
moxifloxacin could be detected after oral or dermal
application plus UVA irradiation in either guinea pigs
or mice. In guinea pigs, moxifloxacin at a dose of 30
and 100 mg/kg did not induce any UV-dependent increase in skin reddening when compared with mock or
vehicle-treated animals. In contrast, animals treated
with other test compounds like ciprofloxacin, sparfloxacin, lomefloxacin, promethazine, 8-MOP or Bay y 1318
plus irradiation developed a weak to strong, sometimes
long-lasting, sunburn-like skin reaction, even at a doses
of 30 mg/kg.
N.J. Neumann et al. / Journal of Photochemistry and Photobiology B: Biology 79 (2005) 25–34
The mouse model used in this study was a modified
local lymph node assay (LLNA) designed to discriminate photo irritant from photoallergic reactions (socalled UV-IMDS). Photoirritancy was observed more
or less for all test compounds except olaquindox. However, photoallergic responses were only observed in animals treated with enoxacin, chlorpromazine, and
olaquindox. In contrast to the other test methods, this
model was able to detect both photo irritant and photoallergic reactions [32–34,36]. With respect to photoirritancy, these in vivo findings were corroborated by the
results of the in vitro phototoxicity assays, i.e., 3T3
mouse fibroblasts (NRU) as well as the PHET. Additionally, the acute guinea pig skin test revealed comparable results.
The UVA dependent IMDS test was also able to
differentiate the photoreactions induced by chlorpromazine according to the route of administration. These
findings have been confirmed by observations made
in humans after treatment with chlorpromazine. Indeed, we found differences in the reaction pattern after
UVA irradiation determined by the route of administration. Thus, the results obtained after different
administrations of chlorpromazine clearly demonstrated the validity of this test procedure. Moreover,
the calculated differentiation index (DI) as shown in
Figs. 2 and 3 was linked to the tested concentrations
of the compounds. Thus, the value of the DI per se
was not a measure of the ability to induce photoirritating or photoallergic reactions. As can be seen from Fig.
3 the calculated DI increased with the amount of compound administered. There was a dose dependent relationship between DI and the applied dose in a narrow
range. This holds especially true if a test compound
that induced predominantly one of the endpoints, measured i.e. photo irritation or photoallergy. Compounds,
able to induce the two endpoints, could not show a
clear dose dependent relation ship of DI as this parameter was calculated by the ratio of acute skin reaction
over proliferation of the lymph node cells.
The fluoroquinolones represent a major class of antibacterials with great therapeutic potential. Photosensitivity reactions have been recognised as an unwanted
adverse effect of some of these agents [14–19,37]), but
the incidence and intensity of these reactions were found
to vary considerably from one compound to another.
Where photoreactions are relatively low in incidence,
mild, reversible and clinically manageable, the benefits
of the antimicrobial drug may well outweigh its potential for adverse photosensitivity effects [37].
5. Conclusion
In summary, the ranking of the observed photo irritant (phototoxic) potentials in the two in vitro assays
33
were similar to the human data. But it has to be
stressed that this concordance was exclusively based
on the photo irritating properties of the tested agents.
However, in vitro assays were not able to detect compounds that had photoallergic properties. Moreover,
the ranking of skin reactions in guinea pigs was linked
to the photo irritating potential. In contrast, the modified lymph node assay in mice, i.e. the UVA dep.
IMDS test was able to discriminate between both endpoints and thus the results of this assay were comparable to photoreaction patterns described in humans.
In our study, the two in vitro assays appeared to be
useful tools to detect photo irritant reactions. Therefore, it seems to be reasonable to recommend these
in vitro models as screening tools in the relevant
guidelines [38]. However, these assays are limited to
the evaluation of phototoxic potentials. As long as
no in vitro model for photoallergy is available, the
UV-IMDS should be considered to evaluate photoallergic properties of a supposed photoreactive agent
[36,21,22].
6. Abbreviations
8-MOP
DI
DMEM
HET
HR
IMDS
8-methoxypsoralen
differentiation index
DulbeccoÕs Modified Eagle Medium
henÕs egg test
hemorrhage
integrated model for the differentiation of
skin reactions
LLN
local lymph node
LLNA
local lymph node assay
LN
lymph nodes
MD
membrane discoloration
NRU
neutral red uptake
NSAIDS non-steroidal anti-anflammatory drugs
OECD
Organisation for Economic Co-operation
and Development
PBS
phosphate buffer solution
PHET
photo henÕs egg test
PIF
Photo-Irritation-Factor
PSS
physiological salt solution
YS
yolk sac
References
[1] J.E. Allen, Drug-induced photosensitivity, Clin. Pharm. 12 (1993)
580–587.
[2] J.H. Epstein, B.U. Wintroub, Photosensitivity due to drugs,
Drugs 30 (1985) 42–57.
[3] J.W. Gould, M.G. Mercurino, C.A. Elemets, Cutaneous photosensitivity diseases induced by exogenous agents, J. Am. Acad.
Dermatol. 33 (1995) 551–573.
34
N.J. Neumann et al. / Journal of Photochemistry and Photobiology B: Biology 79 (2005) 25–34
[4] M. Takigawa, Y. Miyachi, Mechanisms of contact photosensitivity in mice. I. T-cell regulation of contact photosensitivity to
tetrachlorosalicylanilide under the genetic restriction of the major
histocompatibility complex, J. Invest. Dermatol. 79 (1982) 108–
115.
[5] Y. Miyachi, M. Takigawa, Mechanisms of contact photosensitivity in mice. II. Langerhans cells are required for successful
induction of photosensitivity to TCSA, J. Invest. Dermatol. 78
(1982) 363–365.
[6] K. Marutani, M. Matsumoto, Y. Otabe, M. Nagamuta, K.
Tanaka, A. Miyoshi, T. Hasegana, H. Nagano, S. Matsubara, R.
Kamide, Reduced phototoxicity of a fluoroquinolone antibacterial agent with a methoxy group at the 8 position in mice
irradiated with long-wavelength UV light, Antimicrob. Agents
Chemother. 37 (10) (1993) 2217–2223.
[7] A. Blotz, J. Blümel, H.-J. Ahr, H.-W. Vohr, Analyses of dermal
photoreactivity after oral application of fluoroquinolones with the
IMDS, Immunobiology 199 (1998) 3–5.
[8] U. Akter, M. Niwa, T. Nose, T. Kaida, H. Natsuno, O. Kozawa,
T. Uematsu, Effects of several agents on UVB- and UVA plus
systemic fluoroquinolone-induced erythema of guinea pig skin
evaluated by reflectance colorimetry, Free Radic. Biol. Med. 24 (78) (1998) 1113–1119.
[9] N.J. Neumann, B. Hormey, H.-W. Vohr, P. Lehmann, Photoreactions, in: P. Elsner, H.F. Merk, H.I. Maibach (Eds.), Cosmetics.
Controlled Efficacy Studies and Regulations, Springer Verlag,
Berlin, 1999.
[10] H.-W. Vohr, G. Wasinska-Kempka, H.J. Ahr, An investigation
into the phototoxic potential of moxifloxacin, in: Moxifloxacin in
Practice, 1999.
[11] N.J. Traynor, M.D. Barrat, W.W. Lovell, G. Fergusson, N.K.
Gibbs, Comparison of an in vitro cellular phototoxicity model
against controlled clinical trials of fluoroquinolone skin phototoxicity, Toxicol. In Vitro 14 (2000) 275–283.
[12] A. Blotz, J. Michel, A. Moysan, J. Blümel, L. Dubertret, H.J.
Ahr, H.-W. Vohr, Analyses of cutaneous fluoroquinolones photoreactivity using the integrated model for the differentiation of
skin reactions, J. Photochem. Photobiol. 58 (2000) 46–53.
[13] P.A. Jones, A.V. King, L.K. Earl, R.S. Lawrence, An assessment
of the phototoxic hazard of personal product ingredient using in
vitro assays, Toxicol. In Vitro 17 (2003) 471–480.
[14] K. Rosen, G. Swanbeck, Phototoxic reactionsfrom some common
drugs provoked by a high-intensity UV-A lamp, Acta Dermatovenerol. 62 (1982) 246–248.
[15] W. Christ, B. Esch, Use-risk consideration of anti-infective agents
from the point of view of the licensing authority, Infection 19
(1991) S65–S71.
[16] J.M. Domagala, Structure-activity and structure-side-effect relationships for the quinolone antibacterials. J. Antimicrob. Chemother. 33, 685–706 (published erratum appears in J. Antimicrob.
Chemother. 34 (1994) 851).
[17] M. Vousden, J. Fergusson, J. Richards, N. Bird, A. Allen,
Evaluation of phototoxic potential of gemifloxacin in healthy
volunteers compared with ciprofloxacin, Chemotherapy 45 (1999)
512–520.
[18] I. Man, J. Murphy, J. Ferguson, Fluoroquinolone phototoxicity:
a comparison of moxifloxacin and lomefloxacin in normal
volunteers, J. Antimicrob. Chemother. 43 (Suppl. B) (1999) 77–82.
[19] I. Man, N.J. Traynor, J. Fergusson, Recent developments in
fluoroquinolone phototoxicity, Photodermatol. Photoimmunol.
Photomed. 15 (1) (1999) 32–33.
[20] H. Spielmann, M. Balls, B. Döring, H.G. Holzhütter, S. Kalweit,
G. Klecak, H. LÔEplattenier, M. Liebsch, W.W. Lovell, EEC/
COLIPA project on in vitro phototoxicity testing: first results
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
obtained with a Balb/c 3T3 cell phototoxicity assay, Toxicol. In
Vitro 8 (1994) 793–796.
Commission Directive 2000/33/EC of 25 April 2000 adapting to
technical progress for the 27th time Council Directive 67/548 E on
the classification, packaging and labelling of dangerous substances., 2000. Annex V method B41 Phototoxicity – In vitro 3T3
NRU phototoxicity test. Official Journal of the European Communities L136, 90.
CPMP, Committee for Proprietary medical Products, Note for
Guidance on Photosafety Testing, CPMP/SWP/398/01, 2002.
Th Maurer, Experimental contact photoallergenicity: guinea pig
models, Photodermatology 1 (1984) 221–231.
H.J. Draize, Intracutaneous sensitization test on guinea pigs, in:
Apraisal of the Safety of Chemicals in Food, Drugs and
Cosmetics, Association of Food and Drug Officials of the United
States, Austin, Texas, 1959.
P.A. Duffy, Irritancy testing-a cultured approach, Toxicol In
Vitro 3 (1989) 157–158.
M. Rosenbruch, Toxizitätsuntersuchungen am bebrüteten Hühnerei, Derm Beruf Umwelt, 38 (1990) 5–11.
R.G. Freeman, W. Murtishaw, J.M. Knox, Tissue culture
techniques in the study of cell photobiology and phototoxicity,
Invest. Dermatol. 54 (1970) 164–169.
K. Maier, R. Schmitt-Landgraf, B. Siegmund, Development of an
in vitro test system with human skin cells for evaluation of
phototoxicity, Toxicol. In Vitro 5/6 (1991) 457–461.
N.J. Neumann, E. Hölzle, P. Lehmann, M. Rosenbruch, P.
Klaucic, G. Plewig, Photo henÕs egg test: a model for phototoxicity, Br. J. Dermatol. 136 (1997) 326–330.
N.J. Neumann, A. Klaucic, E. Hölzle, P. Lehmann, Evaluation of
the phototoxic potential of ciprofloxacin in the photo henÕs egg
test, Arch. Dermatol. Res. 287 (1995) 384.
N.J. Neumann, E. Hölzle, M. Wallerand, S. Vierbaum, T.
Ruzicka, P. Lehmann, The photoprotective effect of ascorbic
acid, acetylsalicylic acid, and indomethacin evaluated by the
photo henÕs egg test, Photodermatol. Photoimmunol. Photomed.
15 (1999) 166–170.
H. Vohr, B. Homey, C. Schuppe, P. Kind, Photoreactions
detected in a modified local lymph node assay in the mouse,
Photodermatol. Photoimmunol. Photomed 10 (1994) 57–64.
B. Homey, C. Schilling, J. Blümel, H.C. Schuppe, T. Ruzicka,
H.J. Ahr, P. Lehmann, H-W. Vohr, Development of an
integrated model for the differentiation of chemical-induced
allergic and irritant skin reactions, Toxicol. Appl. Pharm. 153
(1998) 83–94.
H.-W. Vohr, J. Blümel, A. Blotz, B. Homey, H.J. Ahr, An intralaboratory validation of IMDS: discrimination between
(photo)allergic and (photo) irritant skin reactions in mice, Arch.
Toxicol. 73 (2000) 501–509.
J-H. Epstein, L.A. Brunsting, M.C. Petersen, B.E. Schwarz, A
study of photosensitivity occurring with chlorpromazine therapy,
J. Invest. Dermatol. 28 (1957) 329–338.
H. Spielmann, L. Müller, D. Averbeck, M. Balls, S. BrendlerSchwab, J.V. Castell, R. Curren, O. de Silva, N.K. Gibbs, M.
Liebsch, W.W. Lovell, H.F. Merk, J.F. Nash, N.J. Neumann,
W.J. Pape, P. Ulrich, H.-W- Vohr, The second ECVAM Workshop on Phototoxicity Testing. The Report and Recommendations of ECVAM Workshop 42. Alternatives Lab. Anim. 28(6)
(2000) 777–814.
Vassileeva, Antimicrobial photosensitive reactions, Arch. Intern.
Med. 158 (1998) 1993–2000.
OECD (2002) OECD Guideline for the Testing of Chemicals; In
Vitro 3T3 NRU Phototoxicity Test, Guideline Draft 432, Paris,
22th March, 2002.
Martin Rosenbruch