Food Chemistry 136 (2013) 1590–1596

Contents lists available at SciVerse ScienceDirect

Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem

Endocrine disruptor activity in bottled mineral and flavoured water

Monika Plotan 1, Caroline Frizzell 1, Victoria Robinson, Christopher T. Elliott, Lisa Connolly ⇑
Institute of Agri-Food and Land Use, School of Biological Sciences, Queen’s University Belfast, Belfast BT95AG, Northern Ireland, United Kingdom

a r t i c l e i n f o a b s t r a c t

Article history: A panel of reporter gene assays (RGAs) coupled with a single solid phase extraction (SPE) step was devel-
Available online 14 February 2012 oped and used to screen bottled mineral water for the presence of four classes of endocrine disruptors
(EDs), oestrogens, androgens, progestagens and glucocorticoids.
Keywords: Fourteen brands of bottled mineral water in triplicate (42 samples) were analysed. Overall, hormonal
Oestrogens activity was found in 78% of the samples. Oestrogenic, androgenic, progestagenic and glucocorticoid
Androgen activity was found in 38%, 38%, 36% and 55% of the samples, respectively at an average concentration
Progestagen
of 10 ng/l 17b-estradiol equivalent (EEQ), 26 ng/l testosterone equivalent (TEQ), 123 ng/l progesterone
Glucocorticoid
Reporter gene assay
equivalent (PEQ) and 13.5 ng/l hydrocortisone equivalent (HEQ).
Bioassay The level of oestrogenic, androgenic and progestagenic activity observed is not considered a matter of
Food safety concern for the consumers’ health. It is unknown whether the glucocorticoid levels observed are safe. The
Drinking water ED source, long term exposure and mixture effects remain to be investigated.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction samples showed oestrogenic activity up to 23.1 ng/l 17b-estradiol

An endocrine disruptor (ED) is ‘‘an exogenous substance or
mixture that alters function(s) of the endocrine system and conse-
quently causes adverse health effects in an intact organism, or its
progeny, or (sub) populations’’ (OECD, 2010). EDs can interact with
the endocrine system by mimicking or blocking natural hormones
or their pathways, EDs may also influence the production, secre-
tion or metabolism of endogenous hormones or/and their nuclear
receptors (Tabb & Blumberg, 2006). Consequently, exposure to
EDs may lead to detrimental health effects including interference
with both the male and female reproductive systems, causing a
spectrum of disorders throughout life, including sexual precocity,
hormone related cancers e.g. testicular and breast cancer, repro-
ductive tract abnormalities and infertility (Diamanti-Kandarakis
et al., 2009). Other impacts of EDs may include effects on thyroid
function, obesity and metabolism (Diamanti-Kandarakis et al.,
2009).
It is now widely accepted that one of the major routes of ED
exposure for humans is through the diet (Connolly, 2009). Conse-
quently, with the consumption of bottled mineral water increasing
worldwide (Doria, 2006) the presence of oestrogenic EDs in partic-
ular has become more widely investigated (Pinto & Reali, 2009;
Wagner & Oehlmann, 2009; Wagner & Oehlmann, 2010). Pinto
and Reali (2009) used an in vitro yeast bioassay to screen PET
bottled mineral water and reported that 10% of the surveyed
⇑ Corresponding author. Tel.: +44 9097 6668; fax: +44 9097 6513.
E-mail address: l.connolly@qub.ac.uk (L. Connolly).
1
These authors contributed equally.
equivalent (EEQ). Wagner and Oehlmann (2009) also used an
in vitro yeast bioassay to provide evidence of oestrogenicity in glass
and PET bottled mineral water with 60% of samples surveyed
showing oestrogenic activity up to a maximum of 75 ng/l EEQ.
They also reported that this contamination partly originates from
compounds leaching from the PET bottles into mineral water as
indicated by a molluscan model, whereby mud snails (Potamopyr-
gus antipodarum) that are highly sensitive to oestrogens were
raised in glass and PET bottles and reproductive output measured.
The results of the in vitro yeast based assay confirmed the oestro-
genic results of the in vivo molluscan model in all except one PET
bottled sample which showed a minimal response in the yeast
assay but induced a more significant response in the mud snail
experiment. Therefore, this result may imply that the in vivo assay
is more sensitive than the in vitro assay, or that bottled water con-
tains a mixture of EDs, potentially of other classes. For example, if
EDs such as anti-androgens are present, they will remain unde-
tected by the oestrogen receptor specific in vitro assay, but may
still have effects on reproduction (Tillmann, Schulte-Oehlmann,
Duft, Markert, & Oehlmann, 2001). Further confirmation of the
presence of oestrogenic compounds in bottled mineral water has
again been reported by Wagner and Oehlmann (2010) using an
improved extraction procedure and the E Screen bioassay.
Three potential sources of oestrogenic ED contamination have
been outlined by Wagner and Oehlmann (2009); contaminated
source water, contamination with plastic components or deter-
gents used in the production process or substances migrating from
packaging materials. EDs found in the environment and having the
potential to be present in resources used for drinking water

0308-8146/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodchem.2012.01.115
 M. Plotan et al. / Food Chemistry 136 (2013) 1590–1596
production include: natural hormones and their metabolites, syn-
thetic hormones, phyto- and mycooestrogens, drugs with hor-
monal side effects, heavy metals, pesticides and their
metabolites, and industrial and household chemicals and their
by-products (Burkhardt-Holm, 2010). Most research so far has
focused on the detection of compounds in bottled water that bind
to the oestrogen receptor, however, ligands for other hormone
receptors may also be present. Multiple ED classes have previously
been shown to exist in wastewater effluents and surface waters
using a panel of reporter gene cell lines (Van der Linden et al.,
2008). This study showed that androgen, progestagen and gluco-
corticoid hormonal activities were present at similar or much
higher levels than that of oestrogens. The identity of many of these
substances have yet to be established. Cleaning reagents with ED
activity may originate in the production process and have been
observed in bottled water previously (Diana & Dimitra, 2011).
The widespread use of compounds such as parabens and perfluoro-
nited compounds (PFCs) has lead to their detection in drinking
water resources, lower levels of PFCs have been observed in bottled
water (Richardson and Ternes, 2005). Low levels of bisphenol A has
recently been reported in bottled mineral water, with bottle clo-
sures, the water itself or recycled PET being suggested as possible
sources of contamination (Bach, Dauchy, Chagnon, & Etienne,
2012). Phthalates also have the potential to leach from plastics
(Casajuana & Lacorte, 2003) and individual phthalates such as
DBP and butyl benzyl phthalate (BBP) have been reported as
weakly oestrogenic (Jobling, Reynolds, White, Parker, & Sumpter,
1995), while others such as the plasticiser DEHP has been shown
to be non-oestrogenic. However, all of these phthalates have been
shown to have anti-androgenic effects (Parks et al., 2000; Sohoni &
Sumpter, 1998; Takeuchi et al., 2005). Therefore not only oestro-
genic EDs but a range of ED classes may be present in bottled min-
eral water, this remains to be investigated.
The present study aimed to investigate the presence and activ-
ity profile of a broader range of EDs in bottled mineral water sam-
ples by combining a suitable sample pre-treatment step prior to
analysis in oestrogen, androgen, progestagen and glucocorticoid
responsive RGAs.
2. Material and methods
2.1. Chemicals and reagents
All cell culture reagents were supplied by Invitrogen Ltd.
(Paisley, UK). Reference standards including 17b-estradiol, testos-
terone, progesterone and hydrocortisone, HPLC grade water and
tert-Butyl-methyl-ether were obtained from Sigma (Poole, Dorset,
UK). Falcon tissue culture flasks were supplied by BD Biosciences
(Oxford, UK). Specialised RGA plates were supplied by Greiner
Bio-One (Stonehouse, UK). A luciferase kit (Promega, E1501) was
obtained from Promega (Southhampton, UK). Oasis HLB glass
cartridges (5 cc/200 mg) were supplied by Waters Chromatogra-
phy Ireland (Dublin, Ireland). Methanol and acetone were obtained
from BDH (Poole, Dorset, UK).
2.2. Reporter gene assay analysis
Oestrogen (MMV-Luc), androgen (TARM-Luc), progestagen
(TM-Luc) and glucocorticoid (TGRM-Luc) responsive mammalian
reporter cell lines were previously produced by stable transfection
with the MAR-Vit-Luc vector into the MCF-7 cell line in the case of
oestrogen, and with the MMTV-Luc vector into the T47D cell line
for each of the other three hormone specific cell lines (Willemsen,
Scippo, Maghuin-Rogister, Martial, & Muller, 2002; Willemsen
et al., 2004). Additionally for the androgen cell line the pSV-AR0
1591
expression vector was inserted, as the T47D cell line does not con-
tain the androgen receptor gene. The RGA procedure was followed
as described in a previous study by Plotan et al. (2011). The
progestagen and glucocorticoid responsive RGAs were performed
exactly as for the androgen assay using their corresponding stan-
dards. 17b-Estradiol, testosterone, progesterone and hydrocorti-
sone were applied as the standards for the oestrogen, androgen,
progestagen and glucocorticoid RGAs, respectively.
2.3. Cytotoxicity testing
All extracts produced for validation of the assay and the sam-
ples used in the screening study were tested for cytotoxicity on
all four cell lines, MMV-Luc, TARM-Luc, TM-Luc and TGRM-Luc.
The thiazolyl blue tetrazolium bromide (MTT) based cytotoxicity
assay was used to check the toxicity of the applied samples, the
protocol used was as described previously by Frizzell et al. (2011).
2.4. Oasis HLB cartridge SPE
Solid phase extraction (SPE) was used for the extraction of EDs
from the bottled mineral water samples. Waters Oasis HLB glass
columns (5cc/200 mg) were used. The extraction procedure in-
cluded conditioning with 3 ml of tert-Butyl methyl ether, methanol
and HPLC grade water; loading of the sample (10, 20, 50 and
250 ml); washing with 3 ml of 5% methanol in water and elution
with 6 ml of 10% methanol in tert-Butyl methyl ether. Eluent was
evaporated under a nitrogen stream at 40 °C and resuspended in
50 ll methanol.
2.5. Extraction recoveries
Pure HPLC grade water (250 ml) was spiked with 17b-estradiol,
testosterone, progesterone and hydrocortisone at 0.5, 2.9, 5 and
100 ng/ml, respectively, and extracted using the SPE procedure
described in Section 2.4. The recoveries were determined based
on the RGA results for these extracts which were compared with
the spiked concentrations results for each of the standards. The
extraction recoveries for each RGA was determined from seven
independent experiments. All concentrations (ng/ml) presented
in the manuscript describe final well concentrations.
2.6. Influence of sample volume on the HLB extraction
In order to investigate the influence of the sample volume on
the HLB extraction and subsequent RGA results, a series of volumes
of HPLC grade water was extracted and applied to the androgen
responsive cell line. Two samples of each of four volumes of water
were measured: 10, 20, 50 and 250 ml. One was left unspiked and
the other was spiked with 29 ng testosterone (to give a 2.9 ng/ml
final well concentration in each case), both were then extracted
and applied to the androgen RGA. The influence of the sample vol-
ume was measured based on the extraction recoveries determined
(Section 2.5) and this was investigated in duplicate.
2.7. Repeatability study
Repeatability of the oestrogen, androgen, progestagen and glu-
cocorticoid RGAs was investigated by triplicate analysis of (un)-
spiked HPLC grade water, this was performed in a single
experiment. Three water samples (250 ml) were left unspiked
and the other three water samples of the same 250 ml volume
were spiked with one of the four hormone standards at the same
concentrations: 17b-estradiol (0.5 ng/ml), testosterone (2.9 ng/
ml), progesterone (5 ng/ml) and hydrocortisone (100 ng/ml). These
1592 M. Plotan et al. / Food Chemistry 136 (2013) 1590–1596

water samples were extracted (SPE method, Section 2.4) and ap-
plied to their corresponding RGAs.
2.8. Reproducibility study
In order to determine the reproducibility of the method devel-
oped, the same study as described in the previous section (repeat-
ability study, Section 2.7) was performed over three independent
experiments within a two week period. Each experiment involved
two 250 ml HPLC grade water samples, one left unspiked and the
other was spiked with one of the four steroid hormones (17b-estra-
diol, testosterone, progesterone and hydrocortisone at 0.5, 2.9, 5
and 100 ng/ml, respectively). Each set of samples was extracted
and tested by their corresponding RGAs on the same day.
The water samples EEQ, testosterone equivalent (TEQ), proges-
terone equivalent (PEQ) or hydrocortisone equivalent (HEQ) were
determined based on the samples % activity (% activity = (sample
RLU negative control RLU)/(positive control RLU negative con-
trol RLU)  100%), hormone spiking concentration and dilution fac-
tor; expressed in ng/ml or ng/l.
Recoveries were determined as a ratio of the calculated concen-
tration of the spiked extract (EEQ, TEQ, PEQ, HEQ) and the value of
the spiking concentration before extraction, this was multiplied by
100 to achieve a % extraction recovery.
3. Results and discussion
3.1. Standard curves and RGA sensitivity

2.9. Bottled mineral water samples
For the purpose of the study a range of bottled mineral water
was purchased randomly from various sources in the region of
Belfast (Northern Ireland, UK): supermarkets, off-sales, health
shops, restaurants and pharmacies. Fourteen brands of bottled
mineral waters (samples coded 1–14) were purchased in triplicate,
giving a total of 42 samples. Half of the samples were UK products
from England (5), Wales (1) and Scotland (1). The remaining prod-
ucts were imported from Italy (3), Ireland (2) and France (2). All of
the water samples were bottled in either plastic (samples 1–10) or
glass (samples 11–14) packaging. One sample was carbonated,
sample number 11. Three samples were flavoured waters (samples
8–10), containing sweetener (sucralose) and various flavourings.
These three samples were also promoted to enhance wellbeing,
having additives such as L-carnitine, green tea extract, blueberry
or pomegranate. Each of the triplicate samples possessed different
batch numbers except for the flavoured waters (samples 8–10) and
samples 3, 5 and 12, all of which came from the same batch. All 42
samples were kept according to the manufacturers guidelines:
sealed and stored in a dry place and out of direct sunlight, until
testing commenced.
Each sample was tested for agonist activity in all four RGAs and
additionally in the androgen RGA for antagonist activity. The
250 ml volume of each of the unspiked water samples were ex-
tracted and applied to all four RGAs and MTT assays. Similarly an-
other 250 ml of each water sample was spiked with testosterone to
give a final concentration of 2.9 ng/ml that the cells were exposed
to, this was extracted and applied to the androgen responsive
assays only. For each experiment a set of negative and positive con-
trols (17b-estradiol, testosterone, progesterone and hydrocortisone
at 0.5, 2.9, 5 and 100 ng/ml, respectively) was extracted and
applied to both the RGAs and MTT assay. All the experimental
points were tested in triplicate and in three independent
experiments.
A procedure was developed to detect the presence of oestrogen-
ic, androgenic, progestagenic and glucocorticoid compounds in 14
brands of bottled mineral water. Standard curves were established
using 17b-estradiol, testosterone, progesterone and hydrocortisone
for the oestrogen, androgen, progestagen and glucocorticoid RGAs,
respectively. The sensitivity of each RGA can be compared by the
EC50 values which were 0.006 ng/ml, 0.8, 0.3 and 53.5 ng/ml for
the oestrogen, androgen, progestagen and glucocorticoid RGAs,
respectively.
3.2. SPE recoveries
The percentage recovery was determined for each hormone. The
highest recoveries were obtained for 17b-estradiol (94 ± 6%) and
hydrocortisone (89 ± 14%) followed by testosterone (86 ± 7%) and
progesterone (80 ± 4%).
3.3. Influence of sample volume on SPE
Four sample volumes (ranging between 10–250 ml as recom-
mended by the manufacturer) were assessed for use with the Oasis
HLB column for the detection of low levels of EDs in the bottled min-
eral water. Results for all 4 sample volumes investigated are pre-
sented in Fig 1 and show that levels of recovery were similar in
each volume (ranging between 73% and 90%). Therefore it was

2.10. Data analysis

Standard curves were fitted based on the sigmoidal dose-

responsive equation, using SlideWrite Plus V6 software, where Y
is the response, X is the logarithm of concentration, while bottom
is 0 and top is 100 percent of the maximum achieved response:
Y bottom + (top bottom)/(1 + 10(logEC50 X)).
Fold induction was measured by calculating the ratio of the re-
sponse when compared with the negative control (n-fold). The
concentration that produced a 50% increase in maximal response
(EC50) was used as a measurement of the assays sensitivity.
Coefficient of variation (% CV) was calculated as the ratio of the
standard deviation (STDV) and average of sample repetitions, pre- Fig. 1. Four HPLC grade water volumes were spiked with 29 ng testosterone (final
sented in %. well concentration of 2.9 ng/ml), extracted and tested in the androgen RGA.
 M. Plotan et al. / Food Chemistry 136 (2013) 1590–1596 1593

decided to use a volume of 250 ml in sample analysis to
maximise the detection of EDs that may be present in very low
concentrations.
3.4. Repeatability
The repeatability of the analysis of EDs in bottled mineral water
samples for all four RGAs was investigated. Three unspiked and
three spiked (with a respective standard steroid) 250 ml HPLC
grade water samples were extracted and analysed in their appro-
priate RGA. Very good repeatability was observed as shown in
Table 1 and confirmed by low coefficients of variation (CV). Per-
centage CV (% CV) values for the negative controls were lowest in
the progestagen and oestrogen assays, 10% and 14%, respectively.
A higher % CV was observed for the glucocorticoid assay (17%)
and the highest for the androgen assay (24%). The positive controls
presented lower % CV values than the negative controls, ranging
between 2% and 3% for the androgen and oestrogen assays, and
up to 11–13% for the glucocorticoid and progestagen assays,
respectively.
3.5. Reproducibility
A reproducibility study was performed by repeating the same
study outlined in Section 3.4 in three independent experiments.
The reproducibility results showed a higher degree of variation
than those obtained in the repeatability study (Table 1). Variability
(% CV) for the negative controls was lower than those obtained for
the positive controls, in the range of 9–30%. The lowest deviation
was observed in the glucocorticoid and oestrogen assay (9% and
13%, respectively) and the highest observed in the androgen and
progestagen assay (23% and 30%, respectively). Spiked samples
presented the lowest % CV in the oestrogen assay (19%), higher in
the glucocorticoid and androgen assay (33% and 40%, respectively)
and to the highest in the progestagen assay (83%). ‘Day 1’ results
presented much higher fold induction values than the other days
in almost all the cell lines, except for the oestrogen assay positive
control which showed a low % CV. These higher RLU readings may
be due to the use of a different batch of luciferase substrate. If the
results from ‘Day 1’ were removed from the calculations, the % CV
values for positive controls would be in a significantly lower range
of 3–24%.
3.6. Bottled mineral water results
All fourteen brands of bottled mineral water in triplicate (42
samples) were screened for agonistic activity in the oestrogen,
androgen, progestagen and glucocorticoid assays. Antagonist activ-
ity was also screened in the androgen assay due to the anticipation
that some androgen antagonists e.g. pesticides, phthalates may be
present. The overall ED activity profile of the 14 brands of bottled
mineral water (42 samples) is presented in Table 2. These results
are discussed separately for each ED class.
3.6.1. Oestrogens
Oestrogenic agonist activity was detected in 9 brands of bottled
mineral water as shown in Table 2; including two brands of water
that presented agonist activity in all 3 triplicate samples (brand
numbers 8 and 10). The samples for each of these brands possessed
identical batch numbers and were flavoured waters, containing
sweetener and various natural flavourings and also promoted as
enhancing wellbeing. Some of these additives may be responsible
for the oestrogenic activity observed. Of the 42 samples, 38% pre-
sented an EEQ ranging between 1–34 ng/l with a mean of 10 ng/l
(Table 3). In comparison, the study performed by Wagner and
Oehlmann (2009) reported that 60% of the 20 brands of bottled
mineral waters tested contained oestrogenic contamination with
mean EEQ activity levels of 18 ng/l. In a more recent study, (Wag-
ner & Oehlmann, 2010), used different sample preparation meth-
ods and detected an oestrogenic response in 61% of 18 samples
analysed, corresponding to 1.9–12.2 pg/l EEQ. Another study by
Pinto and Reali (2009) reported that 10% of samples surveyed
showed oestrogenic activity up to 23.1 ng/l EEQ. Within the pres-
ent study, one brand of bottled mineral water (brand number
12), packaged in glass, showed oestrogenic activity in two of the
triplicate samples with the same batch number. This observation
is consistent with that of Wagner and Oehlmann (2009, 2010)
where some water samples packaged in glass bottles also exhibited
oestrogenic activity.
3.6.2. Androgens
The screening for androgenic EDs included both agonistic and
antagonistic activity. Androgen agonists were detected in 11
brands and androgen antagonists in 6 brands of bottled mineral
water as shown in Table 2. Antagonist activity was found in all
three replicates of brands 8, 9 and 10, the flavoured water samples.
Of the 42 samples, 38% presented TEQ agonist activity at concen-
trations ranging between 3 and 63 ng/l and 29% presented TEQ
antagonist activity at concentrations ranging between 13 and
84 ng/l (Table 3). Interestingly, in many samples that displayed
antagonism, agonism was simultaneously observed (Table 2).

Table 1
Repeatability and reproducibility of the RGAs presented as fold inductions of the relevant hormones spiked in HPLC grade water.

Repeatability
Negative control Positive control
Rep. 1 Rep. 2 Rep. 3 Average STDEV % CV Rep. 1 Rep. 2 Rep. 3 Average STDEV % CV
Oestrogen 1.3 1.1 1 1.1 0.15 13.5 4.4 4.4 4.2 4.3 0.1 2.7
Androgen 0.8 0.5 0.6 0.6 0.15 24.1 7.7 7.4 7.7 7.6 0.2 2.3
Progestagen 1 0.9 1.1 1.0 0.1 10 10.9 10.7 8.6 10.1 1.3 12.7
Glucocorticoid 1.2 1 1.4 1.2 0.2 16.7 74.5 92 89.4 85.3 9.4 11.1
Reproducibility
Negative control Positive control
Day 1 Day 2 Day 3 Average STDEV % CV Day 1 Day 2 Day 3 Average STDEV % CV
Oestrogen 1.3 1.0 1.2 1.2 0.15 13.1 3.8 4.5 5.5 4.6 0.9 18.6
Androgen 0.8 0.5 0.7 0.7 0.15 22.9 14.3 7.7 7.4 9.8 3.9 39.8
Progestagen 1.8 1.1 1.1 1.3 0.4 30.3 38 10.7 9.4 19.4 16.1 83.4
Glucocorticoid 1.4 1.2 1.2 1.3 0.12 9.1 124 63 89.4 92.1 30.5 33.2

Abbreviation: Rep. – repetition.

1594 M. Plotan et al. / Food Chemistry 136 (2013) 1590–1596

Table 2
Induction levels of ED agonists and antagonists detected in triplicate samples of 14 types of bottled mineral water as tested by RGA. Detected activity was equivalent to reference
hormone, which depending on the cell line was 17b-estradiol at 0.5 ng/ml, testosterone at 2.9 ng/ml, progesterone at 5 ng/ml, hydrocortisone at 100 ng/ml, for the oestrogen,
androgen, progestagen and glucocorticoid RGAs, respectively.

Cell line Detected Positive control [fold Bottled mineral water (samples 1–14, n = 3)
activity induction]
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Oestrogen Agonist 3.4 jss sss sss dss dds dss dss djj dds ddd sss dds sss sss
Androgen Agonist 13 dss sss jjs djs sss jjs Nss dss Nss dss sss jjs jjs dss
Antagonist sss sss sss sss sss dss dss djj jNN NNN sss sss dss sss
Progestagen Agonist 7 dss sss dss djs dss jss Nss sss Nss dss dss jss djs dds
Glucocorticoid Agonist 376 djs jss dss ddd djs ddN jNs sss ddj dss jss dss dss djs

Symbols describe the fold induction for: (s) no activity; (d) agonist 1.3–2, antagonist 2–4; (j) agonist 2.1–3.5, antagonist 5–6; (N) agonist 4.5–7.5, antagonist 7–9.

3.6.3. Progestagens
Progestagen agonist activity was detected in 12 brands of bot-
tled mineral water (Table 2). Of the 42 samples, 36% showed pro-
gestagenic activity. The highest ED levels were detected by the
progestagen RGA, presenting a PEQ within the range of 11–
392 ng/l (Table 3).
3.6.4. Glucocorticoids
The largest number of samples presenting endocrine disrupting
activity was detected in the glucocorticoid RGA with more than
half of the samples (55%) inducing a positive response (Table 2). Al-
most all brands (13 out of 14) presented glucocorticoid activity
with HEQ activity in the range of 3–81 ng/l (Table 3).
3.7. Exposure to EDs via bottled mineral water
This study reports the presence of EDs with oestrogen, andro-
gen, progestagen and glucocorticoid activity in bottled mineral
water. It must be remembered that the assays used within this
and other studies (Wagner & Oehlmann, 2009, 2010; Pinto & Reali,
2009) measure all xenooestrogens or xenohormones present in the
samples. Therefore, hormonal activity observed may be due to
individual compounds (including as yet unknown compounds with
hormonal potency) or mixtures of various compounds. The identity
of these compounds remain to be confirmed.
Assuming that the average body weight (bw) of the bottled
mineral water consumer is 60 kg and the average daily intake of
mineral water is one 500 ml bottle, the exposure level to oestro-
genic EDs via bottled drinking water can be calculated as being
0.08 ng/kg bw/day (Table 3). This exposure level is lower than
the typical adult eating an omnivorous diet, which was estimated
to deliver 23 ng/person/day EEQ of endogenous oestrogens, an esti-
mate which the authors indicate is likely biased low as this does
not take into account the contribution of phytooestrogens or other
oestrogens which no occurrence or oestrogenic potency data is
available for (Caldwell et al., 2010). Caldwell et al. (2010) have also
indicated that for a young child, typical milk consumption (assumed
to be about 0.42 l/day), exposes them to oestrogen levels approxi-
mately 150 times greater than their exposure to trace concentra-
tions of oestrogens as predicted via drinking water intake. Indeed
it is products such as milk and milk products that supply about
60–80% of ingested female sex steroids (Hartmann, Lacorn, & Stein-
hart, 1998). Additionally, the acceptable daily intake (ADI) for 17b-
estradiol has been established by JECFA (2000) as up to 50 ng/kg
bw/day. Therefore, it may be concluded that the levels of EEQ activ-
ity being reported in bottled mineral water are safe for consumers.
Hartmann et al. (1998), also concluded that in comparison to the
human daily production of steroid hormones, the calculated nutri-
tional supply of oestrogens (100 ng/day), testosterone (50 ng/day)
and progesterone (10,000 ng/day) is not a significant source.
Exposure to androgen agonistic EDs via bottled drinking water
(500 ml) can be calculated as being 0.22 ng/kg bw/day and an inhi-
bition of 0.32 ng/kg bw/day for androgen antagonistic EDs (Ta-
ble 3). While there is no comparative data available in the
literature on the levels of androgens delivered in the typical
omnivorous diet, these exposure levels are well below the ADI
for testosterone established by JECFA (2000) as up to 2 lg/kg bw/
day. Therefore, it may be concluded that the levels of TEQ activity
being reported in bottled mineral water are safe for consumers.
The average daily intake of PEQ via a 500 ml bottle of the inves-
tigated products is approximately 1.24 ng/kg bw/day assuming
that the average body weight of the bottled mineral water con-
sumer is 60 kg. As established by JECFA (2000), the ADI of proges-
terone is up to 30 lg/kg bw/day, which is more than 20,000 times
higher in terms of hormone equivalent levels than that detected in
the present study. Therefore, the PEQ levels found in bottled min-
eral water indicate that they are safe for consumers.
The HEQ was in the range of 3–81 ng/l (Table 3). It should also
be noted that none of 42 tested samples (250 ml) exceeded a fold
induction of 7.5 (Table 3). In comparison, the positive control
hydrocortisone (100 ng/ml) presented a fold induction of 376.
Therefore, the samples tested presented extremely weak activity.
There is no ADI available for glucocorticoids and limited informa-
tion is available on glucocorticoid activity in water. Up to 16 and

Table 3
Summary of detected agonistic and antagonistic EDs in 42 bottled drinking water samples with predicted levels of 17b-estradiol, testosterone, progesterone and hydrocortisone
equivalents.

Oestrogen Androgen Progestagen Glucocorticoid

Agonist Agonist Antagonist Agonist Agonist
Number of samples showing hormonal activity 16 [38%] 16 [38%] 12 [29%] 15 [36%] 23 [55%]
Range of activity equivalent to 17b-estradiol, testosterone, progesterone or 1–34 ng/l 3–63 ng/l 13–84 ng/l 11–392 ng/l 3–81 ng/l
hydrocortisone
Average activity equivalent to 17b-estradiol, testosterone, progesterone or 10 ng/l 26 ng/l 38 ng/l 123 ng/l 13.5 ng/l
hydrocortisone
EDs intake per 500 ml bottle of drinking water equivalent to 17b-estradiol, 5 ng a 0.08 ng/ 13 ng a 0.22 ng/ 19 ng a 0.32 ng/ 62 ng a 1.24 ng/ 7 ng a 0.12 ng/
testosterone, progesterone or hydrocortisone kg bw/day kg bw/day kg bw/day kg bw/day kg bw/day
a
Assumptions: average body weight 60 kg, daily intake of mineral water is 1 bottle (500 ml).
 M. Plotan et al. / Food Chemistry 136 (2013) 1590–1596
2900 ng/l cortisol (hydrocortisone) equivalents have been detected
in surface water and waste water respectively but no activity was
detected in drinking water (Van der Linden et al., 2008).
3.8. ED sources
The possible migration of oestrogenic EDs from the PET bottle
into mineral water is a matter of some current debate. Franz and
Welle (2009) presented migration theories, based on the composi-
tional knowledge of plastic bottles and closures and concluded that
the migration of potential chemical compounds such as nonylphe-
nol, bisphenol A or DEHP cannot be the reason for the level of activ-
ity detected in mineral water by Wagner and Oehlmann (2009),
due to the fact that the oestrogenic potency of these compounds
is too low. However, Sax (2010) suggests that the levels of activity
observed may be contributed to the use of various grades and recy-
cled plastic that may contain previously banned EDs with greater
potency. Sax (2010) also suggests that the magnitude of leaching
may be influenced by storage time and temperature.
The presence of oestrogenic activity in bottled water packaged
in glass containers would suggest that plastic packaging may not
be the source or only source of contamination. The water source it-
self may be oestrogenic, bottle closures may present contamina-
tion or perhaps chemicals used in the production process,
bottling and cleaning may be a source of contamination. Also, the
lack of positive replication (Table 2) in identical batch number
samples (3, 5 and 12) may suggest that contamination during the
production process is responsible for the activity observed in indi-
vidual bottles. Samples number 8–10 included the flavoured fit-
ness waters and presented some of the best replication of results,
indicating that the additives may be responsible for the activity
observed.
The contamination of source waters is another possibility. EDs
may be present in waste waters leading to contamination of
ground waters, rivers, lakes and potentially drinking water re-
sources (López de Alda & Barceló, 2000). The natural mammalian
hormones oestrogens, androgens and progestagens are produced
every day in large quantities by humans and animals and excreted
into the environment. Progestagen excretion by farm animals has
been shown to exceed 45 and 10 times the amount of androgen
and oestrogen excretion respectively. Progesterone is more rapidly
degraded in the environment than oestrogens. However, natural
steroids, synthetic progestagens and their metabolites have been
detected in surface waters at concentrations up to several ng/ml
(Van der Linden et al., 2008). Oestrogenic and progestational chem-
icals are also used in the production of contraceptive pills which
have increased in use in recent years. Some of these, for example
the widely prescribed 17a-ethinylestradiol (EE2), are much more
potent than endogenous oestrogens and are consumed in high con-
centrations, for example 20 lg/day for EE2 (Falconer, 2006). Many
oestrogenic, androgenic and anti-androgenic EDs may also be
widespread in the environment due to their uses as insecticides,
herbicides, industrial chemicals, etc.
Glucocorticoids in the body are involved in many physiological,
cardiovascular, immunological, metabolic and homeostatic pro-
cesses. Next to the natural hormones, glucocorticoids are the most
commonly prescribed drugs worldwide because of their anti-
inflammatory and immunosuppressive action. Glucocorticoids
may be prescribed long-term for illnesses such as asthma (Guilbert
et al., 2006). These compounds are then excreted in urine and fae-
ces into the environment and resource waters. This human origin,
at least in part, has been suggested by the findings of Van der Lin-
den et al. (2008). Therefore, the source of water used for bottled
mineral water production may be the main, but not the only, pos-
sible way of contamination of the final product with EDs.
1595
4. Conclusions
Most of the available literature concerning the presence of EDs
in bottled drinking water is confined to oestrogens (Wagner &
Oehlmann, 2009, 2010; Pinto & Reali, 2009). This study reports
the presence of oestrogenic, androgenic, progestagenic and gluco-
corticoid active EDs in bottled mineral waters.
While hormonal activity was detected in 78% of samples, a com-
parison of the individual ED levels for oestrogens, androgens and
progestagens observed with their individual respective ADI sug-
gests that bottled mineral water consumption should not be of
concern for human health, even though the recommended con-
sumption of 2 l drinking water/day (WHO., 2008) is 3–4 times
higher than that considered in this study. While levels of glucocor-
ticoids detected in this study are quite low, there is no comparative
exposure data for glucocorticoids in the diet or drinking water.
Therefore it is unknown whether the levels observed in this study
are safe or not.
This study also highlights the need to investigate overall biolog-
ical activity, as multiple classes of hormonally active compounds
have been observed in the bottled mineral water product and their
synergistic effect remains to be fully investigated. Additionally,
although dietary exposure to hormonally active compounds may
be low via bottled mineral water, they can represent a consistent
daily exposure for many individuals who consume bottled mineral
water on a daily basis. Special attention to sensitive sub-popula-
tions may need to be considered where exposure to hormonally ac-
tive compounds in the diet is of concern for example, infants,
pregnant woman, prepubescent males, etc. Fetal exposure to EDs
is an area of particular concern. For example, the role of glucocor-
ticoids in the onset of adult diseases has been investigated (Drake,
Tang, & Nyirenda, 2007), with data suggesting that exposure of this
vulnerable group to excess glucocorticoids leads to programming
of cardiovascular, metabolic and neuroendocrine disorders in adult
life (Seckl, 2004). This is an area that warrants further investigation
due to the pivotal role that glucocorticoids play in human health.
Overall the levels of EDs observed in this study may be reassur-
ing. However, it is undesirable to have any level of EDs in our food
chain and efforts to prevent their presence should be made. The
use of in vitro bioassays has aided the study of EDs (Connolly, Rops-
tad, & Verhaegen, 2011) however, the identity of the responsible
compounds and source of contamination requires further investi-
gation. The purity of source waters and contamination during the
production process calls for monitoring. Previous studies on the
leaching of EDs from packaging focuses only on oestrogenic com-
pounds and popular PET bottles (Franz & Welle, 2009; Sax,
2010). The leaching of other classes of EDs such as anti-androgenic
phthalates should also be investigated. The influence of storage
time and conditions on the possible migration of hormonally active
compounds from packaging to the mineral water is also worthy of
investigation. Finally, the addition and stability of sweeteners, fla-
vourings and active compounds by the manufacturer needs to be
assessed for potential ED activity.
References
Bach, C., Dauchy, X., Chagnon, M.-C., & Etienne, S. (2012). Chemical compounds and
toxilogical assessments of drinking water stored in polyethylene terephthalate
(PET) bottles: A source of controversy reviewed. Water Research, 46, 571–583.
Burkhardt-Holm, P. (2010). Endocrine disruptors and water quality: A state-of-the-
art review. Water Resources Development, 26(3), 477–493.
Caldwell, D. J., Mastrocco, F., Nowak, E., Johnston, J., Yekel, H., Pfeiffer, D., Hoyt, M.,
DuPlessie, B. M., & Anderson, P. D. (2010). An assessment of potential exposure
and risk from estrogens in drinking water. Environmental Health Perspectives,
118(3), 338–344.
Casajuana, N., & Lacorte, S. (2003). Presence and release of phthalic esters and other
endocrine disrupting compounds in drinking water. Chromatographia, 57,
649–655.
1596 M. Plotan et al. / Food Chemistry 136 (2013) 1590–1596
Connolly, L. (2009). Endocrine disrupting chemicals: Origins, fates and transmission
into the food chain. In I. Shaw (Ed.), Endocrine Disrupting Chemicals in Food
(pp. 103–125). Cambridge, England: Woodhead Publishing Ltd..
Connolly, L., Ropstad, E., & Verhaegen, S. (2011). In vitro bioassays for the study of
endocrine-disrupting food additives and contaminants. Trends in Analytical
Chemistry, 30(2), 227–238.
Diamanti-Kandarakis, E., Bourguignon, J.-P., Giudice, L. C., Hauser, R., Prins, G. S.,
Soto, A. M., Zoeller, R. T., & Gore, A. C. (2009). Endocrine-disrupting chemicals:
An Endocrine Society Scientific Statement. Endocrine Reviews, 30(4), 293–342.
Diana, A., & Dimitra, V. (2011). Alkylphenols and phthalates in bottled waters.
Journal of Hazardous Materials, 185, 281–286.
Doria, M. (2006). Bottled water versus tap water: Understanding consumer’s
preference. Journal of Water Health, 4, 271–276.
Drake, A. J., Tang, J. I., & Nyirenda, M. J. (2007). Mechanisms underlying the role of
glucocorticoids in early life programming of adult disease. Clinical Science, 113,
219–232.
Falconer, I. R. (2006). Are endocrine disrupting compounds a health risk in drinking
water? International Journal of Environmental Research and Public Health, 3(2),
180–194.
Franz, R., & Welle, F. (2009). Can migration of endocrine disruptors from plastic
bottles be the cause of estrogenic burden recently determined in bottled
mineral water? Deutsche Lebensmittel-Rundschau, 105(5), 315–318.
Frizzell, C., Ndossi, D., Verhaegen, S., Dahl, E., Eriksen, G., Sorlie, M., Ropstad, E.,
Elliott, C. T., Muller, M., & Connolly, L. (2011). Endocrine disrupting effects of
zearalenone, alpha- and beta-zearalenol at the level of nuclear receptor binding
and steroidogenesis. Toxicology Letters, 206, 210–217.
Guilbert, T. W., Morgan, W. J., Zeiger, R. S., Mauger, D. T., Boehmer, S. J., Szefler, S. J.,
Bacharier, L. B., Lemanske, R. F., Strunk, R. C., Allen, D. B., Bloomberg, G. R., Heldt,
G., Krawiec, M., Larsen, G., Liu, A. H., Chinchilli, V. M., Sorkness, C. A., Taussig, L.
M., & Martinez, F. D. (2006). Long-term inhaled corticosteroids in preschool
children at high risk for asthma. The New England Journal of Medicine, 354(19),
1985–1997.
Hartmann, S., Lacorn, M., & Steinhart, H. (1998). Natural occurence of steroid
hormones in food. Food Chemistry, 62(1), 7–20.
JECFA (2000). Toxicological evaluation of certain veterinary drug residues in food.
Geneva, Switzerland. WHO food additives series: 43, IPCS – International
Programme on Chemical Safety.
Jobling, S., Reynolds, T., White, R., Parker, M. G., & Sumpter, J. P. (1995). A variety of
environmentally persistent chemicals, including some phthalate plasticizers are
weakly estrogenic. Environmental Health Perspectives, 105, 582–587.
López de Alda, M. J., & Barceló, D. (2000). Determination of steroid sex hormones
and related synthetic compounds considered as endocrine disrupters in water
by liquid chromatography-diode array detection-mass spectrometry. Journal of
Chromatography A, 892, 391–406.
OECD (2010). Guidence document on the assessment of chemicals for endocrine
disruption. Version 9.
Parks, L. G., Ostby, J. S., Lambright, C. R., Abbott, B. D., Klinefelter, G. R., Barlow, N. J.,
& Grey, L. E. Jr., (2000). The plasticizer diethylhexyl phthalate induces
malformations by decreasing fetal testosterone synthesis during sexual
differentiation in the male rat. Toxicological Sciences, 58, 339–349.
Pinto, B., & Reali, D. (2009). Screening of estrogen-like activity of mineral water
stored in PET bottles. International Journal of Hygiene and Environmental Health,
212, 228–232.
Plotan, M., Elliott, C. T., Scippo, M. L., Muller, M., Antignac, J. P., Malone, E., Bovee, T.
F. H., Mitchell, S., & Connolly, L. (2011). The application of reporter gene assays
for the detection of endocrine disruptors in sport supplements. Analytica
Chimica Acta, 700, 34–40.
Richardson, S. D., & Ternes, T. A. (2005). Water analysis: Emerging contaminants
and current issues. Analytical Chemistry, 77, 3807–3838.
Sax, L. (2010). Polyethylene terephthalate may yield endocrine disruptors.
Environmental Health Perspectives, 118(4), 445–448.
Seckl, J. R. (2004). Prenatal glucocorticoids and long-term programming. European
Journal of Endocrinology, 151, U49–U62.
Sohoni, P., & Sumpter, J. P. (1998). Several environmental oestrogens are also anti-
androgens. Journal of Endocrinology, 158, 327–339.
Tabb, M. M., & Blumberg, B. (2006). New modes of action for endocrine-disrupting
chemicals. Molecular Endocrinology, 20, 475–482.
Takeuchi, S., Iida, M., Kobayashi, S., Jin, K., Matsuda, T., & Kojima, H. (2005).
Differential effects of phthalate esters on transcriptional activities via human
estrogen receptors a and b, and androgen receptor. Toxicology, 210, 223–233.
Tillmann, M., Schulte-Oehlmann, U., Duft, M., Markert, B., & Oehlmann, J. (2001).
Effects of endocrine disruptors on prosobranch snails (Mollusca: Gastropoda) in
the laboratory. Part III: Cyproterone acetate and vinclozolin as antiandrogens.
Ecotoxicology, 10, 373–388.
Van der Linden, S. C., Heringa, M. B., Man, H.-Y., Sonneveld, E., Puijker, L. M.,
Brouwer, A., & Van der Burg, B. (2008). Detection of multiple hormonal activities
in wastewater effluents and surface water, using a panel of steroid receptor
CALUX bioassays. Environmental Science and Technology, 42, 5814–5820.
Wagner, M., & Oehlmann, J. (2009). Endocrine disruptors in bottled mineral water:
Total estrogenic burden and migration from plastic bottles. Environmental
Science and Pollution Research, 16, 278–286.
Wagner, M., & Oehlmann, J. (2010). Endocrine disruptors in bottled mineral water:
Estrogenic activity in the E-Screen. Journal of Steroid Chemistry and Molecular
Biology. doi:10.1016/j.jsbmb.2010.10.007.
WHO. (2008). Guidelines for drinking-water quality, 3rd ed., Geneva, Switzerland.
Willemsen, P., Scippo, M. L., Kausel, G., Figueroa, J., Maghuin-Rogister, G., Martial, J.
A., & Muller, M. (2004). Use of reporter cell lines for detection of endocrine-
disrupter activity. Analytica and Bioanalytical Chemistry, 378, 655–663.
Willemsen, P., Scippo, M. L., Maghuin-Rogister, G., Martial, J. A., & Muller, M. (2002).
Use of specific bioluminescent cell lines for the detection of steroid hormone
(ant) agonist in meat production animals. Analytica Chimica Acta, 473, 119–126.