Endocrine Active UV Filters: Developmental Toxicity and Exposure Through Breast Milk

Abstract

Several UV filters exhibit endocrine activity. Evidence for transdermal passage and presence in the food chain (fish) suggests potential exposure of humans during development. Developmental toxicity was studied in rats for the estrogenic UV filters 4-methylbenzylidene camphor (4-MBC, 0.7, 7, 24, 47 mg/kg/day) and 3-benzylidene camphor (3-BC, 0.07, 0.24, 0.7, 2.4, 7 mg/kg/day) administered in chow to the parent generation before mating, during pregnancy and lactation, and to the offspring until adulthood. Neonates exhibited enhanced prostate growth after 4-MBC and altered uterine gene expression after both filters. 4-MBC and 3-BC delayed male puberty and affected reproductive organ weights of adult offspring. Interactions with the thyroid were noted. Expression and estrogen sensitivity of target genes and nuclear receptor coregulators were altered at mRNA and protein levels in adult uterus, prostate and brain. Female sexual behavior was affected by 4-MBC and 3-BC, estrous cycles by 3-BC. Classical endpoints exhibited LOAELs/NOAELs of 7/0.7 mg/kg/day for 4-MBC and 0.24/0.07 mg/kg/day for 3-BC. Molecular endpoints were affected by the lowest doses. In order to obtain information on human exposure, we conducted a monitoring study on human milk with three series of mother-child pairs (2004, 2005, 2006), with focus on cosmetic UV filters in relation to other endocrine disrupters. Methods for UV filter analysis followed the principles of European standardized methods for pesticide residue analysis (EN 15289). In cohorts 2004 and 2005, 78.8% of women reported use of product(s) containing cosmetic UV filters in a questionnaire, and 76.5% of milk samples contained these filters. Use of UV filters and concentration in human milk were significantly correlated. The results agree with the idea of transdermal passage of UV filters. They also indicate that it may be possible to reduce human exposure during critical periods such as pregnancy and lactation by transiently abstaining from use.

Full text

ENDOCRINE DISRUPTORS: CHEMICALS OF CONCERN 345
CHIMIA 2008,62, No. 5
Chimia 62 (2008) 345–351
© Schweizerische Chemische Gesellschaft
ISSN 0009–4293
Endocrine Active UV Filters: Developmental
Toxicity and Exposure Through Breast Milk
Margret Schlumpf*ab, Karin Kypkec , Claudia C. Vöktd , Monika Birchlerd, Stefan Durrerb, Oliver
Faassb, Colin Ehnesb , Michaela Fuetschb , Catherine Gailleab, Manuel Henselerb , Luke Hofkampe,
Kirsten Maerkelb, Sasha Reolonb , Armin Zenkerf , Barry Timmse, Jesus A. F. Tresguerresg, and
Walter Lichtensteigerab
Abstract: Several UV filters exhibit endocrine activity. Evidence for transdermal passage and presence in the food
chain (fish) suggests potential exposure of humans during development. Developmental toxicity was studied in rats
for the estrogenic UV filters 4-methylbenzylidene camphor (4-MBC, 0.7, 7, 24, 47 mg/kg/day) and 3-benzylidene
camphor (3-BC, 0.07, 0.24, 0.7, 2.4, 7 mg/kg/day) administered in chow to the parent generation before mating,
during pregnancy and lactation, and to the offspring until adulthood. Neonates exhibited enhanced prostate growth
after 4-MBC and altered uterine gene expression after both filters. 4-MBC and 3-BC delayed male puberty and
affected reproductive organ weights of adult offspring. Interactions with the thyroid were noted. Expression and
estrogen sensitivity of target genes and nuclear receptor coregulators were altered at mRNA and protein levels
in adult uterus, prostate and brain. Female sexual behavior was affected by 4-MBC and 3-BC, estrous cycles by
3-BC. Classical endpoints exhibited LOAELs/NOAELs of 7/0.7 mg/kg/day for 4-MBC and 0.24/0.07 mg/kg/day for
3-BC. Molecular endpoints were affected by the lowest doses. In order to obtain information on human exposure,
we conducted a monitoring study on human milk with three series of mother−child pairs (2004, 2005, 2006), with
focus on cosmetic UV filters in relation to other endocrine disrupters. Methods for UV filter analysis followed the
principles of European standardized methods for pesticide residue analysis (EN 15289). In cohorts 2004 and 2005,
78.8% of women reported use of product(s) containing cosmetic UV filters in a questionnaire, and 76.5% of milk
samples contained these filters. Use of UV filters and concentration in human milk were significantly correlated. The
results agree with the idea of transdermal passage of UV filters. They also indicate that it may be possible to reduce
human exposure during critical periods such as pregnancy and lactation by transiently abstaining from use.
Keywords:3-Benzylidene camphor (3-BC) · Developmental toxicity · Human milk ·
4-Methylbenzylidene camphor (4-MBC) · UV filters
1. Endocrine Activity of UV Filters
UV filters are either physical filters like
titanium dioxide and zinc oxide, which
mainly scatter and reflect UV rays, or or-
ganic molecules absorbing light in the UV
range (UVA 400− 320 nm, UVB 320− 280
nm). These organic compounds often pos-
sess single or multiple aromatic structures
capable of absorbing energetic solar pho-
tons and returning to the ground state by
thermally emitting the absorbed energy. [1]
Only substances listed in cosmetic direc-
tives like EU Cosmetics Directive, Swiss
Ordinance for Cosmetics, are allowed for
use in sunscreens and as additives in cos-
metics. Currently 27 UV filters are permit-
ted for cosmetic use in Europe. In spite of
considerable structural similarities with au-
thorized cosmetic UV filters, technical UV
filters in plastics and other products need
not be declared.
Since the introduction of cosmetic UV
filters, the main concern regarding their
use was the efficiency to protect human
skin from adverse effects of UV light while
avoiding dermatological side effects. Ob-
servations made in the 1990s on penetration
of human skin by UV filters,[2,3] and on their
presence in fish[4] indicated the possibility
of systemic effects of these chemicals and
prompted us to investigate their endocrine
activity.When tested on MCF-7 cells in
vitro , a number of UV filters used in sun-
screens exhibited estrogenic activity; some
also stimulated growth of the immature rat
uterus in a short-term in vivo test for estro-
genic activity. [5] Estrogenic activity of UV
filters has subsequently been confirmed in
several in vitro and in vivo tests on mammals
and fish.[6−14] Certain UV filters also dis-
play anti-androgenic activity in vitro,[15,16]
and can affect the thyroid axis.[17−19]
Two UV filters with comparatively high
estrogenic activity are 4-methylbenzylidene
camphor (4-MBC) and 3-benzylidene cam-
phor (3-BC).[5,12] Both compounds exhibit
ER beta preference, but they are also active
* Correspondence: Dr. M. Schlumpfab
Tel.: +41 43 233 9517
Fax: +41 43 268 9573
E-mail: margret.schlumpf@access.uzh.ch
a GREEN Tox GmbH
Langackerstrasse 49
CH-8057 Zürich
b Institute of Anatomy
University of Zurich
Winterthurerstrasse 190, CH-8057 Zurich
c State Institute for Chemical and Veterinary Analysis
of Food
Freiburg, Germany
d University Women’s Hospital
Basel
e University of South Dakota
Vermillion, USA
f University of Applied Sciences Northwestern
Switzerland
Muttenz
g Complutense University
Madrid, Spain
doi:10.2533/chimia.2008.345

ENDOCRINE DISRUPTORS: CHEMICALS OF CONCERN 346
CHIMIA 2008,62, No. 5
in ER alpha, typically in in vivo tests such
as the uterotrophic assay, 4-MBC possibly
because of a hydroxylated metabolite.[20]
Whereas a 90-day dermal exposure study
in adult rats reported to the Scientific Com-
mittee on Consumer Products[21] failed to
disclose estrogenic effects of 4-MBC in
this chronic model, significant estrogenic
and antiestrogenic effects were observed
in 90-day studies on adult ovariectomized
rats that addressed this question more spe-
cifically. Typical estrogen targets affected
by 4-MBC include luteinizing hormone,
leptin, fat depots, bone, and genes such
as insulin-like growth factor-1.[22,23] Inter-
actions with estrogenic mechanisms were
also observed after three months exposure
to benzophenone-2.[24,25] Effect patterns
of UV filters do not fully mimic those of
natural estrogens, which may be explained,
i.a., by partial agonist activity, different af-
finities for estrogen receptor (ER) alpha and
ER beta, or interactions with different hor-
mone axes.[5,12,17]
2. Developmental Toxicity of
4-Methylbenzylidene Camphor
(4-MBC) and 3-Benzylidene
Camphor (3-BC)
Development of the sexual pheno-
type depends on the intervention of sex
hormones. In mammals, testosterone (T)
directs development of the male, but this
action is not exclusively mediated by an-
drogen receptors (AR). In brain, T is con-
verted to estradiol (E2)[26,27] to control male
differentiation. Observations in aromatase
knockout mice indicate that development of
the female brain also depends on E2.[28] Cer-
tain peripheral male tissues such as prostate
convert T to dihydrotestosterone acting on
AR as well as to E2 acting on ER.[29,30] With
the exception of tissues with local conver-
sion of T to E2, endogenous estrogen lev-
els are very low in fetuses of both sexes in
rodents[31] and also comparatively low in
humans,[32] thus facilitating competition of
weak estrogenic chemicals for ER.
2.1. Treatment Design
Potential effects of exposure to 4-MBC
and 3-BC during pre- and postnatal de-
velopment were investigated in a mam-
malian model, the rat.[17,19,33−35] Male and
female rats of the parent generation were
exposed to 4-MBC (47, 24, 7.0, 0.7 mg/kg/
day) or 3-BC (7.0, 2.4, 0.7, 0.24, 0.07 mg/
kg/day) administered in chow at least ten
weeks before mating, females were further
treated during pregnancy and lactation, and
their F1 offspring until adulthood (age of
three months). 47 mg/kg/day 4-MBC cor-
responds to 40% of uterotrophic LOAEL,
3-BC was dosed according to uterotrophic
potency relative to 4-MBC.[12] Schauer et
al.[36] suggested that toxicity of 4-MBC
should be studied with dermal application.
Yet, certain processes of sexual differentia-
tion like sexual brain differentiation have
long been known to be highly sensitive to
pre- and postnatal stress and handling,[37] so
that results obtained with topical applica-
tion of chemicals to the skin of F0 and F1
rats, each for three months, would be ex-
pected to yield questionable results.
2.2. 4-MBC and 3-BC in Rat Milk
For comparison of experimental data
with internal human exposure, the chemi-
cals were determined in rat milk taken from
the stomach on postnatal day 6. Concentra-
tions (ng/g lipid) were as follows (mean ±
SD, number of samples/litters): 4-MBC, 7
mg/kg/day: 208.6 ±108.7 (5/5), 0.7 mg/kg/
day: 86.3 ±40.5 (4/3). 3-BC, 0.24 mg/kg/
day: 132.1 ±75.2 (3/3), 0.07 mg/kg/day:
not detectable (4/4). 4-MBC and 3-BC were
undetectable in milk of controls.
2.3. Postnatal Development
In the higher dose range, both UV fil-
ters reduced survival rate.[17,38] An impair-
ment of the developing immune system was
indicated by a decrease in thymus weight
at postnatal day (PN) 14. A prominent ef-
fect of both UV filters during early post-
natal development is the significant delay
of puberty in males (preputial separation).
Puberty onset in females remained unaf-
fected (Tables 1, 2 [17,34]). 4-MBC and 3-BC
thus mimic the typical effect of E2 only in
males, but differ from E2 in females, where
E2 advances puberty. [39] Body weight at pu-
berty was slightly reduced in females but
normal in males, indicating that the delay of
male puberty did not result from nutritional
effects. Adult body weights were at control
level after 4-MB[19,33,34] and 3-BC except
for a slight reduction after the highest dose
of 3-BC, possibly as a result of estrogenic
activity. [39]
2.4. Low-dose Effects in Neonatal
Uterus and Prostate
Estrogen target gene mRNA levels in
early postnatal uterus (PN 6), determined
by real-time RT PCR,[33] were affected at
doses as low as 0.07 mg/kg/day 3-BC and
0.7 mg/kg/day 4-MBC (Table 2). The same
doses also affected gene expression in sex-
ually dimorphic brain regions at PN 6 (M.
Fuetsch, C. Gaille, unpublished data). The
changes in mRNAs encoding for vascular
epithelial growth factor (VEGF), induc-
ible nitric oxide synthase (iNOS) and, in
part, endothelial cell nitric oxide synthase
(ecNOS) suggest that angiogenesis and
blood flow in uterus may be influenced.
Effects on male accessory sex glands were
investigated by morphometric analysis in
collaboration with L. Hofkamp and B.
Timms. Maternal exposure to 4-MBC re-
sulted in significant increases in the size of
prostate, seminal vesicles and coagulating
gland at PN 1 (day of birth). In line with
other findings on estrogenic chemicals,
marked differences in growth responses
of specific regions of the prostate were
observed (Table 1, Hofkamp et al.[40] and
unpublished data).
2.5. Reproductive Organs of Adult
Offspring: Regulation of Target
Gene Expression
Exposure to 4-MBC and 3-BC af-
fected reproductive organ weights (Tables
1, 2[17,33,34]). Testes of 4-MBC-exposed
offspring showed decreased weight at
PN14[38] and increased relative weight at
the highest dose in adulthood. The adult
finding is reminiscent of neonatal admin-
istration of weak estrogens.[41,42] In con-
trast, the decrease in prostate weight of
4-MBC-exposed offspring resembles the
effect of perinatal administration of the
potent ER agonist diethylstilbestrol.[41,43]
This suggests a differential sensitivity of
male target organs.
The same estrogen target genes were
studied in reproductive organs and brain of
male and female offspring in order to com-
pare effects on gene regulation. In ventral
and dorsolateral prostate and uterus, gene
expression was affected at mRNA and pro-
tein levels in a tissue-specific manner. [33,34]
The 4-MBC-induced decrease in prostate
weight was accompanied by a decrease in
AR, ER alpha, and insulin-like growth fac-
tor-I (IGF-I) (Table 1). In 4-MBC-exposed
uterus, affected target genes include ER al-
pha and progesterone receptor (PR) (Table
2). Effect patterns differed between the two
camphor derivatives also at the molecular
level, in spite of close structural relation-
ship and similar actions in acute assays for
estrogenicity.
Malfunctions may also be caused by
changes in the sensitivity of tissues to natu-
ral estrogens. In order to assess such chang-
es, 4-MBC-exposed offspring were gonad-
ectomized in adulthood, injected two weeks
later with a single dose of E2 (10 or 50 µg/
kg s.c.), and investigated 6 h after the in-
jection. 4-MBC exposure reduced the acute
up-regulation of PR and IGF-I and down-
regulation of ER alpha and AR mRNA
by E2 in uterus, and the down-regulation
of AR and IGF-I mRNA by E2 in ventral
prostate (Table 1[33,34]). The reduced up-
regulation of estrogen target genes in uterus
was accompanied by decreased steroid re-
ceptor coactivator-1 (SRC-1) protein levels,
while reduced down-regulation of genes in
prostate was paralleled by reduced nuclear
receptor corepressor (N-CoR) protein.[33,34]
This identifies nuclear receptor coregula-
tors as targets of endocrine receptors, and
suggests that they are involved in changes
in estrogen sensitivity.

ENDOCRINE DISRUPTORS: CHEMICALS OF CONCERN 347
CHIMIA 2008,62, No. 5
Table 1. Effect of 4-MBC and 3-BC on selected endpoints in male rat offspring
4-Methylbenzylidene camphor
[mg/kg/day p.o.]
3-Benzylidene camphor
[mg/kg/day p.o.]
0.7 724470.07 0.24 0.7 2.4 7
Puberty (preputial separation)
Adult body weight Ø
Delayed
Ø
Delayed
Ø
Delayed
ØØ
Ø
Ø
Ø
Ø
Delayed
Ø
Delayed

Testis
Postnatal Day 14, Testis relative weight
Adult F1, Testis relative weight Ø

Ø

Ø

ØØØØ
Prostate
Postnatal Day 1, Duct number (dorsal) and duct volume (ventral)
Adult F1, Ventral lobe relative weight
Ø
Ø


Ø
Ø
Ø
ØØØ
Gene expression, adult F1 prostate, mRNA/protein
Androgen receptor (AR) dorsolateral prostate (DP)
Androgen receptor (AR) ventral prostate (VP)
AR mRNA down-regulation by estradiol in VP
N-CoR protein, DP
N-CoR protein, VP
Ø/Ø
Ø/Ø

Ø
/Ø
Ø/()


( )
/ 
/ 

( )

/()
Ø/−
/
/

Ø
Ø/
/Ø

Ø
Ø/−
/−
/−
Ø/−
Central nervous system, adult F1, mRNA
Gene expression in Ventromedial Hypothalamic Nucleus
Estrogen Receptor-alpha
Progesterone Receptor
Ø
Ø

Ø

Ø

Ø

Ø
Ø

Ø

 , :significant increase or decrease versus control for p < 0.05 or better. Ø: no statistically significant change. blank or −: not analyzed.
PN 1 = day of birth. Adult F1 offspring: 12 weeks of age, studied under baseline conditions.
Data from Schlumpf et al. [17, 38], Durrer et al. [34], Maerkel et al. [19], Hofkamp et al. [40], O. Faass, M. Fuetsch, C. Ehnes, C. Gaille, unpublished data.
Table 2. Effect of 4-MBC and 3-BC on selected endpoints in female rat offspring
4-Methylbenzylidene camphor
[mg/kg/day p.o.]
3-Benzylidene camphor
[mg/kg/day p.o.]
0.7 724470.07 0.24 0.7 2.4 7
Puberty (vaginal opening)
Adult body weight Ø
Ø
Ø
Ø
Ø
Ø
ØØ
Ø
Ø
Ø
Ø
Ø
Ø
Ø

Ovary
Adult F1, Ovary relative weight ØØ
Uterus
Postnatal Day 6, Uterus relative weight
Adult F1, Uterus relative weight
Ø
Ø
Ø
ØØ
ØØ
ØØØ
Postnatal Day 6 Uterus, Gene expression, mRNA
Estrogen receptor-alpha
ecNOS
iNOS
VEGF

Ø


Ø
Ø

Ø





( )


Adult F1 Uterus, Gene expression, mRNA/protein
Progesterone Receptor (PR-A protein)
PR mRNA up-regulation by estradiol
SRC-1 protein
Ø/
Ø

Ø/Ø

Ø
/Ø

( )
/Ø

Ø/
Ø
Ø/ Ø
( )
Ø/ Ø

/Ø
Ø
Central nervous system, adult F1, mRNA
Gene expression in Ventromedial Hypothalamic Nucleus
Estrogen Receptor-alpha
Progesterone Receptor
PR mRNA up-regulation by estradiol
Female sexual behavior (proceptive and lordosis
behavior)
Estrous cycle




Ø


Ø

Ø


Ø
Ø

irregular
Ø
Ø
irregular



irregular



irregular
 , :significant increase or decrease versus control for p < 0.05 or better. Ø: no statistically significant change. blank: not analyzed.
PN 1 = day of birth. Adult F1 offspring: 12 weeks of age, studied under baseline conditions, females in diestrus.
Data from Schlumpf et al. [17, 38], Durrer et al. [33], Maerkel et al. [19], O. Faass, M. Fuetsch, C. Ehnes, C. Gaille, unpublished data.

ENDOCRINE DISRUPTORS: CHEMICALS OF CONCERN 348
CHIMIA 2008,62, No. 5
2.6. Sexually Dimorphic Gene
Expression in Brain and Female
Sexual Behavior
In consideration of the data on brain
differentiation and fetal estrogen levels as
outlined above, we hypothesized that the fe-
male brain should be sensitive to estrogenic
chemicals. Female sexual behavior was re-
corded in adult female offspring exposed to
one of two doses of 4-MBC (7, 24 mg/kg/
day) or 3-BC (2.4, 7 mg/kg/day). Exposed
females were mated with normal experi-
enced males in the evening of proestrus, at
the onset of the dark phase (16.00), when
behavioral receptivity of gonadally intact
rats is high.[44] Behavior was recorded on
coded videotapes in a room illuminated
by an infrared light source. Va ginal smears
were recorded for at least 10–14 days be-
fore behavioral testing. All four treatments
strongly suppressed female sexual behavior
(Table 2, Faass et al., unpublished data).
The treatments affected proceptive behav-
ior (jumping and ear wiggling, displayed to
attract the male), as well as receptive (lor-
dosis) behavior (decreased lordosis quotient
LQ = number of lordosis responses/number
of mounts x 100). At the same time, the male
attempting to mount was rejected more fre-
quently by the female (rejection behavior).
Detailed analyses of estrous cycles were
performed on additional groups of animals
for 21 and 16 days in 4-MBC- and 3-BC-
exposed offspring, respectively. In 4-MBC-
exposed offspring, female sexual behavior
was disturbed in the presence of normal es-
trous cycles, whereas 3-BC exposure caused
irregular cycles. The two functions thus are
differentially affected.
Gene expression was analyzed by real-
time RT PCR in adult male and female off-
spring in two brain regions involved in the
control of gonadal function and sexual be-
havior, medial preoptic region (MPO) and
ventromedial hypothalamic nucleus (VMH)
(Tables 1, 2, 4-MBC:[19,35] 3-BC: Faass et
al., unpublished data). Both compounds
caused sex- and region-specific changes in
ER, in nuclear receptor coactivator SRC-1,
and in target gene mRNA levels. A drop of
PR mRNA in female VMH down to male
levels emerged as a common feature ob-
served after all doses of 4-MBC and after
the higher two doses of 3-BC (tested for
behavioral effects). Reduced PR mRNA in
female VMH was correlated with impaired
female sexual behavior. A similar relation-
ship had been observed with a polybromi-
nated flame retardant (PBDE 99) and with
a PCB mixture.[45] Lordosis behavior is di-
rectly correlated with the expression of PR
mRNA in VMH of female rats.[46,47] Loss of
sexual dimorphism of PR in female VMH
thus appears to represent a signal of altered
regulation of PR that is linked with behav-
ioral impairment across different endocrine
disrupters.
3. UV Filters in Environment and
Food Chain
There is good evidence that pharmaceu-
ticals and ingredients of personal care prod-
ucts (PPCPs) can spread to the biosphere
and reach the food chain. UV filters may
be directly introduced into surface waters
during swimming or may enter wastewa-
ter from households or industry at several
levels of industrial production or commer-
cial use.[1] Cosmetic compounds, synthetic
perfumes and UV filters, were detected in
high amounts in Swiss sewage sludge.[48]
UV filters and synthetic musks are pres-
ent in surface waters and in biota at various
trophic levels, in particular in fish.[4,49−52]
UV filter levels in fish from rivers receiv-
ing inputs from wastewater treatment plants
(WWTPs) had considerably higher chemi-
cal loads than fish from Swiss lakes with in-
puts from WWTPs,[50] suggesting increased
availability of these contaminants for fish in
rivers. These studies identified WWTPs as
a major source for UV filters in the aquatic
environment and demonstrate the presence
of UV filters in the food chain.
4. Human Exposure: Monitoring
of Human Milk
Assessment of chemical risks requires
information on quality and quantity of
chemicals present in human body during
critical and sensitive life stages such as
pre- and postnatal development. Acute and
short-term (4 d) experiments with percu-
taneous application of 4-MBC to human
volunteers indicated transdermal passage
of the compound,[36,53] but such studies do
not yield information on internal exposure
of the human population under realistic
patterns of cosmetic use. This informa-
tion can be provided by analysis of human
milk, which informs on internal exposure
of mother and fetus and on contamination
of the food provided to the nursing infant.
Most of the existing data relate to organo-
chlorine compounds. Unfortunately, their
trend to decrease in human milk has been
considered as a success in the campaign
against chemical exposure of babies, with-
out asking for possible exposures to other
chemicals from the food web, like cosmet-
ics, pharmaceuticals, industrial or house-
hold products. Thus, phthalate exposure
has recently been linked with alterations
in male genital development and hormone
profiles.[54,55]
5. The Swiss Cohort
Since there was no information on inter-
nal exposure of human populations to UV
filters, and very limited information on ad-
ditional cosmetic ingredients, we started a
monitoring study of human milk at the Uni-
versity Hospital Basel with the approval of
the Basel University Ethics Committee. The
study focused primarily on chemical analy-
sis of UV filters in relation to several other
groups of endocrine-disrupting chemicals
(EDCs and EDC candidates) and consisted
of three different cohorts over three years
(2004, 2005, 2006). So far, the first two co-
horts have been evaluated. For the first time
the questionnaire given to the mothers con-
tained very detailed questions on the use of
cosmetics in pregnancy and lactation. The
aim was to detect a possible correlation
between exposure to certain UV filters and
their presence in human milk.
5.1. Questionnaires
All mothers had to fill out a question-
naire and to give written consent for par-
ticipation in the study. The questionnaires
asked for mother and child data on birth
date, sex, height, weight, sisters and broth-
ers, education, professional career, living
area (urban, suburban, rural), nutritional
and smoking habits of mothers. The ques-
tionnaire then asked for detailed qualitative
and semiquantitative (daily, weekly, month-
ly or less) use of different types and brands
of cosmetic products during pregnancy and
lactation, including sunscreens, lipsticks,
perfumes, deodorants, skin care creams,
body lotions, shower lotions, bubble baths,
hair dyes, make-ups etc.
5.2. Sampling of Human Milk
Sampling was supervised by Claudia
Vökt with the assistance of the study nurse
Monika Birchler. Care was taken to avoid
contamination. The mothers were instruct-
ed to clean breast and nipples thoroughly
with warm tap water before milk sampling.
The milk was obtained using a freshly hot
water-rinsed milk pump (Type Harmony,
Medela AG, Baar). The milk servings were
collected in a clean sterilized bottle (Schott
Duran ISO 4796) stored in the freezer at
− 20 °C. Milk sampling mainly included the
transitory phase of lactation (day 6 to 14 af-
ter birth), occasionally also the first days of
the mature phase of lactation (from 14 days
after birth on), rarely the colostral phase
(first 6 days after birth) (Wünschmann et
al.[56]). The numbers of milk servings per
bottle (around 100 ml) representing the in-
dividual milk sample of each mother taken
for chemical analysis, varied between 4 and
10.
5.3. Chemicals Analyzed
Our intention was to simultaneously
analyze different groups of EDCs in order
to obtain information on their relative im-
portance. Together with Karin Kypke from
the Community Reference Laboratory for
Pesticides in Food of Animal Origin at the

ENDOCRINE DISRUPTORS: CHEMICALS OF CONCERN 349
CHIMIA 2008,62, No. 5
State Institute for Chemical and Veterinary
Analysis of Food in Freiburg/Germany, we
analyzed eight of a total of 27 authorized
cosmetic UV filters (Tables 3, 4) in the same
human milk sample, as we wanted to know
whether these cosmetic UV filters concen-
trate in human milk, and whether there is a
correlation between use of cosmetics con-
taining these chemicals and their presence
in human milk. The same milk samples
were also analyzed for synthetic fragrances
such as nitro musks (musk xylene, musk ke-
tone), polycyclic musks including HHCB
(Galaxolide) and AHTN (Tonalide), mac-
rocyclic musks, several polybrominated
diphenylethers (BDE28, 47, 99, 100, 153,
154), organochlorine pesticides (including
DDT/DDE, methoxychlor, hexachloro-
cyclohexane (HCH), hexachlorobenzene,
toxaphene), seven indicator PCB conge-
ners, and cyclodiene insecticides (aldrin,
dieldrin, chordane, endrine, endosulfan,
heptachlor bromocyclene) (Schlumpf et al.,
in preparation).
5.4. Extraction of UV Filters
The amount of human milk used as a
sample depends on the lipid content of
the milk. Routinely a sample amount of
0.25−0.5 g lipid per sample was analyzed.
Human milk samples were centrifuged and
the UV filters of interest (Table 3) were
extracted from the cream together with
lipid, using sodium sulfate and the solvent
n-hexane/acetone (1:1) at first, followed
by dichlormethane/acetone (1:1). Follow-
ing evaporation of the solvent in a rotary
evaporator, the extract containing the UV
filters was re-dissolved in cyclohexane/
ethyl acetate (1:1), centrifuged and three
internal standards for the eight UV filters
were added. To remove lipid, gel perme-
ation chromatography was performed on
Bio-Beads S-X3 with cyclohexane/ethyl
acetate as eluting solvent. The eluate was
concentrated to a defined volume. Analo-
gous procedures were used for additional
groups of lipophilic xenobiotic substances
analyzed in the same milk sample, like
persistent organochlorine compounds, syn-
thetic musks and PBDEs. These data will
be presented in the final report on all three
cohorts (Schlumpf et al., in preparation).
In rats, the whole stomach of the pup
was homogenized and extracted using ac-
etone and n-heptane in a Dispomix homog-
enization system (Medic Tools). After add-
ing the internal standards, the sample was
shaken 15 min at 0 °C. and centrifuged, the
supernatant collected and dried. The extrac-
tion step was repeated and the UV filters
were separated from lipids by RP-HPLC
using a octadecylsilyl column (Zenker et
al., in preparation).The fraction containing
4-MBC and 3-BC was dried in a vacuum
centrifuge, re-dissolved in ethanol and de-
termined by GC-MS (see below).
5.5. Determination of UV Filters
The method for analysis of UV filters
in human milk samples followed the prin-
ciples of the European standardized meth-
ods for pesticide residue analysis.[57] Deter-
mination of all UV filters except Bp-2 was
done by GC-LRMS (GC: HP 6890; MS: HP
5973; 30 m HP5-MS, 0.25 mm i.d., 0.25
µ mfilm thickness + 2.5 m pre-column)
with MSD-EI detection mode, using se-
lected ion monitoring (SIM mode) and se-
lecting one target and three qualifier ions
as characteristic mass ions. To compensate
for matrix effects matrix-matched calibra-
tion was used. For determination of Bp-2
LC-MSD (LC: HP 1100; MS: Quattro LC,
50 ×2 mm Luna C18 (2), 5 µm Phenom-
enex) with the detection mode ESI pos was
applied, using multiple reaction monitoring
(MRM), eluent: A being 1 mM ammonium
acetate, pH 4.75 and eluent B being metha-
nol, using matrix-matched calibration. The
concentrations of the substances are report-
ed as ng per g of milk lipid (ng/g lipid). The
limit of quantification (LOQ) and the limit
of determination (LOD) for the UV filters:
HMS, 3-BC, BP-3, 4-MBC and OC were
4.0 (ng/g lipid) (for LOQ) and 2.0 (ng/g
lipid) for LOD. For OD-PABA, EHMC and
BP-2, LOQ was 2.0 ng/g lipid and LOD 1.0
ng/g lipid. The mean level for each residue
was calculated with the assumption of zero
level for undetected value and half LOQ for
levels determined between LOD and LOQ.
The levelwas stated as ‘nd’, i.e.undetected,
if it was below LOD.
5.6. Biostatistics
Possible relationships between use of
UV filters and chemical-analytical data
in human milk were analyzed by Va lentin
Rousson, Biostatistics Unit, University
of Zurich, using Pearson Chi-Square and
Fisher’s exact test.
5.7. Use of UV Filters in Cosmetics
and Presence in Human Milk
An analysis of the first two cohorts from
the Basel cohort study, pilot study (2004)
and Study 1 (2005), revealed that during the
periods of pregnancy and lactation, 78.8%
of the women used some cosmetic product
containing UV filters. In 76.5% of human
milk samples, UV filters were detected
(Table 4). Ethylhexyl-methoxicinnamate
(EHMC, previously known as octyl-meth-
oxycinnamate (OMC)) and octocrylene
(OC) were the UV filters most frequently
used according to the questionnaire and
most frequently detected in milk samples.
For these two filters, a significant correla-
tion between use and presence in human
milk could be demonstrated for the indi-
vidual chemical (p = 0.031 for EHMC, p
= 0.046 for OC, Fisher’s exact test). The
correlation was also significant across all
UV filters for use and presence in the cor-
responding milk sample (p = 0.009). Inter-
estingly, only 45.5% of women reported
use of sunscreens with UV filters, whereas
60.6% of the women used other cosmetics
containing UV filters.
These data demonstrate concentration
of UV filters in a relevant proportion of
human milk samples. Except for lipsticks
where oral uptake is probably important,
these results agree with the idea of transder-
mal passage of UV filters from cosmetics,
as proposed from animal and human stud-
ies.[5,36,53,58] However, it should be kept in
mind that there are also other sources for
Table 3. UV fi lters analyzed in human milk
Abbre-
viation
Chemical INCIa) NomenclaturePurity of
reference
chemical
Bp-2 2,2`4,4`- Tetrahydroxi-benzophenone Benzophenone-2 97%
Bp-3 2-Hydroxi-4-methoxi-benzophenone Benzophenone-3 98%
3-BC 3-Benzylidene-bornane-2-on 3-Benzylidene Camphor > 97%
4-MBC 3-(4`-Methyl)benzylidene bornane-2-on 4-Methylbenzylidene Camphor > 99.7%
EHMC
(OMC)
2-Ethylhexyl-4-methoxicinnamate Ethyl-hexylcinnamate
(Octyl-methoxycinnamate)
98%
HMS 3,3,5-Trimethyl-cyclohexyl-salicylate,
Homosalate
Homosalate > 98%
OC 2-Cyano-3,3`-diphenyl-acrylic acid
2`-ethyl-hexylester
Octocrylene 98%
OD-
PABA
4-Dimethylamino-benzoic acid-2 ethyl-
hexyl-ester
Octyl-dimethyl PABA > 98.5 %
a)INCI: International Nomenclature of Cosmetic Ingredients

ENDOCRINE DISRUPTORS: CHEMICALS OF CONCERN 350
CHIMIA 2008,62, No. 5
these compounds, since they are present in
the ecosphere and food chain (see above).
The Basel cohort study shows multiple
chemical exposures of neonates to groups
of identified and candidate EDCs. Accord-
ing to these results, it is evident that expo-
sure (use) of cosmetics containing UV fil-
ters will produce UV filter-positive human
milk samples. Abstinence from use of or-
ganic UV filter containing sunscreens and
cosmetics could therefore be considered as
an important step to diminish the total load
of chemicals in human milk, in order to re-
duce exposure during particularly sensitive
life stages.
6. Comparison of Experimental Rat
Data with Human Exposure
Our data indicate that pre- and postna-
tal exposure to 4-MBC and 3-BC can in-
terfere with sexual development at brain
and reproductive organ levels. Classical
toxicological endpoints such as puberty
and reproductive organ weights exhib-
ited lowest observed adverse effect lev-
els (LOAEL) and no observed adverse
effect levels (NOAEL) of 7 and 0.7 mg/
kg/day for 4-MBC and of 0.24 and 0.07
mg/kg/day for 3-BC, respectively. Mo-
lecular endpoints were affected by the
lowest doses studied. At the LOAEL of
7 mg/kg/day, 4-MBC concentration in rat
milk (208.6 ng/g lipid) was eleven times
the highest value so far found in human
milk (19 ng/g lipid, Table 4). The ratio
at the classical NOAEL and molecular
LOAEL of 0.7 mg/kg/day 4-MBC is 4.5.
This comparatively small ratio indicates
that the potential risk posed by UV filters
warrants further considerations.
Acknowledgements
The investigations were supported by Swiss
NationalResearchProgram 50, EU 5th Framework
Program (EURISKED), Swiss Federal Office for
the Environment, Hartmann-Müller Stiftung, and
the Olga Mayenfisch Stiftung.
Received: April 2, 2008
[1] D. L. Giokas, A. Salvador, A. Chisvert,
Trends Analyt. Chem.2007,26, 360.
[2] C. G. J. Hayden, M. S. Roberts, H. A. E.
Benson, Lancet 1997,350, 863.
[3] T. Felix, B. J. Hall, J. S. Brodbelt, Anal.
Chim. Acta 1998,371, 195.
[4] M. Nagtegaal, T. A. Ternes, W. Baumann,
R. Nagel, UWSF- Z. Umweltchem. Ökotox.
1997,9 , 79.
[5] M. Schlumpf, B. Cotton, M. Conscience,
V. Haller, B. Steinmann,W. Lichtensteiger,
Environ. Health Perspect.2001,109, 239.
[6] H. Holbech, U. Norum, B. Korsgaard, P.
Bjerregard, Pharmacol. Toxicol. 2002,91,
204.
[7] H. Tinwell, P. A. Lefevre, G. J. Moffat,
A. Burns, J. Odum, T. D. Spurway, G.
Orphanides, J. Ashby, Environ Health
Perspect. 2002,110, 533.
[8] R. Schreurs, P. Lauser, W. Seinen, B. van
den Burg, Arch.Toxicol. 2002,76, 257.
[9] M. Inui, T. Adachi, S. Ta kenaka, H. Inui,
M. Nakazawa, M. Ueda, H. Wa tanabe, C.
Mori, T. Iguchi, K. Miyatake, Toxicology
2003,194, 43.
[10] S. O. Mueller, M. Kling, P. A. Firzani, A.
Mecky, E. Duranti, J. Shields-Botella, R.
Delansorne, T. Broschard, P. J. Kramer,
Toxicol. Lett. 2003,142, 89.
[11] H. Jarry, J. Christoffel, G. Rimoldi, L. Koch,
W. Wuttke, Toxicology 2004,205, 87.
Table 4. UV fi lters in human milka
Percentage of
women using product
with compound
n = 34
% of Total
Percentage of
milk samples with
compound
n = 34
% of Total
Levels in Human Milk
Mean
± SD, number of positive
samples
ng/g lipid
Median
ng/g lipid
Range in positive samples
ng/g lipid
UV fi lters
EHMC / OMCb
Octocrylene
Bp-3
4-MBC
OD-PABA
HMS
Bp-2
3-BCc
58.8
38.2
14.7
17.7
2.94
14.7
11.8
0.0
64.7
47.1
18.2
11.8
2.94
0
0
0
28.9 ± 22.1
(22)
18.3 ± 17.97
(16)
49.2 ± 54.9
(6)
15.6 ± 5.9
(4)
50.0 ± 0
(1)
n.d.
n.d.
n.d.
25.0
12.5
19.8
18.4
50.0
n.d.
n.d.
n.d.
2.1 - 78.1
4.7 - 77.5
7.3 - 121.4
6.7 - 19.0
50.0
n.d.
n.d.
n.d.
UV Filters in sunscreens 45.5
UV fi lters in different
cosmetics
60.6
% of women using any
product with UV Filters
% of milk samples
containing any of the UV
fi lters
78.8 76.5
a Combined data from Pilot Study (2004, n = 13) and Study 1 (2005, n = 21). Study 2 (2006, n = 20) not yet incorporated.
b Abbreviations: EHMC = ethylhexylmethoxy cinnamate = OMC = octylmethoxy cinnamate, 4-MBC = 4-methylbenzylidene camphor, 3-BC = 3
benzylidene camphor, Bp-3 = benzophenone-3, Bp-2 = benzophenone-2, HMS = homosalate, OD-PABA = octyldimethylamino benzoic acid.
c Authorised for use in cosmetics in 2006.

ENDOCRINE DISRUPTORS: CHEMICALS OF CONCERN 351
CHIMIA 2008,62, No. 5
[12] M. Schlumpf, H. Jarry, W. Wuttke, R. Ma,
W. Lichtensteiger, Toxicology 2004,199,
109.
[13] H. Klammer, C. Schlecht, W. Wuttke, H.
Jarry,Toxicology 2005,215, 90.
[14] P. Y. Kunz, H. F. Galicia, K. Fent, Toxicol.
Sci. 2006,90, 349.
[15] R. Ma, B. Cotton, W. Lichtensteiger, M.
Schlumpf, Toxicol. Sci. 2003,74, 43.
[16] R. H. Schreurs, E. Sonneveld, J. H. Jansen,
W. Seinen, B. van der Burg, Toxicol. Sci.
2005,83, 264.
[17] M. Schlumpf, P. Schmid, S. Durrer, M.
Conscience, K. Maerkel, M. Henseler, M.
Gruetter,I. Herzog,S. Reolon, R.Ceccatelli,
O. Faass, E. Stutz, W. Lichtensteiger,
Toxicology 2004,205, 113.
[18] C. Schmutzler, I. Hamann, P. J. Hoffmann,
G. Kovacs, L. Stemmler, B. Mentrup, L.
Schomburg, P. Ambrugger, A. Grüters, D.
Seidlowa-Wuttke, H. Jarry, W. Wuttke, J.
Köhrle, Toxicology 2004,205, 95.
[19] K. Maerkel, S. Durrer, M. Henseler, M.
Schlumpf, W. Lichtensteiger, Toxicol.
Appl. Pharmacol. 2007,218, 152.
[20] W. Völkel, T. Colnot, U. M. T. Schauer, T.
H. Broschard, W. Dekant, Toxicol. Appl.
Pharmacol. 2006,216, 331.
[21] Scientific Committee on Consumer
Products (SCCP), Brussels, Opinion on
4-methylbenzylidene camphor (Colipa
no. S60), adopted by the SCCP during the
9th plenary meeting of 10 October 2006
(SCCP/1042/06).
[22] D. Seidlova-Wuttke, J. Christoffel, G.
Rimoldi, H. Jarry, W. Wuttke, Toxicol.
Appl. Pharmacol. 2006,210, 1.
[23] D. Seidlova-Wuttke, J. Christoffel, G.
Rimoldi, H. Jarry, W. Wuttke, Toxicol.
Appl. Pharmacol. 2006,210, 246.
[24] D.Seidlova-Wuttke,H. Jarry,J.Christoffel,
G. Rimoldi, W. Wuttke, To xicology 2004,
205, 103.
[25] D.Seidlova-Wuttke,H. Jarry,J.Christoffel,
G. Rimoldi, W. Wuttke, To xicology 2005,
213, 13.
[26] N. J. MacLusky, F. Naftolin, Science 1981,
211, 1294.
[27] M. E. Lauber, W. Lichtensteiger,
Endocrinology 1994,135, 1661.
[28] J. Bakker, S. Honda, N. Harada, J.
Balthazart, J. Neurosc. 2002,22, 9104.
[29] F. W. George, J. Devl. Physiol. 1993,19,
187.
[30] V. Pezzi, J. M. Mathis, W. E. Rainey, B. R
Carr,J. Steroid Biochem. Mol. Biol.2003,
87, 181.
[31] R. Habert, R. Picon, J. Steroid. Biochem.
1984,21, 193.
[32] T. M. Siler-Khodr, in ‘Fetal and Neonatal
Physiology’, Vol. 1, Ed. R. A. Polin, W.
W. Fox, W. B. Saunders Co., Philadelphia,
1998, p. 89.
[33] S. Durrer, K. Maerkel, M. Schlumpf, W.
Lichtensteiger,Endocrinology 2005,146,
2130.
[34] S. Durrer, C. Ehnes, M. Fuetsch, K.
Maerkel, M. Schlumpf, W. Lichtensteiger,
Environ. Health Perspect. 2007,115,
Suppl. 1 , 42.
[35] K. Maerkel, W. Lichtensteiger, S. Durrer,
M. Conscience, M. Schlumpf, Environ.
Toxicol. Pharmacol. 2005,19, 761.
[36] U. M. D. Schauer, W. Völkel,A. Heusener,
T. Colnot, T. H. Broschard, F. von
Landenberg, W. Dekant, Toxicol. Appl.
Pharmacol. 2006,216, 339.
[37] I. L. Wa rd, Science 1972,175, 82.
[38] M. Schlumpf, L. Berger, B. Cotton, M.
Conscience-Egli, S. Durrer, I. Fleisch-
mann, V. Haller, K. Maerkel, W. Lichten-
steiger, SFÖW-Journal 2001,127, 10.
[39] L. B. Biegel, J. A. Flaws,A. N. Hirshfield,
J. C. O’Connor, G. S. Elliott, G. S. Ladics,
E. K. Silbergeld, C. S. Van Pelt, M. E.
Hurtt, J. C. Cook, S. R. Frame, Toxicol.
Sci. 1998,44, 116.
[40] L.E. Hofkamp, S. Bradley, M. Schlumpf,
B.G. Timms, Posters on the Hill (POH),
2007,Council on UndergraduateResearch,
Washington, DC.
[41] N.Atanassova, C. McKinnell, K. J. Turner,
M. Walker, J. S. Fisher, M. Morley, M.
R. Millar, N. P. Groome, R. M. Sharpe,
Endocrinology 2000,141, 3898.
[42] O. Putz, C. B. Schwartz, G. A. LeBlanc,
R. L. Cooper, G. S. Prins, Biol. Reprod.
2001,65, 1506.
[43] F. S. vom Saal, B. G. Timms, M. M.
Montano, P. Palanza, K. A. Thayer, S.
C. Nagel, M. D. Dhar, V. K. Ganjam, S.
Parmigiani, W. V. Whelshons, Proc. Natl.
Acad. Sci. USA 1997,94, 2056.
[44] S. E. Gans, M. K. McClintock, Horm.
Behav.1993,27, 403.
[45] W. Lichtensteiger, O. Faass, R. Ceccatelli,
M.Schlumpf, OrganohalogenCompounds
2004,66, 3965.
[46] G. Pollio, P. Xue, M. Zanisi, A. Nicolin,
A. Maggi, Mol. Brain Res. 1993,19, 135.
[47] S. Ogawa, U. E. Olazabal, I. S. Parhar, D.
W. Pfaff, J. Neurosci. 1994,14, 1766.
[48] T. Kupper, C. Plagellat, R. C. Brändli,
L. F. de Alencastro, D. Grandjean, J.
Tarradellas, Wa ter Res. 2006,40, 2603.
[49] M. E. Balmer, H. R. Buser, M. D. Müller,
T. Poiger, Environ. Sci. Technol. 2005,39,
953.
[50] H. R. Buser, M. Balmer, P. Schmid, M.
Kohler, Environ. Sci. Technol. 2006,40,
1427.
[51] H. Nakata, H. Sasaki, A. Takemura, M.
Yoshioka, S. Tanabe, K. Kannan, Environ.
Sci. Technol. 2007,41, 2216.
[52] P. Schmid, M. Kohler, E. Gujer, M.
Zennegg, M. Lanfranchi, Chemosphere
2007,67, S16.
[53] N. R. Janjua, B. Mogensen, A. M.
Andersson, J. H. Petersen, M. Henriksen,
N. E. Skakkebaek, H. C. Wulf, J. Invest.
Dermatol. 2004,123, 57.
[54] S. H. Swan, K. M. Main, F. Liu, S. L.
Stewart, R. L. Kruse, A. M. Calafat, C.
S. Mao, J. B. Redmon, C. L. Terrand, S.
Sullivan, J. L. Teague, and the Study for
Future Families Research Team, Environ.
Health Perspect. 2005,113, 1056.
[55] K. M. Main, G. K. Mortensen, M. M.
Kaleve, K. A. Boisen, I. M. Damgaard, M.
Chellakooty, I. M. Schmidt, A. M. Suomi,
H. E. Virtanen, J. H. Petersen, A. M.
Andersson, J. Toppari, N. E. Skakkebaek,
Environ. Health Perspect. 2006,114, 270.
[56] S. Wünschmann, I. Kühn, H.
Heidenreich, S. Fränzle, O. Wappelhorst,
B. Markert‚‘Transfer von Elementen
in die Muttermilch’, Schriftenreihe
Reaktorsicherheit und Strahlenschutz,
Bundesministerium für Umwelt,
Naturschutz und Reaktorsicherheit, Bonn,
2005.
[57] European standardized methods for
pesticide residue analysis: fatty food –
determination of pesticides and PCBs, EN
1528 part 1-4, 1996-10, confirmed, 2001.
[58] T. Soeborg, N. C. Ganderu, J. H.
Kristensen, P. Bjerregaard, K. L. Pedersen,
P. Bollen, S. H. Hansen, B. H. Soerensen,
J. Chromatogr. B2006,834, 117.