This study evaluated the in vitro skin phototoxicity of cosmetic formulations containing photounstable and photostable UV filters (octyl methoxycinnamate, benzophenone-3, avobenzone, octyl salicylate, 4-methylbenzilidene camphor) and vitamin A palmitate using two tests: the 3T3 Neutral Red Uptake Phototoxicity Test and the Human 3-D Skin Model In Vitro Phototoxicity Test. Avobenzone showed pronounced phototoxicity and vitamin A showed a tendency toward weak phototoxic potential. A synergistic effect of vitamin A palmitate on the phototoxicity of combinations containing avobenzone was observed

1. Skin phototoxicity of cosmetic formulations containing photounstable
and photostable UV-filters and vitamin A palmitate
Lorena R. Gaspar a,⇑
, Julian Tharmann b
, Patricia M.B.G. Maia Campos a
, Manfred Liebsch b
a
University of Sao Paulo, Faculty of Pharmaceutical Sciences of Ribeirao Preto, SP, Brazil
b
Federal Institute for Risk Assessment (BfR), ZEBET at the BfR, Berlin, Germany
a r t i c l e i n f o
Article history:
Received 10 January 2012
Accepted 2 August 2012
Available online 10 August 2012
Keywords:
Skin phototoxicity
UV-filters
Vitamin A
3T3 NRU
H3D-PT
a b s t r a c t
The aim of this study was to evaluate the in vitro skin phototoxicity of cosmetic formulations containing
photounstable and photostable UV-filters and vitamin A palmitate, assessed by two in vitro techniques:
3T3 Neutral Red Uptake Phototoxicity Test and Human 3-D Skin Model In Vitro Phototoxicity Test. For
this, four different formulations containing vitamin A palmitate and different UV-filters combinations,
two of them considered photostable and two of them considered photounstable, were prepared. Solu-
tions of each UV-filter and vitamin under study and solutions of four different combinations under study
were also prepared. The phototoxicity was assessed in vitro by the 3T3 NRU phototoxicity test (3T3-NRU-
PT) and subsequently in a phototoxicity test on reconstructed human skin model (H3D-PT). Avobenzone
presented a pronounced phototoxicity and vitamin A presented a tendency to a weak phototoxic poten-
tial. A synergistic effect of vitamin A palmitate on the phototoxicity of combinations containing avobenz-
one was observed. H3D-PT results did not confirm the positive 3T3-NRU-PT results. However, despite the
four formulations studied did not present any acute phototoxicity potential, the combination 2 contain-
ing octyl methoxycinnamate (OMC), avobenzone (AVB) and 4-methylbenzilidene camphor (MBC) pre-
sented an indication of phototoxicity that should be better investigated in terms of the frequency of
photoallergic or chronic phototoxicity in humans, once these tests are scientifically validated only to
detect phototoxic potential with the aim of preventing phototoxic reactions in the general population,
and positive results cannot predict the exact incidence of phototoxic reactions in humans.
1. Introduction
The level and quality of UV protection provided by sunscreen
products have improved considerably over the past three decades.
Modern sunscreen products should provide broad-spectrum UV
protection, offering uniform UVB/UVA protection, because this as-
sures that the natural spectrum of sunlight is attenuated without
altering its quality. Modern sunscreens may contain at least two
UV filters, one with optimal performance in the UVA region and
the other one in the UVB region. However, the presence of different
UV filters, which usually leads to synergistic effects regarding both
the final performance and photostabilization of the sunscreen, can
also accelerate their decomposition if a photoreaction occurs be-
tween the single components (Osterwalder and Herzog, 2010;
Gonzalez et al., 2007; Chatelain and Gabard, 2001; Lhiaubet-Vallet
et al., 2010).
Despite the wide range of UVB filters, appropriate UVA filters
are rare; among them avobenzone is probably the most important
representative. This active ingredient is present in numerous com-
mercial sunscreen and cosmetic formulations. Avobenzone
strongly absorbs UVA, but presents significant degradation under
UV exposure reducing its UVA protecting effect (Paris et al.,
2009; Bouillon, 2000).
The reactive intermediates of photounstable filter substances
come into direct contact with the skin, where they may behave
as photo-oxidants or may also promote phototoxic or photoallergic
contact dermatitis. The interaction of photodegradation products
with sunscreen excipients or skin components like sebum may
lead to the formation of newmolecules with unknown toxicological
properties (Cambon et al., 2001; Deleo et al., 1992; Rieger, 1997;
Schrader et al., 1994; Nohynek and Schaefer, 2001). Consequently,
there is an increasing concern about the phototoxicity and photo-
allergy of UV filters.
Phototoxicity is defined as a toxic response from a substance
applied to the body which is either elicited or increased (apparent
at lower dose levels) after subsequent exposure to light, or that is
induced by skin irradiation after systemic administration of a sub-
stance (OECD, 2004). It is as a non-immunological light-induced
0887-2333
http://dx.doi.org/10.1016/j.tiv.2012.08.006
⇑ Corresponding author. Address: Faculdade de Ciências Farmacêuticas de
Ribeirão Preto, Universidade de São Paulo, Av. do Café s/n, 14040-903 Ribeirão
Preto, SP, Brazil. Tel.: +55 16 3602 4315.
E-mail address: lorena@fcfrp.usp.br (L.R. Gaspar).
Toxicology in Vitro 27 (2013) 418–425
Contents lists available at SciVerse ScienceDirect
Toxicology in Vitro
journal homepage: www.elsevier.com/locate/toxinvit
Ó 2012 Elsevier Ltd. Open access under the Elsevier OA license.
Ó 2012 Elsevier Ltd. Open access under the Elsevier OA license.

2. skin response (dermatitis) to a photoactive chemical, and the skin
response is characterized by erythema and sometimes edema,
vesiculation, and pigmentation. Phototoxic reactions are compara-
ble with primary irritation reactions in that they may be elicited
after a single exposure, thus no induction period is required (Marz-
ulli and Maibach, 1985).
Photoallergic contact dermatitis is thought to arise when UV
radiation interacts with a chemical to form a hapten or antigen,
which in turn triggers a type IV hypersensitivity reaction (Bryden
et al., 2006).
As organic UV filters are used in increasing amounts, there is
gradual emergence of reports of allergic and photoallergic reac-
tions to UV filters on human skin. Epidemiological studies per-
formed using human photopatch test, showed that avobenzone
and many other UV-filters were the causal agents of these allergic
and photoallergic reactions (Schauder and Ippen, 1997; Lodén
et al., 2011). Among the organic UV absorbers, octocrylene, benzo-
phenone-3 and avobenzone most frequently elicited photoallergic
contact dermatitis. On the other hand, despite cinnamates and
salicylates are used in large quantities, reports of allergic reactions
are relatively low (Kerr et al., 2012; Kerr and Ferguson, 2010).
Another tendency in photoprotection in the topical application
and systemic administration of antioxidants acting as photoprotec-
tives, which could maintain or restore a healthy skin barrier (Pinnell,
2003). Among the frequently used antioxidants in anti-aging prod-
ucts we can point out vitamin A, C and E derivatives. Vitamin A
palmitate acts on epithelization in dry and rough skin, as well as
on keratinization considered being abnormal (Maia Campos et al.,
1999). In addition, it also absorbs UV radiation between 300 and
350 nm, with a maximum at 325 nm (Antille et al., 2003), which
can suggest that it may have a biologically relevant filter activity
as well. However some studies have shown that vitamin A and its es-
ter undergo photo-oxidation to give a variety of photodecomposi-
tion products and reactive oxygen species (Xia et al., 2006).
Therefore, since some studies show that vitamin A generates toxic
photoproducts or allergens when exposed to UV radiation, the US
FDA selected vitamin A palmitate by the National Toxicology Pro-
gram (NTP) as a high priority compound for phototoxicity and
photocarcinogenicity studies (Xia et al., 2006; Tolleson et al., 2005).
In Europe, since the year 2000, in vivo testing in animals for
acute phototoxic potential is no longer permitted, since a success-
fully validated in vitro alternative method has been accepted for
regulatory purposes. Due to its high sensitivity and specificity,
the validated 3T3 Neutral Red Uptake Phototoxicity Test (3T3-
NRU-PT) is the core test, which is usually the only phototoxicity
test required when the substance is not considered phototoxic
(Liebsch et al., 2005). Reconstructed human skin models closely
resemble the native human epidermis due to the presence of a bar-
rier function similar to the barrier function of human epidermis.
Thus, the reconstructed human skin models are proposed as an
additional tool for verification of positive results of the 3T3 NRU
PT, with respect to bioavailability in human skin, and/or for testing
of substances incompatible with the 3T3 NRU PT (Liebsch et al.,
2005; Kejlová et al., 2007). Human in vivo photopatch method
can also be performed, but they must be carried out only after prior
risk assessment in vitro studies and in compliance with the ethical
principles avoiding unnecessary risks to human subjects. Some
studies report a good correlation among 3T3-NRU-PT, Human
3-D Skin Model and human in vivo photopatch tests (Kejlová
et al., 2007; Spielmann et al., 1998). However, despite the proposed
tests are scientifically validated to detect phototoxic potential with
the aim of preventing phototoxic reactions in the general popula-
tion, the extrapolation of in vitro results to the human situation
may be performed only to a limited extent. These limitations are
in part due to the higher permeability of the skin tissues compared
to human skin in vivo (Kand’árová, 2006),
There are also some other concerns involving the predictability
of phototoxicity testing in animals and humans (Maibach and
Marzulli, 2004). For example, Marzulli and Maibach (1970) dis-
cussed the correlation between skin permeability and bergapten
phototoxicity performed in animals and humans. They found that
animals with more permeable skin (rabbits and hairless mice)
were more reactive to bergapten than monkey and swine that have
less permeable skin. In addition, they found that stripped skin had
more pronounced biological effects than intact skin or less perme-
able forearm skin.
Nevertheless, even human photopatch tests need to be stan-
dardized in order to investigate photoallergic reactions and obtain
consistent results. Such points are related to experimental design,
irradiation sources, specify exposure time and distance of source to
the skin, as well as UV dose (Maibach and Marzulli, 2004). In 2004
a group of interested European Contact Dermatologists/Photobiol-
ogists met to produce a consensus statement on methodology, test
materials and interpretation of photopatch testing (Bruynzeel
et al., 2004). In 2012, this group provided current information on
the relative frequency of photo-allergic contact dermatitis to com-
mon photoallergic organic UV-filters and they also stated the rele-
vance of such investigations as well as of some cross-reactions
between some UV-filters combinations (EMCPPTS, 2012). This
way, it is of great importance to investigate the phototoxic poten-
tial of new combinations of UV-filters and antioxidant substances
like vitamin A. However, for ethical reasons before in vivo testing
on human volunteers and to avoid confirmatory testing in animals,
3T3 NRU-PT and H3D-PT are offering an attractive in vitro alterna-
tive approach, since H3D-PT is characterized by skin barrier
function.
Therefore, the aim of this study was to evaluate the in vitro skin
phototoxicity of cosmetic formulations containing photounstable
and photostable UV-filters and vitamin A palmitate, assessed by
two in vitro techniques: 3T3 Neutral Red Uptake Phototoxicity Test
and Human 3-D Skin Model In Vitro Phototoxicity Test.
2. Materials and methods
2.1. Chemicals
UV-filters samples were supplied by Symrise (Germany): ben-
zophenone-3, butyl methoxydibenzoylmethane (avobenzone), eth-
ylhexyl methoxycinnamate, Octocrylene, methylbenzylidene
camphor, ethylhexyl salicylate. Vitamin A palmitate (retinylpalmi-
tate) was supplied by DSM (Switzerland). Positive controls Chlor-
promazinehydrochloride and Bergamot oil were purchased from
SIGMA AG (Germany).
2.2. Formulations
Four UV filter combinations often used in SPF 15 sunscreen
products were chosen for this study. The combined UV filters were
added to a formulation containing 0.5% of hydroxyethyl cellulose,
3% of glycerin, 0.05% of BHT, 3.5% of phosphate-based self-
emulsifying wax (cetearyl alcohol, dicetyl phosphate, ceteth-10
phosphate), 6% of C12–C15 alkyl benzoate, 3% of propyleneglycol,
4% of a blend of polyglyceryl-10 myristate, diphenylmethicone,
trietilexanoine, 2% of cyclopentasiloxane, 0.8% of phenoxyethanol
and parabens and distilled water. The combinations were: 7% of oc-
tyl methoxycinnamate (OMC), 2% of benzophenone-3 (BP-3) and
1.5% of octyl salicylate (OS) (formulation 1); 10% of OMC, 2% of avo-
benzone (AVB) and 2% of 4-methylbenzilidene camphor (MBC)
(formulation 2); 7% of OMC, 4% of BP-3 and 5% of octocrylene
(OC) (formulation 3); 5% of OMC, 2% of AVB and 7% of OC (formu-
lation 4) (Gaspar and Maia Campos, 2006).
L.R. Gaspar et al. / Toxicology in Vitro 27 (2013) 418–425 419

3. 2.3. Combinations under study
For the 3T3 Neutral Red Uptake Phototoxicity Test, a stock solu-
tion was prepared in DMSO for each UV-filter and the vitamin un-
der study. This stock solution was diluted in eight different
concentrations in EBSS ranging from 0.1 to 316 lg/mL in a geomet-
ric progression (constant factor of 3.16).
Four different combinations under study were also analyzed,
these combinations contained the UV-filters under study in the
same proportion (1:1:1) (Comb 1, Comb 2, Comb 3, Comb 4) or
the same proportion used in the formulations under study (Comb
1=, Comb 2=, Comb 3=, Comb 4=). The different combinations of
UV-filters in the presence of vitamin A, in different proportions
were also analyzed. The stock solutions of the combinations in
DMSO were diluted in 8 different concentrations in EBSS ranging
from 3.16 to 178 lg/mL in a geometric progression (constant factor
of 1.78).
For the EpiDerm Skin Phototoxicity test, all combinations were
diluted in C12–C15 alkyl benzoate.
2.4. Source of irradiation
The UV light source used in phototoxicity tests in cell culture
(3T3 NRU) and in human 3-D skin model (H3D-PT) was a doped
mercury metal halide lamp (SOL 500, Dr. Hönle, Germany) which
simulates the spectral distribution of natural sunlight. Aspectrum
almost devoid of UVB (<320 nm) was achievedby filtering with
50% transmission at a wavelength of335 nm (Filter H1, Dr. Hönle,
Germany). The emittedenergy was measured before each experi-
ment with a calibrated UVA meter (Type No. 37, Dr. Hönle,
Germany)(OECD, 2004; Kejlová et al., 2007).
2.5. Phototoxicity test in cell culture (3T3 NRU)
The 3T3 Neutral Red Uptake Phototoxicity Test was performed
according to INVITTOX Protocol No. 78 (Liebsch and Spielmann,
1998), using 3T3 Balb/c fibroblasts (L1, ECACC No. 86052701).
For this purpose, after the evaluation of the fibroblasts sensibility
to the UVA radiation, two 96-well plates were used for each
substance or combination, one to determine the cytotoxicity
(absence of radiation) and another for the phototoxicity (presence
of radiation). For that, firstly 100 lL of a cell suspension of 3T3
fibroblasts in Dulbecco’s Modification of Eagle’s Medium (DMEM)
containing New Born Calf Serum and antibiotics (1 Â 105
cells/mL,
1 Â 104
cells/well) was dispensed in two 96-well plates. After a
24 h period of incubation (7.5% CO2, 37 °C), plates were washed
with 150 lL of Earle’s Balanced Salt Solution (EBSS) and different
concentrations of the test chemicals or combination were applied
in sextuplicate in the 96-well plates. After 1 h incubation, the
+UVA plate was irradiated for 50 min with 1.7 mW/cm2
(=5 J/cm2
) of UVA radiation from UV-sun simulator, type SOL-500
(Dr. Hönle, Germany). The ÀUVA plate was kept in a dark box for
50 min. The test solutions were replaced by culture medium and
plates were incubated overnight. Neutral Red medium was added
in each well and after an incubation period, cells were washed with
EBSS and a desorb (ethanol/acetic acid) solution was added. Then,
neutral red extracted from viable cells formed a homogeneous
solution and the +UVA and ÀUVA plates were analyzed in a micro-
liter plate reader at 540 nm.
For concentration–response analysis Phototox Version 2.0 soft-
ware (obtained from ZEBET, Germany) was employed. A test sub-
stance is predicted as having a potential phototoxic hazard if
the photoirritation factor (PIF), calculated as the ratio of toxicity
for each substance with and without UV light, is higher than 5
(Spielmann et al., 1998). Using the Phototox software, a second
predictor of phototoxicity, the mean photoeffect (MPE) was also
calculated. The MPE is a statistical comparison of the dose–re-
sponse curves obtained withand without UV and a test substance
is predicted as phototoxic if MPE is higher than 0.1 (Holzhütter,
1997). According to the Organisation for Economic Cooperation
and Development (OECD) Test Guideline 432, a test substance with
a PIF >2 and <5 or an MPE >0.1 and <0.15 is predicted as ‘‘probably
phototoxic’’ (OECD, 2004; Kejlová et al., 2007). Results are the
mean of at least two independent experiments ± SEM.
Chlorpromazine was used as positive control for phototoxicity
test in cell culture. According to the validation procedures, the test
meets acceptance criteria, if for chlorpromazine EC50 (+UVA), i.e.
the concentration inhibiting cell viability by 50% of untreated con-
trols, is within the range of 0.1–2.0 lg/mL, and the chlorpromazine
EC50 (ÀUVA) is within the range of 7.0–90.0 lg/mL (OECD, 2004).
2.6. Phototoxicity test in human 3-D skin model (H3D-PT)
The EpiDerm Skin Phototoxicity Test was conducted according
to Liebsch et al. (1999) and Kejlová et al. (2007). 3D skin models,
Epi-Derm EPI-200 (0.63 cm2
), were supplied by MatTek, USA. Be-
fore dosing, the tissues were preincubated in fresh medium for
1 h to release transport stress related compounds and debris. After
that, the medium was replaced by fresh medium and the tissue
was incubated over night (18–24 h) (37 °C, 5% CO2). The test
formulations and substances were applied overnight (16–20 h) in
a volume of 15 lL of each formulation per tissue or 25 lL of each
combination diluted in C12–C15 alkyl benzoate per tissue. One
set of tissues was irradiated with a nontoxicdose of 6 J/cm2
(as
measured in the UVA range). One day after the treatment and
UVA exposure the cytotoxicitywas detected as reduction of mito-
chondrial conversion of MTT to formazan. The optical density of
the formazan extract was determined at 540 nm by means of Spec-
trophotometer Infinite 200 (Tecan Trading AG, Switzerland). The
results of meantissue values in the presence and absence of UV
light werecompared and a test substance was considered to be
phototoxic, if one or more test concentrations of the (+UVA) part
of the experiment revealed a decrease in viability exceeding 30%
when compared with identical concentrationsof the (ÀUVA) part
of the experiment (Liebsch et al., 1997). Bergamot oil was used
as positive control (Kejlová et al., 2007).
3. Results
3.1. Phototoxicity test in cell culture (3T3 NRU)
The results obtained in the 3T3 Neutral Red Uptake Phototoxic-
ity test showed that only avobenzone was considered phototoxic,
since it presented mean MPE of 0.327 and mean PIF of 11.478
(Table 1). Despite vitamin A palmitate presented a borderline mean
MPE (0.106), some obtained values were classified as phototoxic or
probably phototoxic, thuson the basis of these borderline results,
this vitamin was submitted to a UV dose/response study to confirm
its phototoxic potential.
The results obtained when avobenzone and vitamin A palmitate
were submitted to 3T3 NRU Phototoxicity test under various inten-
sities of UVA (2, 4 and 8 J/cm2
) showed that avobenzone presented
a pronounced phototoxicity enhancement (increased MPE) with
higher UVA doses, showing that its phototoxicity was UVA dose
dependent (Table 2). However when vitamin A was analyzed, no
dose response effect was observed. Thus, the obtained results
showed that vitamin A presented a tendency to a weak phototox-
icpotential that was not confirmed in the dose response study
(Table 2).
When the combinations under study were analyzed, the photo-
toxicity test showed that only the combinations containing
420 L.R. Gaspar et al. / Toxicology in Vitro 27 (2013) 418–425

4. avobenzone, comb 2 (OMC, AVB, MBC) and comb 4 (OMC, AVB, OC),
presented phototoxic potential (Table 4). The other combinations,
comb 1 (OMC, BP-3 and OS) and comb 3 (OMC, BP-3 and OC), did
not present any phototoxic potential, even when combined with
vitamin A (Table 3).
Both combination 2= and 4= (containing the different UV-filters
in the same proportion used in the formulations under study) were
not considered phototoxic (MPE lower than 0.15). There was an
enhancement of MPE values, when vitamin A palmitate was added
to these combinations (comb 2a= and comb 4a=), however these
combinations where still considered not phototoxic (Table 4).
When combinations 2 and 4 (containing the different UV-filters
in the proportion 1:1:1) were evaluated, there was an enhance-
ment of MPE values, which were closer to borderline phototoxicity
values. When vitamin A palmitate was added to these combina-
tions, comb 2A and comb 4A had their MPE enhanced to 0.310
and 0.229, respectively, indicating a synergistic effect of vitamin
A palmitate on phototoxicity of these combinations containing
avobenzone. When a lower concentration of vitamin A was added
to these UV-filters combinations, comb 2a and 4a (containing a
proportion of UV-filters/vitamin A 1:0.1), a reduction of MPE val-
ues was observed (0.169 and 0.181, respectively), however these
combinations were still considered phototoxic (Table 4).
3.2. Phototoxicity test in human 3-D skin model (H3D-PT)
In order to evaluate the relevance of positive results obtained in
the 3T3-NRU-PT with respect to bioavailability in human skin, the
four formulations under study, containing or not vitamin A palmi-
tate, as well as the combinations 2 and 4, containing avobenzone
were submitted to the H3D-PT test.
The results of the phototoxicity assay using the human skin
model are given in Figs. 1 to 3 as the mean% solvent control MTT
conversion (n = 2) in the presence and absence of UV light. Un-
treated control tissues gave a mean OD value in the MTT assay of
1.983 without UV and there was no significant effect of solvent
treatment (C12–15 alkyl benzoate (mean OD value 1.854) on
MTT conversion. In addition, the UV exposure did not have any
effect on MTT conversion indicating that the cultures were of
satisfactory viability (85%).
Bergamot oil was phototoxic only in the highest concentration
tested (10% in C12–15 alkyl benzoate) as expected (Kejlová et al.,
2007), with a reduction in MTT conversion in the presence of UV
to approximately 40% of that of control tissues.
Fig. 2 shows that no phototoxicity was detected with the appli-
cation of the formulations 1, 2, 3 and 4, since none of the (+UVA)
tissues revealed a decrease in viability exceeding 30% when com-
pared with the (ÀUVA) tissues. The presence of vitamin A palmi-
tate did not alter tissue viability.
Fig. 3 shows that no phototoxicity was detected with the appli-
cation of the combinations studied, since none of the (+UVA) tis-
sues revealed a decrease in viability exceeding 30% when
compared with the (ÀUVA) tissues, except combination 2 in
the highest concentration tested (10% in C12–15 alkyl benzoate),
with a reduction in MTT conversion in the presence of UV to
approximately 53% of the ÀUV tissues (Fig. 3A). There was a slight
dose-related reduction in MTT conversion with the enhancement
of concentrations of combination 2 tested. The enhancement of
vitamin A palmitate concentration did not reduce tissue viability
(Fig. 3D) or protected the tissues from UVA-induced damage.
Previous studies showed that bergamot oil from different com-
panies was classified as phototoxic in the 3T3 NRU PT and pre-
sented borderline results in H3D PT, which was also dependent
on the solvent used (Kand’árová, 2006; Kejlová et al., 2007).
Despite the higher permeability of Human 3-D Skin Model com-
pared to human skin in vivo, these authors found a good correlation
of the photopotency of bergamot oils diluted in sesame oil, when
Human 3-D Skin Model and human in vivo photopatch tests result
were compared; however they stated that the extrapolation of
in vitro results to the human situation may be performed only to
a limited extent.
However other studies showed that phototoxicity prediction of
avobenzone did not have a good correlation when the monolayer
(3T3 NRU) and the reconstructed human skin (H3D PT) were
compared, which could be due to its low skin penetration, which
reduces its viable epidermis availability (Kandˇárová et al., 2005,
2006; Trauer et al., 2006).
Table 1
Phototoxicity of isolated UV-filters and vitamin under study.
Substance Run PIF MPE
Octyl methoxycinnamate (OMC) 1 1.303 0.031
2 1.196 À0.067
Octocrylene (OC) 1 1.664 À0.003
2 1.818 À0.040
Octyl salicylate (OS) 1 1.756 0.109
2 1.043 0.061
3 1.748 0.016
4-Methylbenzilidene camphor (MBC) 1 1.057 À0.197
2 1.514 À0.109
Benzophenone-3 (BP-3) 1 – À0.057
2 1.415 0.025
Avobenzone (AVB) 1 3.428 0.177b
2 21.034 0.358b
3 9.973 0.425b
Vitamin A palmitate (Vit A) 1 – 0.133a
2 – À0.024
3 – 0.034
4 – 0.282b
– Not determined.
a
Probably phototoxic.
b
Phototoxic (according to OECD TG 432).
Table 2
Phototoxicity of avobenzone and vitamin A palmitate under various intensities of
UVA radiation (2, 4 and 8 J/cm2
). (n = 1).
Substance UVA dose (J/cm2
) PIF MPE
Vitamin A palmitate 2 – –0.050
4 – 0.075
8 – 0.023
Avobenzone 2 0.826 0.017
4 3.027 0.101a
8 12.140 0.294b
– Not determined.
a
Probably phototoxic.
b
Phototoxic (according to OECD TG 432).
Table 3
Phototoxicity of UV-filters and vitamin under study in combination: combinations 1
and 3.
Combination 1
(OMC, BP-3 and OS)
PIF MPE Combination 3
(OMC, BP-3 and OC)
PIF MPE
Comb 1 1.25 0.016 Comb 3 1.240 À0.022
Comb 1= 1.103 0.003 Comb 3= 1.145 0.031
Comb 1A 1.315 0.027 Comb 3A 2.011 0.032
Comb 1a 1.157 0.007 Comb 3a 1.383 À0.004
Comb 1a= 1.452 0.057 Comb 3a= 2.12 0.073
Comb 1 and 3: containing the different UV-filters in the proportion (1:1:1).
Comb 1= and 3=: containing the different UV-filters in the same proportion used in
the formulations under study.
Comb 1A and 3A: containing a proportion of UV-filters/vitamin A 1:1.
Comb 1a and 3a: containing a proportion of UV-filters/vitamin A 1:0.1.
Comb 1a= and 3a=: containing the different UV-filters and vitamin A in the same
proportion used in the formulations under study.
L.R. Gaspar et al. / Toxicology in Vitro 27 (2013) 418–425 421

5. Hayden et al. (2005) observed that avobenzone had the highest
toxicity in culture human keratinocytes, however it did not pene-
trate the skin and thus the concentration of the UV-filter detected
on viable epidermis after topical application was at least 5-fold
lower than the toxic concentration on keratinocytes monolayer.
Other authors observed that 1 h after application of a sunscreen
formulation containing avobenzone (Parsol™ 1789), this UV-filter
were located in the upper 30% of the horny layer and did not reach
the living cells (Lademann et al., 2009).
Organic UV-filters are among the most common agent groups
currently responsible for photo-allergic contact dermatitis. A mul-
ticenter photopatch test study conducted with 1031 patients in 30
European centers found that the UV-filters octocrylene, benzophe-
none-3 and avobenzone most frequently elicited photo-allergic
contact dermatitis and they also reported some cross-reactions be-
tween some UV-filters combinations (EMCPPTS, 2012).
This way, although the extrapolation of the positive results ob-
tained in the present study to the human situation may be per-
formed only to a limited extent, they are valid to investigate the
phototoxic potential of new combinations of UV-filters and antiox-
idant substances like vitamin A.
The results obtained in the present study showed that
despite the four formulations studied did not present any acute
0
20
40
60
80
100
120
140
0.1 1 10 control
MTT(%ofuntreatedcontrol)
[%]
UV-
UV+
Fig. 1. Bergamot oil (positive control) phototoxicity (human skin model) and
negative control (indicator of tissue viability). Results are the mean of two
independent experiments ± SEM.
(A) Formulations 1 and 2 (B) Formulations 3 and 4
0
20
40
60
80
100
120
140
MTT(%ofuntreatedcontrol)
[%]
UV-
UV+
0
20
40
60
80
100
120
140
Form 1 Form 1A Form 2 Form 2A
Form 3 Form 3A Form 4 Form 4A
MTT(%ofuntreatedcontrol)
[%]
Fig. 2. Phototoxicity (human skin model) of the formulations containing (form 1A, form 2A, form 3A and form 4A) or not (form 1, form 2, form 3 and form 4) vitamin A
palmitate. Results are the mean of two independent experiments ± SEM.
Table 4
Phototoxicity of UV-filters and vitamin under study in combination: combinations 2 and 4.
Combination 2 (OMC, AVB, MBC) Run PIF MPE Combination 4 (OMC, AVB, OC) Run PIF MPE
Comb 2 1 1.924 0.043 Comb 4 1 2.011 0.18b
2 2.925 0.17b
2 1.594 0.149a
3 2.9 0.167b
3 2.9 0.158b
Comb 2= 1 – 0.105a
Comb 4= 1 – 0.156b
2 1.007 0.002 2 1.959 0.114a
Comb 2A 1 5.047 0.322b
Comb 4A 1 4.643 0.367b
2 3.15 0.297b
2 2.5 0.154b
Comb 2a 1 4.313 0.248b
Comb 4a 3 1.724 0.165b
2 2.48 0.149a
1 2.295 0.13a
3 1.992 0.111a
2 3.3 0.25b
Comb 2a= 1 – 0.071 Comb 4a= 3 1.944 0.163b
2 3.113 0.122a
1 – 0.14a
2 1.139 0.088
Comb 2 and 4: containing the different UV-filters in the proportion (1:1:1).
Comb 2 = and 4=: containing the different UV-filters in the same proportion used in the formulations under study.
Comb 2A and 4A: containing a proportion of UV-filters/vitamin A 1:1.
Comb 2a and 4a: containing a proportion of UV-filters/vitamin A 1:0.1.
Comb 4a = and 4a=: containing the different UV-filters and vitamin A in the same proportion used in the formulations under study.
– not determined.
a
Probably phototoxic.
b
Phototoxic (according to OECD TG 432).
422 L.R. Gaspar et al. / Toxicology in Vitro 27 (2013) 418–425

6. phototoxicity potential due to their reduced penetration, the com-
bination 2 containing octyl methoxycinnamate (OMC), avobenzone
(AVB) and 4-methylbenzilidene camphor (MBC) presented an indi-
cation of phototoxicity that should be better investigated. On the
other hand, some previous studies of our group performed in
humans showed that some formulations containing vitamins could
induce allergic responses after a 2-week period of application,
mainly when applied on the face (Gaspar et al., 2008). Thus,
although no acute phototoxicity was detected in the H3D PT mod-
el, the formulations may have photoallergic or chronic phototoxic-
ity and nowadays the development of additional models and
endpoints is a challenge among researchers to avoid underpredic-
tions and to increase the sensitivity of the these in vitro assays to
replace animal use.
(A) Combination 2 (B) Combination 2A
(C) Combination 2a, 2= and 2a= (D) Vitamin A palmitate
(E) Combination 4 (F) Combination 4A
0
20
40
60
80
100
120
140
0.1 0.316 1 3.16 10
MTT(%ofuntreatedcontrol)
[%]
0
20
40
60
80
100
120
140
0.1 1 10
MTT(%ofuntreatedcontrol)
[%]
0
20
40
60
80
100
120
140
Ass2a (10%)Ass2= (10%) Ass2a=
(10%)
MTT(%ofuntreatedcontrol)
[%]
0
20
40
60
80
100
120
140
1,000 10,000
MTT(%ofuntreatedcontrol)
[%]
0
20
40
60
80
100
120
140
0.010 0.100 0.316 1.000 10.000
MTT(%ofuntreatedcontrol)
[%]
0
20
40
60
80
100
120
140
0.100 1.000 10.000
MTT(%ofuntreatedcontrol)
[%]
Fig. 3. Phototoxicity (human skin model) of the combinations of UV-filters and vitamin A palmitate. (A) Combination 2: containing the three different UV-filters in the same
concentration (1:1:1). (B) Combination 2A: containing UV-filters/vitamin A palmitate in the same proportion (1:1). (C) Combination 2a: containing UV-filters/vitamin A
palmitate in a different proportion (1:0.1); combination 2=: containing UV-filters and vitamin A palmitate in the same proportion used in formulation 2 (1:1); combination
2a=: containing UV-filters and vitamin A palmitate in the same proportion used in formulation 2A. (D) Vitamin A palmitate. (E) Combination 4: containing the three different
UV-filters in the same concentration (1:1:1). (F) Combination 4A: containing UV-filters/vitamin A palmitate in the same proportion (1:1). Results are the mean of two
independent experiments ± SEM.
L.R. Gaspar et al. / Toxicology in Vitro 27 (2013) 418–425 423

7. 4. Conclusions
The results obtained when avobenzone and vitamin A palmitate
were submitted to 3T3 NRU Phototoxicity test showed that avo-
benzone presented a pronounced phototoxicity enhancement that
was UVA dose dependent. However when vitamin A was analyzed,
the obtained results showed that vitamin A presented a tendency
to a weak phototoxic potential that was not confirmed in the
UVA dose response study.
When combinations 2 and 4 containing avobenzone were eval-
uated, there was an enhancement of MPE values, which were clo-
ser to borderline phototoxicity values. A synergistic effect of
vitamin A palmitate on the phototoxicity of combinations contain-
ing avobenzone was observed.
The results of the phototoxicity assay using the human skin
model (H3D-PT) did not confirm the positive results obtained in
the 3T3-NRU-PT; however despite the four formulations studied
did not present any acute phototoxicity potential, the combination
2 containing octyl methoxycinnamate (OMC), avobenzone (AVB)
and 4-methylbenzilidene camphor (MBC) presented an indication
of phototoxicity that should be better investigated. Thus, although
no acute phototoxicity was detected in the H3D PT model, the for-
mulations may have photoallergic or chronic phototoxicity and
thus additional studies must be performed in terms of the fre-
quency of photoallergic or chronic phototoxicity in humans, since
the proposed tests cannot predict the exact incidence of phototoxic
reactions in humans.
Conflict of interest
The authors do not recognize any actual or potential conflict of
interest including any financial, personal or other relationships
with other people or organizations that could inappropriately
influencethe work.
Acknowledgements
The study was supported by a Grant project of the Fundação de
Amparo à Pesquisa do Estado de São Paulo (FAPESP) No. 08/58920-
0 and by Federal Institute for Risk Assessment (BfR).
References
Antille, C., Tran, C., Sorg, O., Carraux, P., Didierjean, L., Saurat, J.H., 2003. Vitamin A
exerts a photoprotective action in skin by absorbing ultraviolet B radiation. J.
Invest. Dermatol. 121, 1163–1167.
Bouillon, C., 2000. Recent advances in sun protection. J. Dermatol. Sci. 23, S57–S61.
Bruynzeel, D.P., Ferguson, J., Andersen, K., Gonçalo, M., English, J., Goossens, A.,
Holzle, E., Ibbotson, S.H., Lecha, M., Lehmann, P., Leonard, F., Moseley, H.,
Pigatto, P., Tanew, A., 2004. European taskforce for photopatch testing.
Photopatch testing: a consensus methodology for Europe. J. Eur. Acad.
Dermatol. Venereol. 18, 679–682.
Bryden, A.M., Moseley, H., Ibbotson, S.H., et al., 2006. Photopatch testing of 1155
patients: results of the UK multicentre photopatch study group. Br. J. Dermatol.
155, 737–747.
Cambon, M., Issachar, N., Castelli, D., Robert, C., 2001. An in vivo method to assess
the photostability of UV filters in a sunscreen. J. Cosmet. Sci. 52, 1–11.
Chatelain, E., Gabard, B., 2001. Photostabilization of butyl
methoxydibenzoylmethane (Avobenzone) and ethylhexyl methoxycinnamate
by bis-ethylhexyloxyphenol methoxyphenyltriazine (Tinosorb S), a new UV
broadband filter. Photochem. Photobiol. 74, 401–406.
Deleo, V.A., Suarez, S.M., Maso, M.J., 1992. Photoallergic contact dermatitis. Results
of photopatch testing in New York, 1985–1990. Arch. Dermatol. 128, 1513–
1518.
European Multicentre Photopatch Test Study (EMCPPTS) Taskforce. 2012. A
European multicentre photopatch test study. Br. J. Dermatol. 166, 1002–1009.
Gaspar, L.R., Camargo Jr., F.B., Gianeti, M.D., Maia Campos, P.M.B.G., 2008.
Evaluation of dermatological effects of cosmetic formulations containing
Saccharomyces cerevisiae extract and vitamins. Food Chem. Toxicol. 46, 3493–
3500.
Gaspar, L.R., Maia Campos, P.M.B.G., 2006. Evaluation of the photostability of
different UV filters associations in a sunscreen. Int. J. Pharm. 307, 123–128.
Gonzalez, H., Tarras-Wahlberg, N., Stromdahl, B., Juzeniene, A., Moan, J., Larko, O.,
Rosen, A., Wennberg, A.M., 2007. Photostability of commercial sunscreens upon
sun exposure and irradiation by ultraviolet lamps. BMC Dermatol. 7, 1–9.
Hayden, C.G.J., Cross, S.E., Anderson, C., Saunders, N.A., Roberts, M.S., 2005.
Sunscreen penetration of human skin and related keratinocyte toxicity after
topical application. Skin Pharmacol. Physiol. 18, 170–174.
Holzhütter, H.G., 1997. A general measure of in vitro phototoxicity derived from
pairs of dose–response curves and its use for predicting the in vivo
phototoxicity of chemicals. Altern. Lab. Anim. 25, 445–462.
Kand’árová, H., 2006. Evaluation and Validation of Reconstructed Human Skin
Models as Alternatives to Animal Tests in Regulatory Toxicology, p. 158. PhD
Thesis – Department of Biology, Chemistry and Pharmacy – Freie Universität
Berlin, Berlin.
Kandˇárová, H., Kejlová, K., Jírová, D., Tharmann, J., Liebsch, M., 2005. ECVAM
feasibility study: can the prevalidated in vitro skin model phototoxicity assay
be upgraded to quantify phototoxic potency of topical phototoxins? ALTEX 22,
156, Special Issue.
Kejlová, K., Jírová, D., Bendová, H., Kandárová, H., Weidenhoffer, Z., Kolárová, H.,
Liebsch, M., 2007. Phototoxicity of bergamot oil assessed by in vitro techniques
in combination with human patch tests. Toxicol. In Vitro 21, 1298–1303.
Kerr, A., Ferguson, J., 2010. Photoallergic contact dermatitis. Photodermatol.
Photoimmunol. Photomed. 26, 56–65.
Kerr, A.C., Ferguson. J., Haylett, A.K., et al., 2012. A European multi-centre
photopatch test study (EMCPPTS). Br. J. Dermatol. http://dx.doi.org/10.1111/
j.1365-2133.2012.10857.x.
Lademann, J., Jacobi, U., Surber, C., Weigmann, H.-J., Fluhr, J.W., 2009. The tape
stripping procedure – evaluation of some critical parameters. Eur. J. Pharm.
Biopharm. 72, 317–323.
Lhiaubet-Vallet, V., Marin, M., Jimenez, O., Gorchs, O., Trullas, C., Miranda, M.A.,
2010. Filter–filter interactions. Photostabilization, triplet quenching and
reactivity with singlet oxygen. Photochem. Photobiol. Sci. 9, 552–558.
Liebsch, M., Barrabas, C., Traue, T., Spielmann, H., 1997. Entwicklungeines in vitro
tests auf dermalephototoxizität in einem modell menschlicher epidermis
(EpiDermTM). ALTEX 14, 165–174.
Liebsch, M., Spielmann, H., 1998. INVITTOX Protocol No. 78: 3T3 NRU Phototoxicity
Assay. European Commission DG-JRC, ECVAM, SIS Database, Last update
October 2002.<http://ecvamsis.jrc.it/invittox/static/index.html>.
Liebsch, M., Spielmann, H., Pape, W., Krul, C., Deguercy, A., Eskes, C., 2005. UV-
induced effects. Altern. Lab. Anim. 33, 131–146.
Liebsch, M., Traue, D., Barrabas, C., Spielmann, H., Gerberick, G.F., Cruse, L.,
Diembeck, W., Pfannenbecker, U., Spieker, J., Holzhütter, H.-G., Brantom, P.,
Aspin, P., Southee, J., 1999. Prevalidation of the Epiderm™ phototoxicity test. In:
Clark, D.G., Lisansky, S.G., Macmillan, R. (Eds.), Alternatives to Animal Testing II.
COLIPA, Brussels, pp. 160–167.
Lodén, M., Beitner, H., Gonzalez, H., Edström, D.W., Akerström, U., Austad, J.,
Buraczewska-Norin, I., Matsson, M., Wulf, H.C., 2011. Sunscreen use:
controversies, challenges and regulatory aspects. Br. J. Dermatol. 165, 255–262.
Maia Campos, P.M.B.G., Ricci, G., Semprini, M., Lopes, R.A., 1999. Histopathological,
morphometric and stereological studies of dermocosmetic skin formulations
containing vitamin A and/or glycolic acid. J. Cosmet. Sci. 50, 159–170.
Maibach, H., Marzulli, F., 2004. Photoirritation (Phototoxicity) Testing in Humans.
Dermatotoxicology, Sixth ed., pp. 871–876.
Marzulli, F.N., Maibach, H.I., 1970. Perfume phototoxicity. J. Soc. Cosmet. Chem. 21,
695–715.
Marzulli, F.N., Maibach, H.I., 1985. Phototoxicity from topical agents. In: Maibach,
H.I., Lowe, N.J. (Eds.), Models in Dermatology, vol. 2. S. Karger, Basel, p. 349.
Nohynek, G.J., Schaefer, H., 2001. Benefit and risk of organic ultraviolet filters.
Regul. Toxicol. Pharmacol. 33, 285–299.
OECD, 2004. OECD Guidelines for the Testing of Chemicals Test No. 432: In Vitro 3T3
NRU Phototoxicity Test (Original Guideline, adopted 13th April 2004) <http://
www.oecd.org/>.
Osterwalder, U., Herzog, B., 2010. The long way towards the ideal sunscreen - where
we stand and what still needs to be done. Photochem. Photobiol. Sci. 9, 470–
481.
Paris, C., Lhiaubet-Vallet, V., Jiménez, O., Trullas, C., Miranda, M.A., 2009. A blocked
diketo form of avobenzone: photostability, photosensitizing properties and
triplet quenching by a triazine-derived UVB-filter. Photochem. Photobiol. 85,
178–184.
Pinnell, S.R., 2003. Cutaneous photodamage, oxidative stress, and topical
antioxidant protection. J. Am. Acad. Dermatol. 48, 20–22.
Rieger, M.M., 1997. Photostability of cosmetic ingredients on the skin. Cosmet. Toil.
112, 65–72.
Schauder, S., Ippen, H., 1997. Contact and photocontact sensitivity to sunscreens.
Review of a 15-year experience and of the literature. Contact Dermatitis 37,
221–232.
Schrader, A., Jakupovic, J., Baltes, W., 1994. Photochemical studies on trans- 3-
methylbutyl 4-methoxycinnamate. J. Soc. Cosmet. Chem. 45, 43–52.
Spielmann, H., Balls, M., Dupuis, J., Pape, W.J.W., Pechovitch, G., De Silva, O.,
Holzhütter, H.G., Clothier, R., Desolle, P., Gerberick, F., Liebsch, M., Lovell, W.W.,
Maurer, T., Pfannenbecker, U., Potthast, J.M., Csato, M., Sladowski, D., Steiling,
W., Brantom, P., 1998. The international EU/COLIPA In vitro phototoxicity
validation study: results of phase II (blind trial), Part 1: the 3T3 NRU
phototoxicity test. Toxicol. In Vitro 12, 305–327.
Tolleson, W.H., Cherng, S., XIA, Q., Boudreau, M., Yin, J.J., Wamer, W.G., Howard, P.C.,
Hongtao, Y., Fu, P.P., 2005. Photodecomposition and phototoxicity of natural
retinoids. Int. J. Environ. Res. Public Health 2, 147–155.
424 L.R. Gaspar et al. / Toxicology in Vitro 27 (2013) 418–425

8. Trauer, S., Kandˇárová, H., Schmidt, E., Traue, D., Schaefer, U., Lehr, C.-M., Liebsch, M.,
2006. In vitro skin absorption of buthyl methoxydibenzoylmethane determined
in human epidermis and in a reconstructed human epidermis model. Naunyn
Schmiedebergs Arch. Pharmacol. 372 (Suppl. 1), 130.
Xia, Q., Yin, J.J., Wamer, W.G., Cherng, S.H., Boudreau, M.D., Howard, P.C., Yu, H., Fu,
P.P., 2006. Photoirradiation of retinyl palmitate in ethanol with Ultraviolet light
– formation of photodecomposition products, reactive oxygen species, and lipid
peroxides. Int. J. Environ. Res. Public Health 3, 185–190.
L.R. Gaspar et al. / Toxicology in Vitro 27 (2013) 418–425 425