ARTICLE IN PRESS
Water Research 38 (2004) 4491–4501
www.elsevier.com/locate/watres
Contribution of known endocrine disrupting substances to the
estrogenic activity in Tama River water samples from Japan
using instrumental analysis and in vitro reporter gene assay
Takuma Furuichia,, Kurunthachalam Kannanb,
John P. Giesyc, Shigeki Masunagaa
a
Graduate School of Environment and Information Sciences, Yokohama National University, 79-7 Tokiwadai, Hodogaya-ku,
Yokohama 240-8501, Japan
b
New York State Department of Health, Wadsworth Center, Empire State Plaza, P.O. Box 509, Albany, NY 12201-0509, USA
c
Department of Zoology, National Food Safety and Toxicology Center, and Institute for Environmental Toxicology,
Michigan State University, East Lansing, MI 48824, USA
Received 27 January 2004; received in revised form 29 July 2004; accepted 7 August 2004
Abstract
To quantitatively characterize the substances contributing to estrogenic activity in river water, in vitro bioassay using
MVLN cells and instrumental analysis using liquid chromatograph–mass spectrometer (LC/MS) or liquid
chromatograph–tandem mass spectrometer (LC/MS/MS) were applied to river water extracts taken from various
locations in the Tama River, Japan. Tama River water samples were extracted using solid phase extraction and
the crude extracts were fractionated by high-performance liquid chromatography (HPLC) into 10 fractions. The
sixth fraction contained nonylphenol (NP) and octylphenol (OP) at concentrations in the range of 51.6–147 and
6.9–81.9 ng/L, respectively (concentrations corresponding to the original sample volumes). No estrogenic activity,
expressed as 17b-estradiol equivalents (E2-EQB), however, was observed in this fraction (o0.6 ng-E2eq/L).
Instrumentally determined estrogenic activity (E2-EQC), which is the concentrations of NP and OP multiplied by
their corresponding relative potency, was below the detection limit of the MVLN cell bioassay. Estrogenic activities
were detected only in HPLC fraction nos. 7, 8 and 9. Estrone (E1), estradiol (E2) and bisphenol A (BPA) were detected
in these fractions. Estriol (E3) and ethynylestradiol (EE2) were not detected (o0.2 ng/L) in these fractions. The
calculated E2-EQC for BPA was below the detection limit of bioassay. The E2-EQC for E1 and E2 were on the same
order as the estrogenic activity determined by the bioassay (E2-EQB). The ratios of E2-EQC and E2-EQB for E1 and E2
in the three factions collectively (nos. 7–9) were 0.49–0.97 and 0.29–1.12, respectively. Above results indicated that the
major causal substances to the estrogenic activity in the Tama River were E1 and E2.
r 2004 Elsevier Ltd. All rights reserved.
Keywords: Endocrine disruptor; 17b-estradiol; Estrone; MVLN cells
1. Introduction
Corresponding author. Tel.: +81-45-339-4351; fax: +81-
45-339-4373.
E-mail address: d01td002@ynu.ac.jp (T. Furuichi).
A number of estrogenic compounds have been found
in aquatic environments, and their effects on normal
0043-1354/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2004.08.007
ARTICLE IN PRESS
4492
T. Furuichi et al. / Water Research 38 (2004) 4491–4501
endocrine functions of aquatic organisms are of concern
(Purdom et al., 1994; Folmar et al., 1996; Harries et al.,
1996, 1997; Jobling et al., 1998). In the UK and the
USA, male rainbows trout and roaches exposed to
sewage treatment plant (STP) effluents contained
elevated levels of vitellogenin (VTG), a precursor for
female egg yolk protein, and estrogenic substances
present in river water were related to this observation
(Purdom et al., 1994; Folmar et al., 1996; Harries et al.,
1996, 1997; Jobling et al., 1998). Similar observations
have been reported in male roaches from selected rivers
in Japan (Nakamura and Iguchi, 1998; Wanami, 2002).
Based on the laboratory studies, the Ministry of
Environment of Japan reported that one of the
estrogenic substances, nonylphenol (NP), decreased
fertility and increased abnormal sexuality in Japanese
Medaka. About 4.5% of the public water bodies in
Japan contained NP levels in excess of the predicted no
effect concentration of 0.608 mg/L, a benchmark value
selected by the Ministry of Environment of Japan based
on a medaka partial life cycle study. Desbrow et al.
(1998) identified 17b-estradiol (E2), estrone (E1), and
ethynylestradiol (EE2) as the major compounds responsible for the estrogenic activity in some STP effluents in
the UK based on an in vitro reporter gene assay. In a
previous study, we examined soluble and particulate
fractions of river water and sediment collected from the
Tama River using in vitro MVLN cell bioassay as well
as chemical analysis (Masunaga et al., 2000). Estrogenic
activity was found only in soluble fraction of the
samples and was found to be contributed mainly by
E1. Due to the higher detection limit of E2 at the time,
the contribution of E2 to estrogenic activity could not be
calculated. 17b-estradiol equivalent (E2-eq) of each
chemical was calculated based on the chemical analysis
(E2-EQC), by multiplying chemical concentration with
bioassay specific relative potency factor (REP: potencies
relative to 17b-estradiol). The E2-eq of bioassay analysis
(E2-EQB) did not exceed the corresponding E2-EQC.
The presence of anti-estrogenic, response-inhibiting,
and/or cytotoxic substances in the samples might have
suppressed the response in MVLN cells. Thus, further
experiments were conducted to eliminate cytotoxic or
interfering substances in water samples prior to the
measurement of estrogenic activity.
In vitro estrogen receptor mediated, chemically
activated luciferase gene expression assay, using human
breast cancer cells (MVLN), is one of the useful tools to
determine total estrogenic activity in environmental
extracts. Although in vitro assay provides information
on cumulative response of estrogenic activity caused by
co-existing substances, the extent of contribution from
specific substance cannot be determined. Chemical
analysis has the advantage of identifying and quantifying selected target estrogenic substances, but it does not
provide any information on the estrogenic potentials of
Extract
Gel permeation chromatography (GPC) fractionation
F1
F2
F3
F4
F5
F6
OP
NP
F7
F8
E2
E1
E2
EE2
BPA
E3
BPA
F9 F10
E1
Bioassay with MVLN cell
Internal
standard
LC/MS/MS ( Thermoquest TSQ7000 )
or
LC/MS ( Agilent Technology MSD1100 )
Fig. 1. Fractionation procedure for the determination of
estrogenic activity and the target substances in river water.
E1=estrone; E2=17b-estradiol; E3=estriol; EE2=ethynylestradiol; NP=4-nonylphenol; OP=4-tert-octylphenol; BPA=bisphenol A. LC/MS=liquid chromatograph–mass spectrometer;
LC/MS/MS=liquid chromatograph–tandem mass spectrometer.
target compounds. The use of both in vitro bioassay and
instrumental analysis, therefore, is useful for the
identification and characterization of estrogenic activity
in environmental samples.
In this report, we examined the use of gel permeation
chromatography (GPC) as a sample fractionation
procedure (Fig. 1) followed by in vitro gene expression
bioassay using MVLN cells, and instrumental analysis
for the identification of estrogenic activity and compounds responsible for the activity in the Tama
River. The study also aimed to determine the relative
contributions of various potentially estrogenic
substances quantitatively, using liquid chromatograph–
mass spectrometer (LC/MS) or liquid chromatograph–
tandem mass spectrometer (LC/MS/MS).
2. Materials and methods
2.1. Sample collection
The Tama River flows through the west suburban
Tokyo into Tokyo Bay. The drainage area of the river is
1240 km2 with a population of about 5 million.
Residential areas are located along the middle reaches
of the river and nine STPs discharge effluents into the
river. The sewerage system is widely spread over 90% of
the catchment area. The amount of industrial effluent
disposal into the river is relatively small.
ARTICLE IN PRESS
T. Furuichi et al. / Water Research 38 (2004) 4491–4501
The grab sampling of river water was conducted at
five sites; Kitatama No. 1 STP effluent, and near Hino
Bridge, Sekido Bridge, Hara Bridge, and Maruko Bridge
on 9th April 2002. All the samples were collected in
stainless-steel buckets and brought to the laboratory.
The samples were stored at 20 1C in a freezer until
pretreatment.
2.2. Sample extraction
A concentration system with flow control (Nippon
Waters Corporation, Tokyo, Japan) equipped with
1.0 mm (pore size) grass fiber filter (pre-heated at
400 1C for 4 h prior to use) and a solid phase extraction
cartridge filled with hydrophilic copolymer including
N-vinylacetamide (EDS-1) (Showadenko Corporation,
Tokyo, Japan; pre-cleaned with 10 mL methanol and
10 mL nanopure water), was prepared. After the pH of
the samples was adjusted to 3 with acetic acid, 1000 mL
of water samples was passed through this system at a
flow rate of below 10 mL/min. Blank sample (1000 mL)
was prepared by passing nanopure water through the
same system. After passage of the samples, residual
water within the cartridges was removed by centrifugation for 5 min at 1000 rpm. Cartridges were then eluted
with 10 mL of 5 mM trimethylamine in methanol into
glass centrifuge tubes. The eluates were concentrated
under a gentle stream of nitrogen at 30 1C. After drying,
300 mL of methanol was added to the tubes as the crude
extract. One hundred mL of the extract was prepared for
identifying the estrogenic activity in the samples.
Fractionation of the extracts (200 mL) was carried out,
after the presence of estrogenic activity was confirmed
by the bioassay.
2.3. Fractionation of extract
The extracts were fractionated into 10 fractions using
GPC (GF-310HQ) (300 mm 7.6 mm j, Showadenko
Corporation, Tokyo, Japan). GPC was selected because
it is effective in removing matrix substances in samples,
has good recovery of target substances and also because
mobile phase other than water can be used. Highperformance liquid chromatography (HPLC) system
(HP 1100 series, Agilent, CA, USA) and electronic
fraction collector (Model 203B, Gilson, WI, USA) were
used with a column oven temperature of 40 1C. Isocratic
elution with 5mM trimethylamine in methanol at a flow
rate of 0.5 mL/min was used. Two hundred mL of crude
extract was injected to the HPLC system. Each fraction
was collected using the fraction collector at every 4 min.
A total of 10 fractions were collected from each sample
extract. Fractionated samples were concentrated under a
gentle stream of nitrogen at 30 1C. After drying, 200 mL
of methanol was added for the bioassay. After assaying,
170 mL of fractionated samples were concentrated under
4493
a gentle stream of nitrogen at 30 1C. After drying,
methanol was added at 85 mL each of 17b-estradiol-d4,
estrone-d2, and estriol-d2 (50 mg/L) as internal standards
for instrumental analysis.
2.4. Target substances
17b-estradiol (98%), estrone (99%), estriol (E3, 99%),
and Ethynylestradiol (EE2, 98%) were obtained from
Sigma-Aldrich Chemicals (MO, USA). 4-nonylphenol
(NP, 99%), 4-tert-octylphenol (OP, 98%) and bisphenol
A (BPA, 99%) were obtained from Wako Chemical
Industries Ltd. (Osaka, Japan). Internal standards,
E1-d2, E2-d4 and E3-d2 were obtained from CDN
isotope (Quebec, Canada). All solvents used in the
experiments were pesticide grade.
2.5. Instrumental analysis
Target substances in water samples were determined
either by LC/MS/MS or LC/MS. For LC/MS/MS
analysis, fractionated samples were injected into a LC/
MS/MS (TSQ 7000; Thermo Quest, CA, USA) equipped
with C18 reverse phase column (RP-18 GP, 2.0 mm
ID 150 mm, 5 mm, Kanto Chemical, Tokyo, Japan)
(Ishii et al., 2000). A linear 30-min gradient of
acetonitrile/water mobile phase from 30:70 to 70:30 (v/
v) was used. The flow rate of mobile phase was 0.2 mL/
min. Five mL of the fractionated samples was injected.
E2, E1, E3 and EE2 were determined by selected ion
monitoring using atmospheric-pressure-electrospray ionization (ESI). For LC/MS analysis, we used
HP1100LC-MSD (Agilent, CA, USA) with reverse
phase C-18 column (Zorbax 2.1 mm ID 150 mm,
Agilent Technology, CA, USA). Acetonitrile/water
mobile phase was used. The initial condition was 20%
acetonitrile and was increased to 80% in 10 min, and
kept isocratic for 10 min, and then increased to 100%
acetonitrile in 5 min. The flow rate was 0.2 mL/min and
10 mL of the sample was injected. NP, OP and BPA were
determined by selected ion monitoring with ESI negative
ionization mode.
2.6. Bioassay
MVLN cells are derived from human breast carcinoma MCF-7 cell line and are stably transfected with
luciferase reporter gene under the control of estrogen
responsive elements (EREs) of the Xenopus VTG A2
gene for the detection of ER-meditated activity (Pons et
al., 1990). This cell was obtained from INSERM (The
French Institute of Health and Medical Research,
Montpellier, France). The cell culture media, Dulbecco’s
modified Eagle medium-F12 without phenol red (Sigma
Aldrich, MO, USA) was supplemented with 10%
dextran/charcoal-stripped fetal bovine serum (Hyclone,
ARTICLE IN PRESS
T. Furuichi et al. / Water Research 38 (2004) 4491–4501
4494
UT, USA), 1 mM sodium pyruvate and 1 mg/mL insulin
(Sigma Aldrich, MO, USA). The cell culture was plated
in 60 interior wells of a 96-well microplate (Perkin
Elmer, MA, USA). Culture fluid (250 mL) was dispensed
into each well at a density of approximately 75,000 cells/
mL. The 36 exterior wells of each plate were filled with
250 mL of culture media. Cells were incubated overnight
in a CO2 incubator at 37 1C prior to dosing. Test wells
were dosed with 2.5 mL of sample extracts or fractionates. In the bioassay experiment, methanol was used as a
solvent. Solvent control wells were dosed with methanol
and blank wells were not dosed. A minimum of three
control wells and three blank wells were prepared for
each plate. All the samples were tested in triplicate. For
standard calibration, six different concentrations of E2
stock solution (0.1–100 nM in vial) were dosed in
triplicate for at least every three plates. After 3 days of
exposure, medium was removed and washed with
phosphate buffer solution (PBS), which contained MgCl2
and CaCl2. After dispensing 50 mL of PBS to each well,
luciferin solution was added to the wells. Luciferase
activity was then measured by luminescence microplate
reader (Lumicount, Perkin Elmer, MA, USA).
2.7. Bioassay data analysis
Estrogenic activity, expressed as mean relative luminescence unit (RLU) of the three replicates, was
converted into the ratio of the mean maximum response
observed for E2 standard curve generated daily. The
maximum response ratio of E2 was set as 1. The
estrogenic activity in samples derived from bioassay
analysis was expressed as E2 equivalent quantity (E2EQB, ng E2-eq/L), which was estimated on the basis of
the logistic model (Nakagawa and Koyanagi, 1992;
Yanagawa, 2002)
Y¼
expðb0 þ b1 X Þ
:
1 þ expðb0 þ b1 X Þ
(1)
The logistic model has two parameters, b0 and b1,
which describe intercept and slope, respectively. Y stands
for estrogenic activity in a sample relative to E2 maximum
response. X is the concentration ratio of the sample. To
perform the logistic modeling (Eq. (1)), nonlinear leastsquares estimation (Q) was used as a loss function
Q¼
n
X
ðY Y^ Þ2 :
(2)
i¼1
For fitting logistic model to the data obtained, a QuasiNewton method was applied. Then, E2 equivalent
concentration (E2-EQB) in the sample was estimated at
a point exhibiting 0.5 (50%) of the maximum E2 response.
Estrogenic activity with 95% confidence interval (CF) was
calculated using the variance obtained during the model
fitting. A statistical software, STATISTICA (Statsoft,
Tulsa, OK, USA) was used for the logistic model fitting.
Significant responses were defined as those greater than
three times the standard deviation of the response
obtained from solvent control.
2.8. Determination of E2 equivalents derived from
chemical analysis (E2-EQC)
We assumed that the estrogenic responses measured in
the bioassay were additive over different substances
present in the samples. The E2 equivalent quantity from
chemical analysis, expressed as E2-EQC (ng E2-eq/L),
was calculated by the multiplication of measured
concentrations with REP and subsequent summation.
Based on this calculation, we can identify, quantitatively, how much the target substances are contributing
to the total estrogenic activities in the samples.
3. Results and discussion
3.1. Recovery of target substances
The recoveries of target substances through the
extraction procedure using solid phase cartridge were
examined. The recoveries of target compounds spiked to
water samples were between 81.8% and 108% except OP
(28%) (n=4), for which recoveries were low (Table 1).
3.2. Estrogenic activity in crude extract
The estrogenic activity of crude water samples
measured using MVLN cell bioassay ranged between
0.45 and 0.93 of the maximum induction elicited by E2
(Fig. 2). Particularly, Kitatama No. 1 STP effluent
showed the highest estrogenic activity, which was as
great as 0.93. Nakamura and Iguchi (1998) reported the
presence of more female roaches relative to males, and
that male roaches had high concentrations of VTG and
abnormal testis, at locations downstream of Kitatama
No. 1 STP effluent. This may suggest that estrogenic
Table 1
Mean percentage recovery of target substances (abbreviation)
in pure water (n=4)
Target substance
(abbreviation)
Percentage
recovery (%)
C.V. (%)
Estrone (E1)
17b-Estradiol (E2)
Estriol (E3)
Ethnylestradiol (EE2)
Bisphenol A (BPA)
4-Nonylphenol (NP)
4-t-Octylphenol (OP)
95.5
104.6
94.5
95.9
103.3
83.5
28.0
7.4
8.8
1.8
5.1
13.6
6.3
17.5
ARTICLE IN PRESS
T. Furuichi et al. / Water Research 38 (2004) 4491–4501
4495
Fig. 2. Dose–response curves of estrogenic activity for 17b-estradiol (E2) and a crude extract of Tama River water. Data points
represent the mean value of three measurements; error represents the standard deviation.
Fig. 3. Estrogenic activity of each fraction from Tama River sample by MVLN assay. Activity represents the mean value of triplicate
measurements. The activity in the samples is shown relative to 17b-estradiol maximum response. Samples in well were 3.3 times
concentrated against their original volume.
chemicals in Kitatama No. 1 STP effluent may impact
the local fish population.
3.3. The estrogenic activity of GPC fractionated samples
To identify the potential contribution of known
estrogenic substances to the estrogenic activity in the
water samples, all crude extracts were fractionated into
10 fractions (Fig. 1). The estrogenic activity was
observed only in Fraction nos. 7 (F7), 8 (F8), and 9
(F9) in all of the sample extracts (Fig. 3). Thus, the
substances responsible for estrogenic activity are expected to be present in F7, F8 and F9. Some other
fractions, such as Hino Bridge F1–F6 and Maruko
Bridge F1–F6, showed negative estrogenic activity
compared with the activity observed in solvent control
samples. This might be an indication of existence of
inhibiting substances in these fractions. When the target
substances spiked to water samples were fractionated by
the GPC column procedure, they were separated as
follows: E1 in F8 and F9; E2 in F7 and F8; E3 in F7;
BPA in F7 and F8; EE2 in F8 (Fig. 1). This indicates
that E1, E2, E3, EE2, and BPA might be contributing
to the observed estrogenic activity. No activity
was observed in F6 of the samples (E2-EQBo0.6 ng
E2-eq/L).
Concentrations of target substances found in each
fraction, as determined by LC/MS or LC/MS/MS, are
listed in Table 2. NP and OP were detected in all of the
F6 fractions at a concentration range of 51.6–147 and
6.9–81.9 ng/L, respectively. The REPs (potencies relative
to 17b-estradiol) for NP and OP measured by MVLN
ARTICLE IN PRESS
4496
T. Furuichi et al. / Water Research 38 (2004) 4491–4501
Table 2
Concentrations of target substances in river water samples (ng/L)
Sample site
Target substance
Concentration (ng/L)
F6
F7
F8
F9
Total concentration
Hino Bridge
E1
E2
E3
EE2
OP
NP
BPA
—
—
—
—
20.7
78.0
—
—
1.5
o0.2
—
—
—
11.7
10.6
1.1
—
o0.2
—
—
4.8
17.2
—
—
—
—
—
—
27.8
2.6
o0.2
o0.2
20.7
78.0
16.5
Sekido Bridge
E1
E2
E3
EE2
OP
NP
BPA
—
—
—
—
6.9
51.6
—
—
3.2
o0.2
—
—
—
10.1
6.4
0.5
—
o0.2
—
—
17.9
10.7
—
—
—
—
—
—
17.1
3.7
o0.2
o0.2
6.9
51.6
28.0
Kitatama No.1 STP Effluent
E1
E2
E3
EE2
OP
NP
BPA
—
—
—
—
81.9
147.0
—
—
12.3
o0.2
—
—
—
12.1
22.0
2.4
—
o0.2
—
—
49.5
85.6
—
—
—
—
—
—
107.6
14.7
o0.2
o0.2
81.9
147.0
61.7
Hara Bridge
E1
E2
E3
EE2
OP
NP
BPA
—
—
—
—
19.1
96.3
—
—
4.4
o0.2
—
—
—
8.2
21.3
0.8
—
o0.2
—
—
25.0
26.3
—
—
—
—
—
—
47.6
5.2
o0.2
o0.2
19.1
96.3
33.2
Maruko Bridge
E1
E2
E3
EE2
OP
NP
BPA
—
—
—
—
47.5
144.0
—
—
5.9
o0.2
—
—
—
73.9
18.7
1.8
—
o0.2
—
—
76.3
25.8
—
—
—
—
—
—
44.5
7.7
o0.2
o0.2
47.5
144.0
150.2
Refer to Table 1 for abbreviations of target substances.
cells were 2.8 106 and 6.7 106, respectively. The
concentrations of NP and OP multiplied by the REP,
were summed to obtain E2-EQC. The E2-EQC of NP
and OP in all of the samples ranged from 1.9 104 to
9.6 104 ng E2-eq/L, which were below the sensitivity
of MVLN cells (o0.6 ng E2-eq/L), and consistent with
the results of bioassay E2-EQB measured by MVLN
cells. This suggested that NP and OP were not
responsible for the estrogenic activity measured by
MVLN cells. A similar observation reported that NP
and OP accounted for less than 1% of the total
concentration of E2-EQB in water sample collected
from river, STP effluent and lake in USA (Snyder et al.,
2001).
E2-EQB of F7 were calculated by the logistic model,
and ranged from 0.65 to 7.2 ng E2-eq/L (Fig. 4). E2
and BPA were detected in F7 at a concentration range
of 1.5–12.3 and 8.2–73.9 ng/L, respectively. E3 concentrations were below the detection limit (o0.2 ng/L)
(Table 2). The median concentrations of E2-EQC in F7
were 1.2–4 times higher than the E2-EQB. This suggests
that interfering compounds might have inhibited the
response of MVLN cells. The E2-EQC of BPA in F7
obtained by multiplying its REP (BPA: 6.1 10 7) were
ARTICLE IN PRESS
T. Furuichi et al. / Water Research 38 (2004) 4491–4501
4497
Fig. 4. Comparison of 17b-estradiol equivalents between chemical analysis (E2-EQC) and bioassay analysis (E2-EQB) from fraction 6
(F6) to fraction 9 (F9) with 95% CF.
5.0 105–4.5 104 ng E2-eq/L, which was below the
detection limits of bioassay-derived E2-EQB (o0.6 ng
E2-eq/L). Therefore, E2 seems to be the major
contributor to the estrogenic activity in F7. The
E2-EQB in F8 were in the range of 0.48–8.3 ng E2-eq/L
(Fig. 4). Concentrations of E2, E1 and BPA in F8 were
between 0.8 and 2.4, 6.4 and 22.0 ng/L, and 4.8 and
76.3 ng/L, respectively (Table 2). EE2 was not detected
in F8 (o0.2 ng/L) (Table 2). In order to compare the
E2-EQB and E2-EQC in F8, we obtained the REP value
of E1 and its 95% CF relative to E2 (E1: mean=0.19,
95% CF=0.16, +95% CF=0.22). The E2-EQC values
were from the similar level as E2-EQB to that of 3 times
higher than the E2-EQB (Fig. 4). The observed trends
in F8 were quite similar to those in F7. The ratios of
E2-EQC from E1 and E2, based on the chemical analysis,
were between 1.8 and 5. Therefore, both E2 and E1 are
responsible for the estrogenic activity in F8. The E2-EQC
of BPA was 2.9 105–4.7 104 ng E2-eq/L, which
was below the detection limits of bioassay analysis
ARTICLE IN PRESS
T. Furuichi et al. / Water Research 38 (2004) 4491–4501
4498
(o0.6 ng E2-eq/L). In F9, E1 was detected at a
concentration range of 10.7–85.6 ng/L (Table 2). The
E2-EQB was 1–2.3 times higher than the E2-EQC in F9
of all the samples, except for a sample from Hino Bridge,
which was 1.4 times less than the E2-EQC (Fig. 4).
The presence of unknown estrogenic compounds in F9
could be the reason of difference between E2-EQC and
E2-EQB.
3.4. The comparison of E2 equivalents derived from
chemical and bioassay analysis
The E2-EQB and E2-EQC in F7, F8 and F9 obtained
from bioassay and chemical analysis, respectively, were
summed up to determine Total-E2-EQB and Total-E2EQC and then the E2-EQC of E2, and E1 were also
summed up, respectively, to calculate the contribution of
E2 and E1 to Total-E2-EQB. The ratio of the contribution of E1 and E2 to Total-E2-EQB was 0.49–0.97 and
0.29–1.12, respectively (Table 3). Therefore, among all
the target substances analyzed, E1 and E2 are the most
contributing substances for the estrogenic activity
observed in these samples. Snyder et al. (2001) and
Körner et al. (2001) suggested that E2 and EE2 were the
major contributing substances to estrogenic activity in
water samples. In this study, however, EE2 could not be
detected in water (o0.2 ng/L), because contraceptive
pills can be purchased only with a medical prescription
in Japan. Thus, EE2 may be present only at very low
levels in Japanese waters. In water samples collected
near Hara Bridge and Maruko Bridge, the ratio of the
Total-E2-EQB and Total-E2-EQC were 0.85 and 0.94,
respectively (Total-E2-EQC/Total-E2-EQB) (Table 3).
On the other hand, the Total-E2-EQC was 1.4–2.1 times
higher than the Total-E2-EQB in Hino Bridge, Sekido
Bridge and Kitatama No. 1 STP effluent samples. In all
samples, the Total-E2-EQB was 1.1–2.7 times higher
than the bioassay E2-EQB: Crude extract, which was
calculated from the dose–response curve of crude extract
(Table 3). The results suggested that estrogenic substances and inhibiting substances were separated into
different fractions and that estrogenic substances were
not lost significantly during the fractionation procedure.
However, the E2-eq derived from bioassay (Total-E2EQB and E2-EQB in F7, F8) was lower than the E2-eq
calculated from (Total-E2-EQC and E2-EQC) in some
samples. Nagahora et al. (2001) reported that humics
suppressed estrogenic activity measured by in vitro yeast
reporter gene assay. The presence of estrogen antagonist
was suggested by the lack of eliciting estrogenic activity
with MVLN cells in river water and STP effluents
(Snyder et al., 2001). If these inhibitors existed in active
fractions such as F7, F8 and F9, even after fractionation, the result of the Total-E2-EQB and E2-EQB might
have been underestimated.
Körner et al. (2001) also reported that the E2-eq
calculated from GC/MS result was higher than that
determined by in vitro bioassay (MCF-7) by a factor of
two to four in most STP effluent samples. Although the
cumulative estrogenic response can be estimated based
on the concentration of well-defined estrogenic substance mixtures by using MCF-7 bioassay in laboratory
experiment (Charles et al., 2002), the presence of
chemical mixtures in environmental water may cause
some deviation from mass balance approach with
bioassay and chemical analysis.
Villeneuve et al. (2000) suggested that mass balances
analysis between chemical analysis and in vitro bioassay
Table 3
17b-estradiol (E2) equivalent quantity derived from bioassay analysis (Total-E2-EQB) and chemical analysis (Total-E2-EQC), which
was summed from fraction no. 7 (F7) to fraction no. 9 (F9)
Sample site
E2 equivalents (E2-eq) derived from chemical analysis (E2-EQC) and
bioassay analysis (E2-EQB) (ng E2-eq/L)
Contribution ratio of E1, E2 and Total-E2EQC to Total-E2-EQB
E2-EQB:
E1ðE2 EQC Þ
Total E2 EQB
E2ðE2 EQC Þ
Total E2 EQB
Total E2 EQC
Total E2 EQB
Crude extract
Hino Bridge
Sekido
Bridge
Kitatama
No.1 STP
Effluent
Hara Bridge
Maruko
Bridge
Total-E2EQB
E1 (E2EQC)
E2 (E2EQC)
Total-E2EQC
- -
- -
- - -
2.7
1.2
5.4
3.3
5.2
3.2
2.6
3.7
7.8
6.9
0.96
0.97
0.48
1.12
1.44
2.09
23.6
25.9
20.4
14.7
35.1
0.79
0.57
1.36
8.3
10.2
16.7
17.2
9.0
8.4
5.2
7.7
14.2
16.1
0.54
0.49
0.31
0.45
0.85
0.94
Contribution ratio of estrone (E1), E2 and Total-E2-EQC to Total-E2-EQB. E2-EQB: Crude extract was derived from the dose–response
curve of crude extract sample.
ARTICLE IN PRESS
T. Furuichi et al. / Water Research 38 (2004) 4491–4501
using a single point estimate such as EC50 may lead to
erroneous conclusions because of the differences in the
slopes of dose–response curves between standard and
sample. Although we did not find a significant difference
in the slopes of the dose–response curves between E2
and samples in this study, the application of the logistic
model may provide better estimates of E2-EQB.
In the previous study, we reported that the E2-EQC
were 10–1000 times higher than those calculated based
on the bioassay in Tama River water samples (Masunaga et al., 2000). The earlier study was conducted using
crude extracts without fractionation. On the other hand,
in this study, the E2-EQC was 1.5–5.6 times higher than
E2-EQB: Crude extract. The difference of experimental
method, namely with and without fractionation may be
the cause of the different results. Further research is
needed to assess the mass balance comparison of
bioassay and chemical analysis for samples with complex environmental matrix.
In this study, with the use fractionation, we obtained a
better comparison between E2-EQB and E2-EQC,
indicating that the co-existence of estrogenic compounds
and/or inhibiting compounds in the samples might affect
the resolution and mass balance analysis.
3.5. The levels of E2 and E1
The Tama River has nine STPs in its drainage area,
which have facilities for secondary treatment processes
such as activated sludge. Some of the plants have even
more advanced treatment processes such as ozone
treatment. Concentrations of E1 in F8 and F9 were
higher than that of E2, with a range of 17.1–107.6 ng/L
for E1, 2.6–14.7 ng/L for E2, respectively (Table 2).
Similar observations have been made not only in river
water and STP effluents in Japan (Ministry of Land,
Infrastructure and Transport, 2001; Isobe et al., 2003)
but also in other countries (Baronti et al., 2000; Korpin
et al., 2002). Ternes et al. (1999) reported that E2 was
transformed into E1 in activated sludge. E2-3-glucuronide (E2-3G) and E2-17-glucuronide (E2-17G), which
are metabolite of natural estrogens, were transformed to
E2 and then immediately to E1 in an experimental
activated sludge process. Panter et al. (1999) reported
that male fathead minnows exposed to E2-3G did not
exhibit significant VTG induction. However, male fathead minnows that were exposed to water after the
activated sludge treatment or the continuous-flow
systems, showed a significant induction of VTG. Source
of E1 could be oxidation of E2 and also transformation
from E1-3sulfate (E1-3S) and E1-3G. These observations suggest that transformation of E2 to E1 may occur
in STPs in this study.
E2 concentrations observed in this study were less
than the threshold for VTG induction in adult male
rainbow trout (410 ng/L) (Routledge et al., 1998),
4499
except those from Kitatama No. 1 STP effluent. The
concentration of E1 in Kitatama No. 1 STP effluent,
Hara Bridge and Maruko Bridge were high enough to
induce VTG in adult rainbow trout (444 ng/L) (Routledge et al., 1998). Although threshold E1 concentration for VTG induction was higher than E2, presence of
E1 is of concern because the concentration of E1 was
much higher than that of E2 in the Tama River. Thorpe
et al. (2003) reported that female juvenile rainbow trout
exposed to a binary mixture of E2 and EE2 induced
VTG in an additive manner. In addition to the
estrogenic activity of the individual compounds, additive
or synergistic effect of these estrogenic compounds
should be taken into account. In this context, it is
important to examine total estrogenic activity rather
than measuring the concentration of individual estrogenic compounds in river water. A positive correlation
between E2-EQC, which was calculated from the
concentrations of E1 and E2 and VTG levels in male
carps, was reported in a field study (Wanami, 2002).
About 13% of atrophy of testes has been reported for
male carps in the Tama River (Wanami, 2002).
Although there is no conclusive evidence of a cause–effect relationship between estrogenic substances and
atrophy of testes, it is important to understand the
levels of estrogenic substances in waters to assess the
effect on biota.
4. Conclusion
To quantitatively characterize the substances contributing to estrogenic activity in river water, in vitro
bioassay using MVLN cells and instrumental analysis
using LC/MS or LC/MS/MS were applied to river water
extracts taken from various locations in the Tama River,
Japan. Tama River water samples were extracted using
solid phase extraction and crude extracts were fractionated by HPLC into the various fractions, after which
the samples were tested for estrogenic activity in MVLN
assay. The estrogenic activity in fractionation sample
determined by MVLN assay was compared to a
predicted estrogenic response based on quantitative
measurements of the target substances in the fraction
and their relative potencies. The sixth fraction contained
NP and OP at concentrations in the ranges of 51.6–147
and 6.9–81.9 ng/L, respectively. However, no estrogenic
activity was observed in this fraction (o0.6 ng-E2eq/L).
Instrumentally determined estrogenic activity (E2-EQC),
which is the concentrations of NP and OP multiplied by
their corresponding relative potency, was below the
detection limit of the MVLN cell bioassay. Estrogenic
activities were detected in HPLC fraction nos. 7, 8 and 9.
E1, E2 were detected in these fractions with a range
of 17.1–107.6 ng/L, 2.6–14.7 ng/L, respectively. The
E2-EQC for E1 and E2 were on the same order as the
ARTICLE IN PRESS
4500
T. Furuichi et al. / Water Research 38 (2004) 4491–4501
estrogenic activity determined by the bioassay (E2EQB). The ratios of E2-EQC and E2-EQB for E1 and E2
in the three fractions collectively (nos. 7–9) were
0.49–0.97 and 0.29–1.12, respectively. Estriol (E3) and
ethynylestradiol (EE2) were not detected (o0.2 ng/L) in
these fractions. Although BPA were detected in these
fractions, the calculated E2-EQC for BPA was below the
detection limit of bioassay (o0.6 ng-E2eq/L). Above
results indicated that the major causal substances to the
estrogenic activity in the Tama River were E1 and E2.
Acknowledgments
The financial supports from Grant-in-Aid for the
Creation of Innovations through Business-AcademicPublic Sector Cooperation (no. 12323) and The 21st
Century COE Program ‘‘Bio-Eco Environmental Risk
Management’’, Ministry of Education, Culture, Sports,
Science and Technology are much appreciated. Thanks
are also given for help with chemical analysis to Yoshiaki
Ishii of Environmental Control Center, Co. Ltd.
Reference
Baronti, C., Curini, R., D’Ascenzo, G., Corcia, A.D., Gentili,
A., Samperi, R., 2000. Monitoring natural and synthetic
estrogens at activated sludge sewage treatment plants and in
a receiving river water. Environ. Sci. Technol. 34,
5059–5066.
Charles, G.D., Gennings, C., Zacharewski, T.R., Gollapudi,
B.B., Carney, E.W., 2002. Assessment of interactions of
diverse ternary mixtures in an estrogen receptor-a receptor
assay. Toxicol. Appl. Pharmacol. 180, 11–21.
Desbrow, C., Routledge, E.J., Brighty, G.C., Sumpter, J.P.,
Waldock, M., 1998. Identification of estrogenic chemicals
in STW effluent. 1. Chemical fractionation and in
vitro biological screening. Environ. Sci. Technol. 32,
1549–1558.
Folmar, L.C., Denslow, N.D., Rao, V., Chow, M., Crain, D.A.,
Enblom, J., Marcino, J., Guillette Jr., L.J., 1996. Vitellogenin induction and reduced serum testosterone concentrations in feral male carp (Cyprinus carpoi) captured near a
major metropolitan sewage treatment plant. Environ.
Health Perspect. 104, 1096–1101.
Harries, J.E., Sheahan, D.A., Jobling, S., Matthiessen, P.,
Neall, P., Routledge, E.J., Rycroft, R., Sumpter, J.P., Tylor,
T., 1996. Estrogenic activity in five United Kingdom rivers
detected by measurement of vitellogenesis in caged male
trout. Environ. Toxicol. Chem. 15, 1993–2002.
Harries, J.E., Sheahan, D.A., Jobling, S., Matthiessen, P.,
Neall, P., Sumpter, J.P., Tylor, T., Zaman, N., 1997.
Estrogenic activity in five United Kingdom rivers detected
by measurement of vitellogenesis in caged male trout.
Environ. Toxicol. Chem. 16, 534–542.
Ishii, Y., Okita, S., Torigai, M., Yun, S., 2000. Determination
of estrogens in environmental water samples by LC/MS/
MS. Jpn. Soc. Anal. Chem. 49, 753–758 (in Japanese).
Isobe, T., Shiraishi, H., Yasuda, M., Shinoda, A., Suzuki, H.,
Morita, M., 2003. Determination of estrogens and their
conjugates in water using solid-phase extraction followed by
liquid chromatography-tandem mass spectrometry. J. Chromatogr. A 984, 195–202.
Jobling, S., Nolan, M., Tyler, C.R., Brighty, G., Sumpter, J.P.,
1998. Widespread sexual disruption in wild fish. Environ.
Sci. Technol. 32, 2498–2506.
Körner, W., Spengler, P., Bolz, U., Sbhuller, W., Hanf, V.,
Metzger, J.W., 2001. Substances with estrogenic activity in
effluents of sewage treatment plants in southwestern
Germany. 2. Biological analysis. Environ. Toxicol. Chem.
20, 2142–2151.
Korpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M.,
Zaugg, S.D., Barber, L.B., Buxton, H.T., 2002. Pharmaceutical, hormones, and other organic wastewater contaminations in US streams, 1999–2000: a national
reconnaissance. Environ. Sci. Technol. 36, 1202–1211.
Masunaga, S., Itazawa, T., Furuichi, T., Sunardi, Villeneuve,
D., Kannan, K., Giesy, J.P., Nakanishi, J., 2000. Occurrence of estrogenic activity and estrogenic compounds in the
Tama River, Japan. Environ. Sci. 7, 101–117.
Ministry of Land, Infrastructure and Transport, 2001. Results
of the survey on the endocrine disrupting chemicals in water
environment for fiscal year 2000. Ministry of Land,
Infrastructure and Transport Government of Japan
(in Japanese).
Nagahora, S., Aga, H., Numabe, A., Murata, K., Sakata, K.,
2001. The influence of humic substance to estrogenicity
assay in environmental samples by using solid phase
extraction–yeast two-hybrid system. The 10th Symposium
on Japan Environmental Chemistry. Program and Abstracts, pp. 324–325 (in Japanese).
Nakamura, M., Iguchi, T., 1998. Abnormal fish in Tama River.
Sci. J. KAGAKU 68, 515–517 (in Japanese).
Nakagawa, T., Koyanagi, Y., 1992. Analysis of Experimental
Data by Least-squares Method. University of Tokyo Press,
Tokyo (in Japanese).
Panter, G.H., Thompson, R.S., Beresford, N., Sumpter, J.P.,
1999. Transformation of a non-oestrogenic steroid metabolite to an oestrogenically active substance by minimal
bacterial activity. Chemosphere 38, 3579–3596.
Pons, M., Gagne, D., Nicolas, C.J., Mehtali, M., 1990. A new
cellular model of response to estrogens: a bioluminescent
test to characterize (anti)estrogen molecules. Biotechniques
9, 450–459.
Purdom, C.E., Hardiman, P.A., Bye, V.J., Eno, N.C., Tyler,
C.R., Sumpter, J.P., 1994. Estrogenic effects of effluents
from sewage treatment works. Chem. Ecol. 8, 275–285.
Routledge, E.J., Sheahan, D., Desbrow, C., Brighty, G.C.,
Waldock, M., Sumpter, J.P., 1998. Identification of
estrogenic chemicals in STW effluent. 2. In vivo responses
in trout and roach. Environ. Sci. Technol. 32, 1559–1565.
Snyder, S.A., Villeneuve, D.L., Snyder, E.M., Giesy, J.P., 2001.
Identification and quantification of estrogenic receptor
agonist in wastewater effluents. Environ. Sci. Technol. 35,
3620–3625.
Ternes, T.A., Kreckel, R., Mueller, J., 1999. Behavior and
occurrence of estrogens in municipal sewage treatment
plants—II. Aerobic batch experiments with activated
sludge. Sci. Total Environ. 225, 91–99.
ARTICLE IN PRESS
T. Furuichi et al. / Water Research 38 (2004) 4491–4501
Thorpe, K.L., Cummings, R.I., Hutchinson, T.H., Scholze, M.,
Brighty, G., Sumpter, J.P., 2003. Relative potencies and
combination effects of steriodal estrogen in Fish. Environ.
Sci. Toxicol. 37, 1142–1149.
Villeneuve, D.L., Blankenship, A.L., Giesy, J.P., 2000. Derivation
and application of relative potency estimates based on in vitro
bioassay result. Environ. Toxicol. Chem. 19, 2835–2843.
4501
Wanami, K., 2002. Problems of endocrine disrupting
chemicals in river basin, Tokyo. Study of feminization
of male carp in Tama River. Water Inf. 22, 13–17
(in Japanese).
Yanagawa, T., 2002. Environmental and Health Data—Data
Science of Risk Assessment. Kyoritsu Publishing Co. Ltd.,
Tokyo (in Japanese).