JOURNAL OF APPLIED TOXICOLOGY PARABENS IN HUMAN BREAST TUMOURS 5

J. Appl. Toxicol. 24, 5–13 (2004)

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jat.958

Concentrations of Parabens in Human Breast

Tumours

P. D. Darbre,1,* A. Aljarrah,2 W. R. Miller,2 N. G. Coldham,3 M. J. Sauer4 and G. S. Pope1

1
Division of Cell and Molecular Biology, School of Animal and Microbial Sciences, University of Reading, Reading
RG6 6AJ, UK
2
The Edinburgh Breast Unit Research Group, Paderewski Building, Western General Hospital, Edinburgh EH4 2XU,
UK
3
Department of Bacterial Diseases, Veterinary Laboratories Agency, Weybridge, New Haw, Addlestone, Surrey KT15
3NB, UK
4
Department of TSE Molecular Biology, Veterinary Laboratories Agency, Weybridge, New Haw, Addlestone, Surrey
KT15 3NB, UK

Key words: parabens; xenoestrogens; oestrogenic activity; HPLC–MS–MS; human breast cancer; preservatives; cosmetics.

Parabens are used as preservatives in many thousands of cosmetic, food and pharmaceutical products to which
the human population is exposed. Although recent reports of the oestrogenic properties of parabens have
challenged current concepts of their toxicity in these consumer products, the question remains as to whether
any of the parabens can accumulate intact in the body from the long-term, low-dose levels to which humans are
exposed. Initial studies reported here show that parabens can be extracted from human breast tissue and
detected by thin-layer chromatography. More detailed studies enabled identification and measurement of mean
concentrations of individual parabens in samples of 20 human breast tumours by high-pressure liquid chroma-
tography followed by tandem mass spectrometry. The mean concentration of parabens in these 20 human
breast tumours was found to be 20.6 ± 4.2 ng g−1 tissue. Comparison of individual parabens showed that
methylparaben was present at the highest level (with a mean value of 12.8 ± 2.2 ng g−1 tissue) and represents
62% of the total paraben recovered in the extractions. These studies demonstrate that parabens can be found
intact in the human breast and this should open the way technically for more detailed information to be
obtained on body burdens of parabens and in particular whether body burdens are different in cancer from
those in normal tissues. Copyright © 2004 John Wiley & Sons, Ltd.

survey of 215 cosmetic products found that parabens were

INTRODUCTION used in 99% of leave-on products and in 77% of rinse-off
cosmetics (Rastogi et al., 1995).
Animal studies have shown that parabens are rapidly
The alkyl esters of p-hydroxybenzoic acid (parabens) are absorbed, metabolized and excreted. Parabens are quickly
used widely as preservatives in many thousands of cos- absorbed from the gastrointestinal tract and from blood,
metic, food and pharmaceutical products (Elder, 1984). hydrolysed to p-hydroxybenzoic acid, conjugated and the
These simple esters have proved to be very effective anti- conjugate excreted in the urine (Jones et al., 1956; Heim
microbial agents, with antimicrobial activity increasing with et al., 1957; Tsukamoto & Terada, 1960, 1962, 1964;
the length of the alkyl grouping from methyl to n-butyl Derache & Gourdon, 1963; Phillips et al., 1978; Kiwada
(Murrell & Vincent, 1950), and it is the simplicity and et al., 1979). Parabens also can be absorbed rapidly through
effectiveness of these compounds that have resulted in their intact skin (Whitworth & Jun, 1973; Fischmeister et al.,
widespread use. As such, the human population is exposed 1975; Komatsu & Suzuki, 1979) and this can be influ-
to parabens from a wide variety of sources on a daily basis. enced by the presence of penetration enhancers found in
Parabens are permitted as preservatives in food up to cosmetic preparations (Kitagawa et al., 1997). However,
0.1% and the average daily intake of parabens from food the presence of carboxylesterases in skin and subcu-
by adult humans was estimated in 1984 to be 4–6 mg kg−1 taneous fatty tissues results in varying hydrolysis to p-
(Elder, 1984). In cosmetics, parabens are permitted in hydroxybenzoic acid (Lobemeier et al., 1996) and this can
concentrations of up to 1% (Elder, 1984). In 1984, it was also influence absorption (Bando et al., 1997). However,
estimated that parabens were used in 13 200 different the question remains as to whether any of the parabens
cosmetic formulations (Elder, 1984) and a more recent can enter the body intact from the long-term, low-dose
levels to which humans are exposed. Parabens have a high
* Correspondence to: Dr. P. D. Darbre, Division of Cell and Molecular oil/water partition coefficient and water solubility decreases
Biology, School of Animal and Microbial Sciences, University of Reading, with increase in ester chain length (Elder, 1984). There-
PO Box 228, Whiteknights, Reading RG6 6AJ, UK.
E-mail: p.d.darbre@reading.ac.uk
fore, if any parabens do enter the human body intact, they
Contract/grant sponsor: Seedcorn Fund of the Veterinary Laboratories may be able to accumulate in fatty components of body
Agency. tissues in a similar manner to that of other lipophilic
Copyright © 2004 John Wiley & Sons, Ltd. J. Appl. Toxicol.
Received24,23
5–13
June(2004)
2003
Revised 7 October 2003
Copyright © 2004 John Wiley & Sons, Ltd. Accepted 8 October 2003
6 P. D. DARBRE ET AL.

pollutants that are known to bioaccumulate (Dobson Hardell et al., 1996; Guttes et al., 1998; Stellman et al.,
et al., 1989; Dobson, 1993; Sonawane, 1995; Hardell et al., 1998, 2000). This study has aimed to investigate whether
1996; Guttes et al., 1998; Stellman et al., 1998, 2000; Darbre, parabens also can be detected in human breast tissue,
1998). using available breast tumour material. Initial experiments
Most studies have indicated that parabens are not enabled the extraction of total parabens from human breast
mutagenic (Elder, 1984), but there are reports that they tissue to be visualized by thin-layer chromatography. More
can cause chromosomal aberrations (Ishidate et al., 1978), detailed studies enabled identification and measurement
particularly in the co-presence of polychlorinated biphenyls of individual parabens in human breast tumour samples by
(Matsuoka et al., 1979), and subcutaneous administration high-pressure liquid chromatography (HPLC) followed by
of methylparaben has been reported to cause mammary tandem mass spectrometry (MS/MS).
adenocarcinomas in rats (Mason et al., 1971). At a cellular
level, parabens have been shown capable of disrupting
cellular function through inhibiting secretion of lysosomal
enzymes (Bairati et al., 1994) and causing mitochondrial MATERIALS AND METHODS
dysfunction (Nakagawa & Maldeus, 1998). However, the
recent discovery that parabens possess oestrogenic activity
Human breast tumour material
has challenged the concepts of their toxicity in new ways.
Because parabens can bind to oestrogen receptors, they Samples of human breast tumour material were collected
may be able to mediate unwanted effects at much lower at the Edinburgh Breast Unit and stored in liquid nitrogen.
concentrations and more specifically than through non-
receptor mediated mechanisms.
Chemical standards
The oestrogenic activity of parabens was first reported
in 1998 (Routledge et al., 1998). Since then, parabens have Methylparaben, ethylparaben, n-propylparaben, n-
been shown to bind to oestrogen receptors from different butylparaben and benzylparaben were purchased from
sources, including rodent uterus (Routledge et al., 1998; Sigma (Poole, UK). Isobutylparaben was a gift from Nipa
Blair et al., 2000; Fang et al., 2001) and MCF7 human breast Laboratories (Mid-Glamorgan, UK). All compounds were
cancer cells (Byford et al., 2002; Darbre et al., 2002, 2003). made as stock solutions of 0.1 M in ethanol.
They have been shown to regulate oestrogen-responsive
reporter gene expression in yeast cells (Routledge et al.,
Extraction of parabens from human breast material
1998; Jin-Sung et al., 2000; Nishihara et al., 2000) and in
and analysis by thin-layer chromatography
human breast cancer cells (Byford et al., 2002; Darbre
et al., 2002, 2003), and expression of the endogenous All glassware was pre-washed in 0.1 M NaOH and
oestrogen-regulated genes pS2 (Byford et al., 2002) and extractions were performed using sterile polycarbonate
progesterone receptor (Okubo et al., 2001) in breast tubes (Falcon). Samples of human breast tissue (1 g) were
cancer cells. Parabens can increase the growth of MCF7 chopped finely with a sterile razor and homogenized in
human breast cancer cells (Okubo et al., 2001; Byford 5 ml of hexane using a hand-homogenizer. Samples were
et al., 2002; Darbre et al., 2002, 2003), which can be blocked left in a sealed polycarbonate tube with mixing for 1 h and
with the antioestrogen ICI 182 780 (faslodex) (Byford then spun at 1500 rpm in a bench centrifuge at room tem-
et al., 2002; Darbre et al., 2002, 2003), demonstrating the perature for 2 min. The supernatant was placed in a sterile
growth effects to be oestrogen-receptor-mediated. Their polycarbonate tube, 5 ml of 0.1 M potassium bicarbonate
oestrogenic activity has been demonstrated also in animal was added and the tube was inverted 40 times by hand.
models in vivo in fish (Pedersen et al., 2000) and in The mixture was spun at 1500 rpm at room temperature
increasing uterine weight in immature rats (Routledge for 2 min to separate the phases. The upper yellow hexane
et al., 1998) and immature mice (Darbre et al., 2002, 2003). layer containing phenolic compounds was placed in a new
In line with other environmental oestrogens, butylparaben sterile polycarbonate tube, 5 ml of 0.1 M potassium car-
has been shown also to be able to alter reproductive bonate was added and again the tube was inverted 40 times
function in male rats, including reduction in sperm counts by hand. The mixture was spun at 1500 rpm at room tem-
(Oishi, 2001). In general, the oestrogenic and antimicro- perature for 2 min to separate the phases. The lower aque-
bial activities of the parabens increase with the length and ous layer containing the phenols as potassium salts was
branching of the alkyl ester (Darbre et al., 2002, 2003). taken into a new sterile polycarbonate tube and acidified
Because oestrogen is known to influence the incidence by the addition of 300 µl of concentrated hydrochloric acid
of breast cancer (Lipworth, 1995) and ablation of oestro- to give a pH in the 1–3 range (checked with pH paper).
gen action remains the preferred treatment for hormone- The free phenolic compounds released on acidification were
sensitive breast tumours (Miller, 1996), the presence of extracted into 5 ml of diethyl ether by inverting the tube
oestrogenic chemicals in the breast area could potentially by hand 40 times (Pope et al., 1990). The mixture was spun
influence both the incidence and treatment of breast at 1500 rpm at room temperature for 2 min to separate the
cancer. Parabens are used as preservatives in a range of phases. The upper ether layer was removed and evaporated
cosmetics applied to the underarm and breast area and it to dryness under nitrogen overnight in a fume hood.
has been suggested that regular application of such oestro- The extract was taken up in 50 µl of ethanol and ali-
genic chemicals could influence breast cancer development quots were run against paraben standards (50–400 ng per
(Darbre, 2001, 2003; Harvey, 2003). However, the outstand- track) on thin-layer chromatography plates (DC-Alufolien
ing question remains as to whether parabens can enter Kieselgel 60 F254, Merck; ca. 6 cm wide × 8 cm high) using
and accumulate in the human breast. Previous studies have a solvent of 5% (v/v) ethanol–95% (v/v) chloroform. Para-
identified other environmental oestrogenic chemicals that bens were visualized under ultravoiolet light. For quantita-
can accumulate in fatty tissue of the breast (Dobson, 1993; tion, the image under ultraviolet light was captured digitally
Copyright © 2004 John Wiley & Sons, Ltd. J. Appl. Toxicol. 24, 5–13 (2004)
 PARABENS IN HUMAN BREAST TUMOURS 7

and relative levels of bands were analysed by image analy- extraction carried out alongside, with all procedures iden-
sis using the software packages Transform 3.4 (Fortner) tical except for the omission of tumour material. However,
and Origin 6.0. analysis by HPLCMS/MS was carried out for all samples
on the same day sequentially. Final paraben concentrations
were calculated by subtraction of the values obtained from
Extraction of parabens from human breast tumour
the corresponding blank extraction. Because the blank
material and analysis by HPLCMS/ MS
values showed variation, statistical analysis was performed
Samples of human breast tumour material (0.25 g) were using the paired t-test method (Snedecor & Cochran, 1980).
chopped finely with a sterile razor and homogenized in a
mixture of 6.25 ml of ethanol and 6.25 ml of acetone. This
mixture was left with periodic shaking overnight in a sealed
glass Corex tube. The next day, the mixture was spun at RESULTS
2500 rpm for 10 min on a bench centrifuge at room tem-
perature. The supernatant was removed to a clean Corex
Extraction of parabens from breast tissue and
tube. The pellet was re-extracted with a further 1.5 ml of
detection by thin-layer chromatography
ethanol and 1.5 ml of acetone, spun and the two supernat-
ants pooled. The total supernatant was dried under nitro- In initial exploratory experiments it was possible to
gen at room temperature. To the residue was added 6 ml detect parabens in human breast tissue using the extrac-
of 70% (v/v) aqueous methanol; the mixture was vortexed tion procedures described in the Materials and Methods
and then incubated overnight at −20 °C. The next day, the section, followed by thin-layer chromatography against
mixture was spun at 3200 rpm for 20 min at 4 °C and the paraben standards. Aliquots (10–400 ng) of methylparaben,
supernatant was removed to a clean Corex tube. The pellet ethylparaben, n-propylparaben, n-butylparaben and iso-
was re-extracted with a further 1 ml of 70% (v/v) aqueous butylparaben were run on thin-layer plates and could be
methanol by vortexing and spun again at 3200 rpm for detected under ultraviolet light. Under these conditions all
20 min at 4 °C. The two supernatants were pooled and the paraben standards ran to the same position, which was,
dried under nitrogen for analysis by HPLCMS/MS. on average, 0.47 ± 0.03 of the distance to the solvent front.
The extracts were dissolved in HPLC mobile phase Extracts of human breast tissue contained compounds
(0.25 ml) and the paraben concentration determined by visible under ultraviolet light at the same relative position
HPLCMS/MS. Samples (20 µl) of the final extracts were as the paraben standards. From rough comparison by
chromatographed on a Hypersil Elite C18 column (150 × eye of the relative levels of paraben standards, it was
2.1 mm; 5 µm) at a flow rate of 0.3 ml min−1 and eluted estimated over six separate extractions that the samples
with a linear binary gradient of 15 mM ammonium acetate contained in the region of 10–50 ng paraben per g breast
pH 4.5 (A) and acetonitrile (B) (t = 0 min A 70%, t = tissue. Figure 1 shows the results of one experiment in
15 min A 40%, t = 16 min A 70%, t = 25 min next injec- which three aliquots (97, 194 and 388 ng) of n-butylparaben
tion). The HPLC retention times for the paraben stand- standards were run on thin-layer plates alongside the
ards are provided in Table 1. The parabens were detected extract of 1 g of breast tissue. The relative intensities of
with a Sciex API 2000 triple quadrupole mass spectro- the resulting bands under ultraviolet light were subjected
meter equipped with a heated nebulizer probe operated to image analysis and plotted as a standard curve shown in
in the negative ion mode. Optimal setting of the instru- Fig. 1. The relative intensity of the paraben band extracted
ment for detection by mass reaction monitoring (MRM) from 1 g of tissue was 11 730, which corresponded to 47.1 ng
was established empirically by infusion of paraben stand- paraben g−1 tissue.
ards (1 µg ml−1). The mass transitions selected for MRM It was on the basis of these preliminary results that we
detection utilized the fragmentation of the deprotonated then proceeded to more detailed identification of individual
molecular ion and are listed in Table 1. Chromatographic parabens by HPLCMS/MS
peaks corresponding to individual parabens were detected
automatically and the mass of analyte calculated after
Extraction of parabens from human breast tumours
interpolation from calibration curves prepared over the
and analysis by HPLCMS/MS
working range 1–300 ng ml−1 using the Analyst™ (PE
Biosystems) software package. Retention times and mass transition for MRM detection
Extractions were performed in groups such that each for the six paraben standards are shown in Table 1.
group of two to five tumour extractions had one blank Parabens were extracted from a sample of each of 20
human breast tumours and extracts were analysed by
Table 1—Paraben standards: HPLC retention times and mass
HPLCMS/MS against paraben standards as described in
transition for MRM detection
the Materials and Methods section. Chromatographic peaks
due to methylparaben, ethylparaben, n-propylparaben,
Analyte HPLC retention Mass transition n-butylparaben and isobutylparaben were seen in breast
time (min) (Q1–Q3; m /z) tumour extracts and were well resolved from one another.
for MRM detection No peaks due to benzylparaben at its retention time
of 14.0 min were seen in any of the tumour extracts. A
Methylparaben 4.6 151.1–92.1 typical chromatogram is shown in Fig. 2.
Ethylparaben 7.3 165.1–92.1 At a practical level, extractions were performed in
n-Propylparaben 10.6 179.1–92.1 small groups such that each group contained between
Isobutylparaben 13.4 193.1–92.1
two and five tumour samples together with one blank
n-Butylparaben 13.7 193.1–92.1
extraction. The blank extraction was performed with all
Benzylparaben 14.0 227.3–92.1
procedures identical, except for the omission of tumour
Copyright © 2004 John Wiley & Sons, Ltd. J. Appl. Toxicol. 24, 5–13 (2004)
8 P. D. DARBRE ET AL.

Figure 1. Detection of parabens from human breast tissue by thin-layer chromatography. Three aliquots of n-butylparaben (97, 194 and
388 ng) were run as standards on thin-layer plates alongside the extract of 1 g of breast tissue, and the relative intensities of the resulting
bands under ultraviolet light were subjected to image analysis. The relative intensities of the bands for the three aliquots of n-
butylparaben were plotted as a standard curve as shown. The relative intensity of the paraben band extracted from 1 g of tissue was
11 730, which calculated to an equivalent of 47.1 ng of paraben.

material. The concentrations of parabens in the 20 tumours

as measured by HPLCMS/MS were corrected by subtrac- DISCUSSION
tion of the corresponding blank value. Results are shown
in Table 2. Because the blank values showed variation, the
statistical significance of the mean corrected concentra- Mean concentrations of each of six parabens in extracts of
tions of each paraben in the 20 tumour extracts was tested 20 human breast tumours (in the range 0–12.8 ng g−1 tissue;
by the paired t-test method, thus enabling the confidence Table 3) have been measured with acceptable confidence.
limits of these mean values to be calculated (Table 3). The reasons for the analytical blank values for parabens
The reasons for the blank values for parabens, and their in these studies have not been identified definitively but
variation, are not clear. The MS data indicated that the probably relate to the ubiquitous use of parabens as pre-
blank values were genuinely parabens and not other con- servatives even in laboratory detergents and personal care
taminating compounds. The blank values did not come products of the operators. Analogous problems have been
from the HPLCMS/MS procedure because blanks through encountered with the measurement of phthalate esters
the equipment were entirely negative. The blank values because of their common use as plasticizers and their
came from the extraction procedure itself. In a series of 30 ubiquitous dispersal as impurities in solvents, water, glass-
blanks carried out on individual parts of the extraction ware and many items of clinical and analytical laboratory
procedure, it was not possible to identify any one specific equipment (Lopez-Aviva et al., 1990; Leung & Giang, 1993;
reagent or procedure contributing to the blank value. Colon et al., 2000). More recent work in these laboratories
However, when blank values were subtracted from the (unpublished) has shown that immersion of all glassware
corresponding tumour extract values, 18/20 tumour extrac- in 1.0 M aqueous sodium hydroxide, followed by copious
tions showed values of total paraben above the blank val- rinsing with double-distilled water, prior to use of this glass-
ues. Values for total paraben present in the 20 tumour ware in tissue extraction greatly reduces the blank values
samples were 0–54.5 ng g−1 tissue, with an overall mean of paraben concentrations as measured by HPLCMS/MS.
value of 20.6 ng g−1. Methylparaben was present at the high- This addition to the analytical procedure is therefore
est level, with an average value of 12.8 ± 2.2 ng g−1 tissue. recommended for use in further studies on paraben
This represented 62% of the total paraben recovered in concentrations in tissues.
the extraction. Benzylparaben was not detected in any The total mean paraben level was found to be of the
tumour extract. order of 20 ng g−1 tissue. This adds parabens to the list of
Estimates of recovery of parabens from the extraction environmental oestrogenic chemicals that can be found
procedure were made by spiking samples with benzylpara- to accumulate in the human breast and already includes
ben, because this was the only paraben not detected in polychlorinated biphenyls (PCBs) and organochlorine
any blank or tumour extract. Analysis by HPLCMS/MS of pesticides (OCPs) (Dobson, 1993, Hardell et al., 1996;
three extraction blank samples, each spiked with 200 ng of Guttes et al., 1998; Stellman et al., 1998, 2000). Compar-
benzylparaben, gave an average recovery of this paraben isons between the relative levels of parabens and other
of 48.5% ± 4.8%. pollutants are not easy because several factors have to be
Copyright © 2004 John Wiley & Sons, Ltd. J. Appl. Toxicol. 24, 5–13 (2004)
 PARABENS IN HUMAN BREAST TUMOURS 9

Figure 2. The HPLCMS/MS chromatograms for methylparaben, ethylparaben, n-propylparaben, isobutylparaben and n-butylparaben in
a human breast tumour extract. Tumour tissue was extracted as described in the text, chromatographed on a Hypersil Elite HPLC column
and detected by tandem mass spectrometry in the mass reaction monitoring mode. The annotated arrows indicate the identity of the
peaks evident in the chromatograms. Benzylparaben was not seen.

taken into account, including the source of tissue used and This demonstrates that at least a proportion of the parabens
the number of isomers or congeners. Most studies of present in cosmetic, food and pharmaceutical products can
bioaccumulation of pollutant chemicals are carried out be absorbed and retained in human body tissues without
by using serum or urine and studies using breast adipose hydrolysis by tissue esterases to the common metabolite p-
tissue are few. Furthermore, for parabens there are only hydroxybenzoic acid. These results complement earlier
six commonly used forms whereas for PCBs there are 209 studies in which there was evidence that the oestrogenic
congeners. Studies of breast adipose tissue from women in properties of these parabens in culture of human breast
Long Island, New York, without breast cancer showed a cancer cells were also due to the esters themselves and not
mean body burden for 14 PCB congeners of 267 ng g−1 and to a common metabolite (Byford et al., 2002; Darbre et al.,
for seven OCPs of 707.5 ng g−1 (Stellman et al., 1998). 2002, 2003). However, these studies cannot identify either
However, Table 4 shows that levels in breast tissue of the source of the parabens or whether they entered the
individual pesticide residues and PCB congeners vary human body by an oral or by a topical route. Nor can they
substantially. Although knowledge of total body burdens identify whether the parabens entered the human breast
of these compounds is far from complete, the accumula- by a systemic route or through non-systemic mechanisms
tion of parabens in breast tissue does fall within the broad involving simply local absorption and diffusion from chemi-
range of these other compounds. cal overload of topical preparations applied to the breast
In the present study, paraben concentrations measured area. Recent evaluation of parabens in uterotrophic assays
in tumours were unequivocally of the esters themselves. has shown them to give oestrogenic responses in immature
Copyright © 2004 John Wiley & Sons, Ltd. J. Appl. Toxicol. 24, 5–13 (2004)
 10

Table 2—The HPLCMS/MS analysis of parabens in 20 human breast tumoursa

Tumour extract 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Mean SEM

Benzylparaben 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Isobutylparaben 5.2 0.1 5.2 3.3 2.8 3.5 2.5 1.9 4.9 2.2 1.7 3.9 2.2 2.3 1.2 2.0 1.7 0.0 1.0 0.0 2.4 0.4

Copyright © 2004 John Wiley & Sons, Ltd.

n-Butylparaben 15.3 7.2 29.5 22.4 14.3 15.5 9.8 6.7 10.3 7.4 2.4 14.7 7.1 6.7 7.4 7.2 5.3 3.1 5.7 3.2 10.1 1.5
n-Propylparaben 6.5 7.9 18.6 18.4 10.9 17.9 10.0 10.5 11.6 6.9 9.1 16.5 10.2 17.2 10.1 15.4 4.2 5.3 5.5 5.6 10.9 1.1
Ethylparaben 7.2 3.0 9.6 6.3 4.6 2.3 1.9 1.8 7.0 3.2 1.8 6.4 2.1 4.5 1.1 2.2 1.8 0.7 2.0 2.1 3.6 0.6
Methylparaben 34.2 20.6 53.0 49.9 34.4 37.6 27.3 19.6 35.7 16.0 17.2 36.4 21.6 39.6 18.5 28.8 36.7 8.2 12.0 10.7 27.9 2.8
Total paraben 68.3 38.8 115.9 100.3 67.0 76.9 51.4 40.4 69.5 35.7 32.1 77.9 43.1 70.3 38.3 55.6 49.7 17.3 26.2 21.6 54.8 5.8
Blank value 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Mean SEM
as 1 as 3 as 3 as 6 as 6 as 6 as 10 as 10 as 10 as 10 as 15 as 15 as 18 as 18
Benzylparaben 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Isobutylparaben 0.0 0.0 2.2 2.1 2.3 1.1 1.3 0.4
n-Butylparaben 6.5 18.0 9.3 4.0 8.2 2.2 8.0 2.3
n-Propylparaben 2.0 13.4 12.0 6.8 8.4 4.8 7.9 1.8
Ethylparaben 1.9 2.2 1.9 1.9 0.0 1.6 1.6 0.3
Methylparaben 10.1 27.8 20.5 10.3 11.6 9.9 15.0 3.0
Total paraben 20.5 61.4 45.9 25.0 30.6 19.6 33.8 6.8
P. D. DARBRE ET AL.

Tumour less blank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Mean SEM

Benzylparaben 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Isobutylparaben 5.2 0.1 5.2 3.3 2.8 1.3 0.3 −0.3 2.7 0.2 −0.4 1.8 0.1 0.2 −1.1 −0.3 −0.6 −1.1 −0.1 −1.1 0.9 0.4
n-Butylparaben 8.8 0.7 11.5 4.4 −3.7 6.2 0.5 −2.6 1.0 3.4 −1.6 10.7 3.1 2.7 −0.8 −1.0 −2.9 1.0 3.5 1.0 2.3 1.0
n-Propylparaben 4.5 5.9 5.2 5.0 −2.5 5.9 −2.0 −1.5 −0.4 0.2 2.3 9.7 3.4 10.4 1.7 7.0 −4.3 0.5 0.7 0.8 2.6 0.9
Ethylparaben 5.3 1.2 7.4 4.1 2.4 0.5 0.0 −0.1 5.1 1.3 −0.2 4.5 0.2 2.6 1.1 2.1 1.8 − 0.9 0.4 0.5 2.0 0.5
Methylparaben 24.1 10.5 25.2 22.1 6.6 17.1 6.8 −0.9 15.2 5.7 6.9 26.1 11.3 29.3 6.9 17.2 25.1 −1.8 2.1 0.8 12.8 2.2
Total paraben 47.9 18.4 54.5 38.9 5.6 31.0 5.5 −5.5 23.7 10.7 7.1 52.8 18.1 45.3 7.7 25.0 19.1 − 2.3 6.6 2.0 20.6 4.2

a
Paraben extractions were performed in small groups such that each group contained between two and five tumour samples together with one corresponding blank extraction. The blank
extraction was performed with all procedures identical except for the omission of tumour material. Results are shown in ng g−1 tumour for the 20 extractions and for the corresponding blank
values. The concentrations of parabens in the 20 tumours were then each corrected by subtraction of the corresponding blank value.

J. Appl. Toxicol. 24, 5–13 (2004)

 PARABENS IN HUMAN BREAST TUMOURS 11

Table 3—Confidence limits of mean concentrations (ng g−1) of epithelial cells in the human breast. Although in rodent
parabens in the 20 human breast tumours of Table 2 uterotrophic assays the levels of parabens were adminis-
tered at a higher range of 0.1–10 mg g−1 body weight
Tumour minus blank Mean Confidence limit (Routledge et al., 1998, Darbre et al., 2002, 2003), these
studies did not incorporate any measurements of paraben
Benzylparaben 0.0 0.0–0.0 (95%) levels reached in the uterus at the time of response, which
Isobutylparaben 0.9 0.1–1.7 (90%)
prevents assessment of the concentrations needed for physi-
n-Butylparaben 2.3 0.3–4.3 (95%)
n-Propylparaben 2.6 0.7–4.5 (95%)
ological response.
Ethylparaben 2.0 1.0–3.0 (95%) It is interesting that the paraben detected in greatest
Methylparaben 12.8 8.2–17.4 (95%) amounts was methylparaben. This may reflect the more
widespread use of methylparaben in consumer products
Total paraben 20.6 11.8–29.4 (95%)
(Rastogi et al., 1995). Alternatively, it may reflect the
greater ability of methylparaben to be absorbed into body
tissues and to resist hydrolysis by esterases of human skin
and subcutaneous fat tissue (Lobemeier et al., 1996). By
rodent uterus only when administered subcutaneously contrast, benzylparaben was not found in any of the 20
or topically but not orally (Routledge et al., 1998; Hossaini breast tumours and this may similarly be attributed to its
et al., 2000; Darbre et al., 2002, 2003), which suggests that less frequent use in consumer products.
skin penetration may be an important route for entry to These measurements of paraben concentrations in breast
the body. tumours open the way technically to more detailed
A major issue in studies of accumulation of environ- determinations of paraben levels in human body tissues.
mental pollutants in body tissues is whether the levels This study used 20 breast tumour samples because of the
reached could be sufficiently high to exert any biological availability of the material. However, it will now be im-
action. In four of the 20 tumours, total paraben concen- portant to measure levels in corresponding normal tissue
tration was more than twice the average level and, allow- to determine whether there is any difference between
ing for a 50% recovery of parabens through the analytical normal and cancer tissues. Larger studies also are needed
procedure, the corrected average level of parabens was to give more representative values for body burdens in
ca. 100 ng g−1 tissue. This concentration may be compared different tissues and across the human population. A main
with the level (ca. 150 ng ml−1; 10−6 M) in culture med- problem with human breast tumour samples is the varied
ium at which n-propylparaben, n-butylparaben and iso- infiltration of the tumour with fatty tissue and blood
butylparaben stimulated growth of oestrogen-dependent vessels and it will be important in future work therefore to
MCF7 human breast cancer cells (Okubo et al., 2001; have more precise histological information on the tumours
Byford et al., 2002; Darbre et al., 2002, 2003). It is there- in order especially to be able to relate results to fatty versus
fore not inconceivable that the levels of parabens meas- non-fatty tissue. It would be informative to ascertain
ured in this study could exert oestrogenic effects on whether there are any gradients in the accumulation of

Table 4—Summary of mean concentrations (ng g−1) of individual pesticide residues and PCB congeners in human
breast adipose tissue from control and breast cancer patients from two published studies

Pesticide or PCB Control Breast cancer Control Breast cancer

HCB 206 343 16 18

β HCH 72 84 16 20
Oxychlordane 39 46
trans -Nonachlor 40 51
p,p′-DDE 450 838 374 419
o,p′-DDD 13 16
p,p′-DDT 24 30 12 12
PCB 74 27 30
PCB 99 14 19
PCB 118 58 85 24 30
PCB 138 176 241 22 29
PCB 146 7 9
PCB 153 437 664 63 76
PCB 156 61 64 9 11
PCB 167 1 2
PCB 170 229 259 11 14
PCB 172 2 2
PCB 178 3 4
PCB 180 258 400 34 42
PCB 183 4 6
PCB 187 13 16
Reference Guttes et al. (1998) Guttes et al. (1998) Stellman et al. (2000) Stellman et al. (2000)
Location Hesse, Germany Hesse, Germany New York, USA New York, USA
Number of samples n = 20 n = 45 n = 323 n = 232

Copyright © 2004 John Wiley & Sons, Ltd. J. Appl. Toxicol. 24, 5–13 (2004)
12 P. D. DARBRE ET AL.

parabens across the human breast from axilla to sternum ment to be made of the impact of these weakly oestro-
in case the topical application of cosmetic at one place genic parabens on human health, and whether paraben
influences the levels of parabens detectable. It will also accumulation from currently permitted levels in cosmetics,
be important to know whether there is any difference foods and pharmaceuticals remains acceptable.
between levels detectable in breast tumours compared
with adjacent non-tumour material in order to determine
whether higher levels of paraben accumulation might be Acknowledgements
present in the tumours. Such information, taken together We are grateful for financial support from the Seedcorn Fund of the Veterinary
with that of concentrations in tissues of endogenous steroid Laboratories Agency (P.D.D., N.G.S., M.J.S.) and for statistical advice from
hormones and other xenoestrogens, should enable assess- Dr M. C. Denham, School of Applied Statistics, University of Reading.

REFERENCES

Bairati C, Goi G, Lombardo A, Tettamanti G. 1994. The esters Heim F, Leuschner F, Wunderlich G. 1957. Metabolism of p-
of p-hydroxy-benzoate (parabens) inhibit the release of hydroxybenzoic acid ethyl ester. Klin. Wochenschr. 35: 823–
lysosomal enzymes by mitogen-stimulated peripheral 825.
human lymphocytes in culture. Clin. Chim. Acta 224: 147– Hossaini A, Larsen JJ, Larsen JC. 2000. Lack of oestrogenic
157. effects of food preservatives (parabens) in uterotropic
Bando H, Mohri S, Yamashita F, Takakura y, Hashida M. 1997. assays. Food Chem. Toxicol. 38: 319–323.
Effects of skin metabolism on percutaneous penetration of Ishidate M, Hayashi M, Sawada M, Matsuoka A. 1978.
lipophilic drugs. J Pharm. Sci. 86: 759–761. Cytotoxicity test on medical drugs. Chromosome aberration
Blair RM, Fang H, Branham WS, Hass BS, Dial SL, Moland CL, tests with Chinese hamster cells in vitro. Eisei Shikensho
Tong W, Shi L, Perkins R, Sheehan DM. 2000. The estrogen Hokoku 96: 55–61.
receptor binding affinities of 188 natural and xenochemicals: Jin-Sung P, Lee BJ, Kang KS, Tai JH, Cho JJ, ChoMH, Inoue T,
structural diversity of ligands. Toxicol. Sci. 54: 138–153. Lee YS. 2000. Hormonal effects of several chemicals in
Byford JR, Shaw LE, Drew MGB, Pope GS, Sauer MJ, Darbre PD. recombinant yeast, MCF7 cells and uterotropic assays in
2002. Oestrogenic activity of parabens in MCF7 human mice. Microbiol. Biotechnol. 10: 293–299.
breast cancer cells. J. Steroid Biochem. Mol. Biol. 80: 49–60. Jones PS, Thigpen D, Morrison JL, Richardson AP. 1956.
Colon I, Caro D, Bourdony CJ, Rosario O. 2000. Identification p-Hydroxybenzoic acid esters as preservatives. III. The
of phthalate esters in the serum of young Puerto Rican physiological disposition of p-hydroxybenzoic acid and its
girls with premature breast development. Environ. Health esters. J Am. Pharm. Assoc. Sci. Ed. 45: 265–273.
Perspect. 108: 895–900. Kitigawa S, Li H, Sato S. 1997. Skin permeation of parabens
Darbre PD. 1998. Environmental contaminants in milk: the prob- in excised guinea pig dorsal skin, its modification by
lem of organochlorine xenobiotics. Biochem. Soc. Trans. 26: penetration enhancers and their relationship with n-octanol/
106–112. water partition coefficients. Chem. Pharm. Bull. (Tokyo) 45:
Darbre PD. 2001. Hypothesis: underarm cosmetics are a cause of 1354–1357.
breast cancer. Eur. J. Cancer Prevent. 10: 389–393. Kiwada H, Awazu S, Hanano M. 1979. The study on the biological
Darbre PD. 2003. Underarm cosmetics and breast cancer. J. Appl. fate of paraben at the dose of practical usage in rat. I. The
Toxicol. 23: 89–95. metabolism and excretion of ethyl-p-hydroxybenzoate (ethyl
Darbre PD, Byford JR, Shaw LE, Horton RA, Pope GS, Sauer MJ. paraben) and p-hydroxybenzoic acid. J. Pharmacobio-Dyn. 2:
2002. Oestrogenic activity of isobutylparaben in vitro and in 356–364.
vivo. J Appl. Toxicol. 22: 219–226. Komatsu H, Suzuki M. 1979. Percutaneous absorption of butyl-
Darbre PD, Byford JR, Shaw LE, Hall S, Coldham NG, Pope GS, paraben through guinea pig skin in vitro. J. Pharm. Sci. 68:
Sauer MJ. 2003. Oestrogenic activity of benzylparaben. J 596–598.
Appl. Toxicol. 23: 43–51. Leung S, Giang BY. 1993. Phthalate esters as potential contam-
Derache R, Gourdon J. 1963. Metabolism of a food preservative: inants during GC analysis of environmental samples using
p-hydroxybenzoic acid and its esters. Food Cosmet. Toxicol. electron capture and mass spectrometric detectors. Bull.
1: 189–195. Environ. Contam. Toxicol. 50: 528–532.
Dobson S, Gilbert M, Howe PD. 1989. DDT and its derivatives — Lipworth L. 1995. Epidemiology of breast cancer. Eur. J. Cancer
environmental aspects. Environ. Health Crit. 83: 1–79. Prevent. 4: 7–30.
Dobson S. 1993. Polychlorinated biphenyls and terphenyls. Lobemeier C, Tschoetschel C, Westie S, Heymann E. 1996.
Environ. Health Crit. 140: 1–496. Hydrolysis of parabenes by extracts from differing layers of
Elder RL. 1984. Final report on the safety assessment of methyl- human skin. Biol. Chem. 377: 647–651.
paraben, ethylparaben, propylparaben and butylparaben. Lopez-Aviva V, Milanes J, Constantine F, Beckert WF. 1990.
J. Ame. Coll. Toxicol. 3: 147–209. Typical phthalate ester contamination incurred using EPA
Fang H, Tong W, Shi LM, Blair R, Perkins R, Branham W, method 8060. J. Assoc. Off. Anal. Chem. 73: 709–720.
Hass BS, Xie Q, Dial SL, Moland CL, Sheehan DM. 2001. Mason MM, Cate CC, Baker J. 1971. Toxicology and carcino-
Structure–activity relationships for a large diverse set of genesis of various chemicals used in the preparation of
natural, synthetic, and environmental estrogens. Chem. Res. vaccines. Clin. Toxicol. 4: 185–204.
Toxicol. 14: 280–294. Matsuoka A, Hayashi M, Ishidate M. 1979. Chromosomal aberra-
Fischmeister I, Hellgren L, Vincent J. 1975. Infrared spectroscopy tion tests on 29 chemicals combined with S9 mix in vitro.
for tracing of topically applied ointment vehicles and active Mutat. Res. 66: 277–290.
substances on healthy skin. Arch. Dermatol. Res. 253: 63–69. Miller WR. 1996. Estrogen and Breast Cancer. Chapman and Hall:
Guttes S, Failing K, Neumann K, Kleinstein J, Georgii S, Brunn H. London.
1998. Chlororganic pesticides and polychlorinated biphenyls Murrell WG, Vincent JM. 1950. The 4-hydroxybenzoic acid esters
in breast tissue of women with benign and malignant breast and related compounds. 4. The bacteriostatic action of 4-
disease. Arch. Environ. Contam. Toxicol. 35: 140–147. hydrozybenzoic acid n-alkyl esters. J. Soc. Chem. Ind. 69:
Hardell L, Lindstrom G, Liljegren G, Dahl P, Magnuson A. 1996. 109–113.
Increased concentrations of octachlorodibenzo-p-dioxin in Nakagawa Y, Maldeus P. 1998. Mechanism of p -hydroxybenzoate
cases with breast cancer — results from a case–control ester-induced mitochondrial dysfunction and cytotoxicity
study. Eur. J Cancer Prevent. 5: 351–357. in isolated rat hepatocytes. Biochem. Pharmacol. 55: 1907–
Harvey PW. 2003. Parabens, oestrogenicity, underarm cosmetics 1914.
and breast cancer: a perspective on a hypothesis (Editorial) Nishihara T, Nishikawa JI, Kanayama T, Dakeyama F, Saito K,
J. Appl. Toxicol. 23: 285–288. Imagawa M, Takatori S, Kitagawa Y, Hori S, Utsumi H. 2000.

Copyright © 2004 John Wiley & Sons, Ltd. J. Appl. Toxicol. 24, 5–13 (2004)
 PARABENS IN HUMAN BREAST TUMOURS 13

Estrogenic activities of 517 chemicals by yeast two-hybrid Snedecor GW, Cochran WG. 1980. Statistical Methods (7th edn).
assay. J. Health Sci. 46: 282–298. Iowa State University Press: Iowa.
Oishi S. 2001. Effects of butylparaben on the male reproductive Sonawane BR. 1995. Chemical contaminants in human milk: an
system in rats. Toxicol. Ind. Health 17: 31–39. overview. Environ. Health Perspect. 103: 197–205.
Okubo T, Yokoyama Y, Kano K, Kano I. 2001. ER-dependent Stellman SD, Djordjevic MV, Muscat JE, Gong L, Bernstein D,
estrogenic activity of parabens assessed by proliferation of Citron ML, White A, Kemeny M, Busch E, Nafziger AN.
human breast cancer MCF-7 cells and expression of ERα and 1998. Relative abundance of organochlorine pesticides and
PR. Food Chem. Toxicol. 39: 1225–1232. polychlorinated biphenyls in adipose tissue and serum of
Pedersen KL, Pedersen SN, Christiansen LB, Korsgaard B, women in Long Island, New York. Cancer Epidemiol.
Bjerregaard P. 2000. The preservatives ethyl-, propyl-, and Biomarkers Prevent. 7: 489–496.
butylparaben are oestrogenic in an in vivo fish assay. Stellman SD, Djordjevic MV, Britton JA, Muscat JE, Citron ML,
Pharmacol. Toxicol. 86: 110–113. Kemeny M, Busch E, Gong L. 2000. Breast cancer risk in
Phillips JC, Topp CS, Gangolli SD. 1978. The metabolism of ethyl relation to adipose concentrations of organochlorine pesti-
and n-propyl-p-benzoate (‘parabens’) in male cats. Toxicol. cides and polychlorinated biphenyls in Long Island, New
Lett. 2: 137–142. York. Cancer Epidemiol. Biomarkers Prevent. 9: 1241–1249.
Pope GS, Cunningham JM, Jenkins N, Waites GMH, Watts GE. Tsukamoto H, Terada S. 1960. Metabolism of drugs. XXIII.
1990. Oestradiol-17β in testis and jugular venous plasma of Metabolic fate of p-hydroxybenzoic acid and its derivatives
intact and hemi-castrated prepubertal lambs and in jugular in rabbits. Chem. Pharm. Bull. (Tokyo) 8: 1066–1070.
venous plasma of castrated prepubertal lambs. Anim. Tsukamoto H, Terada S. 1962. Metabolism of drugs. XXVI.
Reprod. Sci. 22: 9–19. Metabolic fate of p-hydroxybenzoic acid and its derivatives
Rastogi SC, Schouten A, De Kruijf N, Weijland JW. 1995. Con- in rabbits. Chem. Pharm. Bull. (Tokyo) 10: 86–90.
tents of methyl-, ethyl-, propyl-, butyl- and benzylparaben in Tsukamoto H, Terada S. 1964. Metabolism of drugs. XLVII.
cosmetic products. Contact Dermat. 32: 28–30. Metabolic fate of p -hydroxybenzoic acid and its derivatives
Routledge EJ, Parker J, Odum J, Ashby J, Sumpter JP. in rabbits. Chem. Pharm. Bull. (Tokyo) 12: 765–769.
1998. Some alkyl hydroxy benzoate preservatives (para- Whitworth CW, Jun HW. 1973. Influence of polysorbate 20 and
bens) are estrogenic. Toxicol. Appl. Pharmacol. 153: 12– sodium cholate on uptake of p-hydroxybenzoates by the frog
19. Rana pipiens. J. Pharm. Sci. 62: 1890–1891.

Copyright © 2004 John Wiley & Sons, Ltd. J. Appl. Toxicol. 24, 5–13 (2004)