Identification and Quantification of Estrogen Receptor Agonists in ...

Identification and Quantification of
Estrogen Receptor Agonists in
Wastewater Effluents
SHANE A. SNYDER,*
DANIEL L. VILLENEUVE,
ERIN M. SNYDER, AND JOHN P. GIESY
Department of Zoology, National Food Safety and
Toxicology Center, and Institute for Environmental
Toxicology, Michigan State University,
East Lansing, Michigan 48824-1311
Total concentrations of several known xenobiotic estrogen
receptor (ER) agonists and natural and synthetic estrogen
were measured in water by use of a combination of
instrumental and bioanalytical approaches. Samples from
3 municipal wastewater treatment plants (WWTPs) in
south central Michigan (upstream and effluent); 4 point
source locations on the Trenton Channel of the Detroit River,
MI; and 5 locations in Lake Mead, NV were analyzed.
Organic compounds were extracted from 5 L water samples
using solid-phase extraction disks and separated into
three fractions based on polarity. Whole extracts and
fractions were tested for ER agonist potency using the
MVLN in vitro bioassay. ER agonist potency was characterized
by comparing the magnitude of induction elicited by the
extract or fraction to the maximum induction caused by 17estradiol
(E2). The greatest concentrations of ER agonists
were associated with the most polar fraction (F3).
Instrumental analyses and further fractionation were used
to identify specific ER agonists associated with bioassay
responses. Bioassay data were compared to extract
concentrations in order minimize variability associated
with the extraction procedure. Concentrations of endogenous
estrogen, E2, and the synthetic estrogen ethynylestradiol
(EE2) ranged from nondetectable to 14.6 ng/mL extract
(nondetectable to 3.66 ng/L water) and represented from
88 to 99.5% of the total estrogen equivalents in the water
samples analyzed. Concentrations of alkylphenols (APs)
ranged from nondetectable to 148 µg/mL extract (nondetectable
to 37 000 ng/L water). In general, alkylphenols
contributed less than 0.5% of the total estrogen equivalents
in the water samples. Both bioassay-directed fractionation
results and comparison of ER agonist concentrations,
adjusted for their known relative potencies, support the
conclusion that E2 and EE2 were the dominant environmental
estrogens in water samples from mid-Michigan and Lake
Mead, NV.
Introduction
Some compounds released into the environment by human
activities can mimic or modulate endogenous hormones and
have been termed “endocrine-disrupting” compounds (1,
* Corresponding author phone: (702)567-2317; fax: (702)564-7222;
e-mail: shane.snyder@lvvwd.com. Current address: Southern Nevada
Water Authority, 243 Lakeshore Road, Boulder City, NV 89005.
Environ. Sci. Technol. 2001, 35, 3620-3625
2). “Endocrine-disrupting” compounds have been defined
as exogenous agents that interfere with the “synthesis,
secretion, transport, binding, action, or elimination of natural
hormones in the body that are responsible for the maintenance
of homeostasis, reproduction, development, and/or
behavior” (3). It has been hypothesized that such compounds
may elicit a variety of adverse effects in both humans and
wildlife, including promotion of hormone-dependent cancers,
reproductive tract disorders, and reduction in reproductive
fitness (1, 4-10). Much of the concern has focused
on compounds that are estrogen receptor (ER) agonists. These
compounds have been variously referred to as “estrogenic”,
“estrogen-like”, “environmental estrogens”, or “xenoestrogens”.
ER agonists and antagonists have the ability to mimic
or block the functions of endogenous estrogen. Effects
consistent with exposure to ER agonists have been observed
in fish exposed to municipal wastewater treatment plant
effluents (11, 12). Nonylphenol (NP), nonylphenol polyethoxylates
(NPEs), octylphenol (OP), and synthetic and
natural steroids were targeted in this investigation because
they are known to be present in wastewater effluents and
have been implicated as ER agonists that can cause adverse,
population-level effects in aquatic organisms (11, 13-16).
Methods for identifying and quantifying ER agonists in
environmental samples are needed in order to assess the
potential for adverse effects through an ER-mediated mechanism
of action. This need was underscored by recent
legislation mandating that chemicals and formulations be
screened for potential to cause estrogen-like biological
responses before they are manufactured or used in certain
processes (Safe Drinking Water Act Amendments of 1995 -
Bill Number S.1316; Food Quality Protection Act of 1996 -
Bill Number P.L. 104-170).
Halogenated aromatic hydrocarbons (HAHs) and polycyclic
aromatic hydrocarbons (PAHs) are known to cause a
wide range of adverse effects, including mortality, wasting
syndrome, hepatotoxicity, immunotoxicity, reproductive
impairment, and carcinogenicity (16-19). Some of these
effects are mediated through the aryl hydrocarbon receptor
(AhR) (17); however, some of these compounds can modulate
the ER as well. HAHs, such as polychlorinated dibenzo-pdioxins
(PCDDs), polychlorinated dibenzofurans (PCDFs),
and some polychlorinated biphenyls (PCBs), have been
reported to act as ER agonists in vitro (20, 21). PAHs have
been reported to be both ER agonists and antagonists in
vitro (22-23).
Although instrumental analyses can be used to identify
and quantify known ER agonists and antagonists in wastewater
treatment plant (WWTP) effluents, in vitro bioassays
provide useful information that can complement instrumental
analyses to provide a more comprehensive characterization
of a sample’s potential to modulate the ER and
result in estrogenic responses. In vitro bioassays provide an
integrated measure of the total potency of complex mixtures
to induce particular biological responses. Thus, in vitro
bioassays can account for both unknown compounds and
potential nonadditive interactions among compounds. This
study is based on a bioassay-directed fractionation approach
to identify compounds able to modulate ER-mediated gene
expression. Furthermore, bioassay-based estimates of total
ER agonist potency were compared to estimates based on
analytical concentrations of known ER agonists and their
relative potencies (REPs) in a potency balance analysis (24,
25) to determine whether the compounds quantified could
account for the magnitude of ER-mediated bioassay response
observed.
3620 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 18, 2001 10.1021/es001254n CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/14/2001

FIGURE 1. Luciferase induction in the MVLN cell bioassay (estrogen responsive) elicited by water extracts. Response magnitude presented
as percentage of the average maximum response observed for a 1000 pM 17-estradiol standard (%-E2-max). Horizontal lines represent
( 3 SD from the mean solvent control response (set to 0%-E2-max): a. Michigan WWTPs; b. Trenton Channel; c. Lake Mead (April); and
d. Lake Mead (September).
Materials and Methods
Sample Collection and Fractionation. A detailed description
of the analytical methodology used in this study was
published previously (26). Briefly, 5 L water samples were
extracted at each field site using solid-phase extraction (SPE)
Empore disks. Organic extracts from these SPE disks were
separated into three fractions based on polarity using normalphase
high-pressure liquid chromatography (NP-HPLC) (26).
In some cases, ER agonists of interest were isolated from F2
and F3 by fractionating these again with reverse-phase HPLC
(RP-HPLC), with fractions collected approximately every 3
min (Supporting Information, Figure 1). NP-HPLC and RP-
HPLC separations were accomplished using silica and C18
analytical columns, respectively. Quality assurance and
quality control measures included replicate samples, field
and laboratory blanks, and spike-recovery experiments,
which were described in detail previously (26).
Cell Culture and Bioassay. An MCF-7 human breast
carcinoma cell line, stably transfected with an ER-controlled
luciferase reporter gene construct (MVLN or MCF-7-luc cells),
was developed and characterized by Dr. M. D. Pons, Institut
National de la Sante et de la Recherche Medicale (27). MVLN
cells were cultured in 75-cm 2 disposable polyethylene tissue
culture flasks (Corning, Corning, NY) containing 20-25 mL
of Dulbecco’s Modified Eagle Medium (DMEM) with Hams
F-12 nutrient mixture (Sigma D-2906; St. Louis, MO) supplemented
with 10% defined fetal bovine serum (Hyclone, Logan,
UT), 27.3 I.U. insulin (Sigma I-1882)/L, and 1.0 mM sodium
pyruvate (Sigma).
In preparation for bioassay, cells were trypsinized from
flasks or plates in which cells were 80-100% confluent. The
number of cells per mL was determined microscopically by
FIGURE 2. Fine fractionation of LV Wash, Lake Mead (April), F3
extract using RP-HPLC with fluorescence detection followed by
luciferase induction in the MVLN cell bioassay (estrogen responsive)
by the corresponding fractions. Response magnitude presented as
percentage of the average maximum response observed for a 1000
pM 17-estradiol standard (%-E2-max). Horizontal lines represent
( 3 SD from the mean solvent control response (set to 0%-E2-max).
use of a hemacytometer. MVLN cells were diluted in
hormone-stripped medium [DMEM with Hams F-12 nutrient
mixture, supplemented with 10% dextran-coated charcoal
filtered fetal bovine serum (Hyclone), 27.3 I.U. insulin (Sigma
I-1882)/L, and 1.0 mM sodium pyruvate (Sigma)] to a
concentration of approximately 1.5 × 10 5 cells/mL. Cells were
seeded into the 60 interior wells of 96-well flat bottom
microplates (Packard Instruments 6005181; Meriden, CT) at
VOL. 35, NO. 18, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3621

125 µL per well (15 000-20 000 cells per well) using a
repeating pipet. To ensure homogeneity, the cell solution
was continuously mixed during seeding. The 36 exterior wells
of each microplate were filled with 125 µL of medium. Cells
were dosed after an overnight incubation to allow for cell
attachment. Extracts or fractions were dissolved in stripped
medium to yield a final concentration of 1.0% extract. A 3-fold
dilution of each extract or fraction was also prepared, yielding
a concentration of 0.33% extract. Test wells were dosed with
125 µL 1.0% or 0.33% extract in medium to yield final in-well
concentrations of 0.50% and 0.165% extract. Solvent control
wells were dosed with 125 µL of medium spiked with 1.0%
of the appropriate solvent to yield a final in-well concentration
of 0.50% solvent. Blank wells received 125 µL of the
appropriate media. Each plate tested included a minimum
of three solvent control wells, three blank wells, and three
replicates of each fraction tested (at both 0.50% and 0.165%
levels). Dosed cells were exposed for 72 h at standard
incubation conditions.
Each test plate was inspected visually and differences in
cell numbers and condition relative to control wells and
conditions normally observed during routine culturing were
noted for each well. Culture medium was then removed, and
each well was rinsed twice with phosphate buffered saline
(PBS) supplemented with 1.0 mM Ca 2+ and Mg 2+ using an
eight channel vacuum manifold. Plates were inspected for
cell loss during washing. Following inspection, 75 µL PBS
supplemented with Ca 2+ and Mg 2+ was added to each well,
followed by 75 µL Luc-lite reagent (Packard Instruments).
Each plate was incubated for 10 min at 30 °C and then scanned
with an ML 3000 microplate reading luminometer (Dynatech
Laboratories, Chantilly, VA). Following the luminometer scan,
125 µL of 1.08 mM fluorescamine (Sigma) in acetonitrile
(ACN) was added to each well, and plates were assayed for
protein after a 15 min incubation at room temperature (28).
Plates were scanned using a Cytofluor 2300 (excitation 400
nm, emission 460 nm), and responses were compared to a
standard curve consisting of six concentrations of bovine
serum albumin (BSA) (Sigma) ranging from 1.5 to 50 µg per
well.
All data were collected electronically and imported into
a spreadsheet (Excel 7.0, Microsoft Inc., Seattle, WA) for data
analysis. Protein content per well was calculated by regression
against the BSA standard curve. Protein data were used as
an index of cell number to detect outliers that were not
apparent by visual inspection. Relative luminescence units
(RLU) were not adjusted for protein. Sample responses in
RLU were expressed as a percentage of the mean maximum
response observed for standard curves developed on the same
day (% E2-max) (29). The greatest response of the two extract
dilutions was reported. However, for each significant response,
the greatest response came from the greater extract
concentration (0.5% in the well).
Potency balance analyses were conducted by comparing
observed bioassay response magnitudes to those predicted
based on the concentrations of known ER agonists present
in an extract (30). Instrumentally determined concentrations
of individual compounds were multiplied by their assayspecific
relative potencies. The sum of the products for all
target compounds present in an extract provided an estimate
of the 17-estradiol equivalents (EEQ) in the extracts. Linear
regression against a 17-estradiol (E2) standard curve was
used to predict the bioassay response magnitude for the
sample. Variability in the predicted bioassay response
magnitude was estimated based on the 95% confidence band
for a first-order polynomial fit to the E2 standard curve (PlotIT,
Scientific Programming Enterprises, Haslett, MI). Comparisons
were predicated on the assumption that EEQs would
behave as if they were 17-estradiol in the bioassay. Violation
of this assumption may have resulted in some error in the
3622 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 18, 2001
predictions (29). To make an accurate comparison, it was
necessary to address the potential for antagonistic and
synergistic interactions. This was done by the fine fractionation
of samples, by which compounds known to be
antagonistic to the measurement of EEQ were separated from
the active compounds.
Results and Discussion
ER Agonist Activity. None of the F1 fractions induced a
significant response in the MVLN assay (Figure 1). Nonpolar
compounds such as PAHs, PCBs, and most organochlorine
(OC) pesticides, if present, would have been contained in F1
(Figure 1) (26). Certain OC pesticides and PAHs such as
chrysene, benz[a]anthracene, and benzo[a]pyrene have been
reported to cause weak ER-mediated responses in vitro (23,
31). Based on the method detection limit (MDL) for E2 in the
MVLN assay, concentrations of ER agonists were present in
F1 at a concentration less than 0.55 ng EEQ/mL. These results
support the conclusion that concentrations of nonpolar ER
agonists in the surface waters and effluents examined were
small.
Weak ER agonists such as NP and OP were present in F2
(26). No F2 extracts elicited a significant response in the
MVLN assay (Figure 1), despite the confirmed presence of
NP and OP (Table 1). These results suggest that the
compounds present in F2 contributed less than 0.55 ng EEQ/
mL. ER agonist potencies of NP and OP, relative to E2, for
luciferase induction in MVLN cells have been reported to be
1.25 × 10 -5 and 1.9 × 10 -5 for NP and OP, respectively (25).
When concentrations of NP and OP present in the samples
(Table 1) were multiplied by their corresponding relative
potencies and summed, it was concluded that these two
compounds contributed less than 0.075 ng EEQ/mL for 15
of the 16 samples tested. Thus, the general lack of significant
induction of the MVLN cells was consistent with the known
concentrations of ER agonists (alkylphenols) present in F2.
The BV-effluent sample extract contained approximately 1.90
ng of EEQ/mL, which would correspond to approximately
8.5 fmol EEQ/well in the MVLN bioassay. Based on regression
against an E2 standard curve, this dose could have elicited
a response as great as 53% E2-max. However, F2 extract of
BV-effluent failed to induce a significant response in the
MVLN bioassay (Figure 1a). This suggests that F2 of the BVeffluent
sample might have contained interfering compounds
that suppressed the ER agonist potency of NP and OP. The
F2 extract was fractionated further, and fine fractions were
analyzed in the MVLN bioassay (Supporting Information,
Figure 2). No significant ER-mediated responses were
observed in the fine fractions. Thus, the hypothesized
interfering compounds, if present, must have properties
similar to NP. The same fine fractionation was applied to the
F2 extract from Black Lagoon (Supporting Information, Figure
3). Once again, no significant estrogen-like activity was
observed. In general, responses of MVLN cells to F2 samples
were in agreement with the potency expected based on the
known concentrations and relative potencies of these
compounds.
F3 samples caused the greatest magnitude of ER agonist
response in the MVLN bioassay. Six of the 16 F3 samples
elicited significant ER-mediated responses in the MVLN
bioassay (Figure 1). The greatest magnitudes of response for
F3 extracts (≈80% E2-max) were observed for samples
collected from the LV Wash and LV Bay in April 1997 (Figure
1c). However, samples from LV Wash and LV Bay collected
in September of 1997 did not elicit significant responses in
the MVLN assay (Figure 1d). The samples collected in
September 1997 were obtained after a large storm event,
which diluted the wastewater entering LV Wash and LV Bay
(26). The difference in bioassay responses for April and
September samples was paralleled by decreases in EEQs in

TABLE 1. Extract Concentrations and 17-Estradiol Equivalents (EEQs) (ng/mL) c
location date NP OP NPE E2 EE2 NP/OP-EEQ a E2/EE2-EEQ b
LV Wash 4/30/97 4560 172
Lake Mead
36000 10.70 1.92 0.061 10.9
LV Bay 4/30/97 3000 108 19400 8.84 2.08 0.040 9.05
9/5/97 640 ND 12710 0.752 1.01 0.008 0.86
LV Marina 9/5/97 ND ND ND 1.08 ND NA 1.08
Saddle Island 4/30/97 ND ND ND ND ND NA NA
Callville Bay 9/5/97 ND ND ND
Trenton Channel
ND ND NA NA
WWTP 8/30/97 1916 20 21600 4.26 ND 0.024 4.26
Chem. 8/30/97 3450 60 29200 3.64 ND 0.044 3.64
B. Lagoon 8/30/97 3740 264 34700 5.18 1.44 0.052 5.30
M. Creek 8/30/97 4740 324 71260
WWTPs
4.24 ND 0.066 4.25
BV-upstream 10/8/97 ND ND ND 2.50 ND NA 2.50
BV-effluent 10/8/97 148000 2350 1160000 14.6 3.04 1.90 14.9
MA-upstream 10/8/97 ND ND ND ND ND NA NA
MA-effluent 10/8/97 2065 64 19400 3.62 1.43 0.027 3.77
ER-upstream 10/8/97 ND ND ND ND ND NA NA
ER-effluent 10/8/97 680 ND ND 1.90 ND 0.009 1.90
a Nonylphenol and octylphenol-derived 17-estradiol equivalents. NP/OP-EEQ ) (NPrelative potency(REP) × NPconcentration) + (OPREP × OPconcentration).
NPREP ) 1.25 × 10 -5 .OPREP ) 1.9 × 10 -5 . A REP estimate was not available for NPE; therefore, it was not considered when deriving EEQ estimates.
b Estradiol and ethynylestradiol-derived 17-estradiol equivalents. E2/EE2-EEQ ) (E2REP × E2concentratration) + (EE2REP × EE2concentration). E2REP ) 1.0.
EE2REP ) 0.10. c ND ) not detectable; NA ) not applicable
FIGURE 3. Fine fractionation of LV Bay, Lake Mead (April), F3 extract
using RP-HPLC with fluorescence detection followed by luciferase
induction in the MVLN cell bioassay (estrogen responsive) by the
corresponding fractions. Response magnitude presented as percentage
of the average maximum response observed for a 1000 pM
17-estradiol standard (%-E2-max). Horizontal lines represent ( 3
SD from the mean solvent control response (set to 0%-E2-max).
the samples (Tables 1 and 2). The ER agonist potency of EE2
relative to E2 for luciferase induction in the MVLN assay
previously has been reported to be approximately 0.1 (25).
Based on the concentrations of E2 and EE2, and their
corresponding relative potencies, samples collected from LV
Wash and LV Bay in April 1997 were estimated to contain
10.9 and 9.05 ng E2/EE2-derived EEQ/mL, respectively. These
concentrations should have yielded doses of approximately
50 and 41 fmol EEQ/well in the MVLN bioassay. Based on
regression against an E2 standard curve, such doses would
be expected to yield responses of approximately 92% and
88% E2-max, respectively. Based on the range of uncertainty
in the predicted responses (Table 2) and the variability of the
observed bioassay responses (Figure 1c), the responses
observed for the samples collected from LV Wash and LV
Bay in April 1997 were not markedly different from predicted
responses. Thus, the known E2 and EE2 composition of F3
of the samples collected from LV Wash and LV Bay appeared
to account for all the ER agonist potency observed. Additional
TABLE 2. Extract 17-Estradiol Equivalents (EEQs) (ng/mL) and
Predicted MVLN Responses e
location date
NP/OP-
EEQ a
predicted
response b
E2/EE2-
EEQ c
predicted
response d
LV Wash 4/30/97
Lake Mead
0.061 0 10.9 92
LV Bay 4/30/97 0.040 0 9.05 88
9/5/97 0.008 0 0.86 35
LV Marina 9/5/97 NA NA 1.08 41
Saddle Island 4/30/97 NA NA NA NA
Callville Bay 9/5/97 NA NA NA NA
Trenton Channel
WWTP 8/30/97 0.024 0 4.26 71
Chem. 8/30/97 0.044 0 3.64 68
B. Lagoon 8/30/97 0.052 0 5.30 76
M. Creek 8/30/97 0.066
WWTPs
0 4.25 71
BV-upstream 10/8/97 NA NA 2.50 59
BV-effluent 10/8/97 1.90 53 14.9 99
MA-upstream 10/8/97 NA NA NA NA
MA-effluent 10/8/97 0.027 0 3.77 68
ER-upstream 10/8/97 NA NA NA NA
ER-effluent 10/8/97 0.009 0 1.90 53
a Nonylphenol and octylphenol-derived 17-estradiol equivalents.
NP/OP-EEQ ) (NPrelative potency (REP) × NPconcentration) + (OPREP × OPconcentration).
NPREP ) 1.25 × 10 -5 .OPREP ) 1.9 × 10 -5 . A REP estimate was not available
for NPE; therefore, it was not considered when deriving EEQ estimates.
b MVLN bioassay response magnitudes predicted based on regression
of NP/OP-derived EEQ against a 17-estradiol standard curve. Units are
%E2-max. c Estradiol and ethynylestradiol-derived 17-estradiol equivalents.
E2/EE2-EEQ ) (E2REP × E2concentration) + (EE2REP × EE2concentration).
E2REP ) 1.0. EE2REP ) 0.10. d MVLN bioassay response magnitudes
predicted based on regression of E2/EE2-derived EEQ against a 17estradiol
standard curve. Units are %E2-max. e NA ) not applicable.
Note: total EEQ ) NP/OP-EEQ + E2/EE2-EEQ. Predicted bioassay
response magnitudes are not additive.
fractionation of the F3 extracts from LV Wash and LV Bay
revealed that all of the ER agonist potency was associated
with the fine fractions (FFs) 3 and 4, which equate roughly
to the retention times of E2 and EE2 (Figures 2 and 3). FFs
3 and 4 from the F3 extract of the LV Wash sample were
collected, combined, and fractionated again by RP-HPLC
using a slower flow rate and solvent gradient to separate E2
and EE2 (Figure 4). ER agonist potency was observed in fine
fractions where E2 and EE2 elute, and the magnitude of
VOL. 35, NO. 18, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3623

FIGURE 4. Further fractionation of LV Wash, Lake Mead, F3 extract
using RP-HPLC with fluorescence detection followed by luciferase
induction in the MVLN cell bioassay (estrogen responsive) by the
corresponding fractions. Response magnitude presented as percentage
of the average maximum response observed for a 1000 pM
17-estradiol standard (%-E2-max). Horizontal lines represent ( 3
SD from the mean solvent control response (set to 0%-E2-max).
Dashed line shows chromatography of E2 and EE2 standards with
no corresponding bioassay results.
induction was consistent with that predicted from EEQs
calculated from the measured concentrations of E2 and EE2
and their relative ER-agonist potencies (Table 2). The greater
ER agonist potency of the water extracts from LV Wash and
LV Bay was most likely due to increased concentrations of
E2 and EE2 as a result of WWTPs discharging into the Las
Vegas Wash serving a larger population of humans.
Significant ER agonist potency was also associated with
F3 extracts of water from three locations on the Trenton
Channel of the Detroit River (B. Lagoon, Chem., and WWTP)
and BV-effluent (Figure 1). From E2 and EE2 concentrations,
B. Lagoon, Chem., WWTP, and BV-effluent samples were
estimated to contain 5.30, 3.65, 4.25, and 14.9 ng EEQ/mL,
respectively (Table 2). Based on regression against an E2
standard curve, these concentrations of EEQ were predicted
to yield responses of 76%, 67%, 71%, and 99% E2-max,
respectively (Table 2). Observed MVLN cell responses for
these F3 samples were, however, less than predicted (Figure
1). Further fractionation and bioanalysis of F3 extracts from
BV-effluent and B. Lagoon indicated that all of the observed
ER agonist potency was contained in FFs 3 and 4 (Supporting
Information, Figures 4 and 5). However, the magnitude of
induction of the FFs was markedly different from that of the
corresponding total F3 extract. This suggests that interfering
compounds and/or unidentified ER (ant)agonists present in
F3 might have modulated the potency of the known ER
agonists.
The potential presence of interfering compounds and/or
unknown ER (ant)agonists was also suggested by the lack of
significant response for several samples. Based on concentrations
of E2 and EE2, six additional F3 samples should
have elicited significant responses in the MVLN bioassay.
Concentrations of EEQs calculated from concentrations of
EE2 and E2 in samples collected from LV Bay (Sept. 1997),
LV Marina, M. Creek, BV-upstream, MA-effluent, and EReffluent
were estimated to range from 0.85 to 4.25 ng EEQ/
mL. Regression against an E2 standard curve would result in
predicted responses of 35-71% E2-max in the MVLN assay
for these samples. Thus, the responses were less than
predicted for these samples. The reason for this observation
is unknown at this time.
MVLN responses for whole extracts were similar to those
for F3. In those cases where F3 elicited a significant response,
the whole extract also elicited a significant response (Figure
3624 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 18, 2001
1). In five of the six cases, the whole extract response was
slightly less than the response elicited by F3. This suggests
that F1 and F2 may have contained some interfering
compound(s) or ER antagonists that modulated the potency
of the known ER agonists in the samples. However, no
significant ER antagonist responses were observed (Figure
1). Because the decreases were slight, however, the results
suggest that the bulk of potential interfering (antagonistic)
compounds were present in F3. The extract of BV-effluent
was the only sample for which the whole-extract response
was greater than the corresponding F3 response. It was also
the only sample for which the concentrations of NP and OP
in F2 were predicted to yield significant ER agonist activity.
NP and OP accounted for 11% of the total EEQ calculated for
the BV-effluent extract. Thus, although F2 of the BV-effluent
sample failed to elicit a significant response, NP and OP may
have contributed to the response of the whole extract, such
that the total extract response was greater than the F3
response. In general, however, NP and OP accounted for less
than 1% of the total concentrations of sample EEQs present
in samples.
No significant ER activity was observed for blank samples,
including field blanks, laboratory blanks, and solvent blanks.
Water concentrations of these compounds have been described
previously (26).
Summary
The potency balance calculations based on instrumental
analyses and bioassay-directed fractionation support the
conclusion that E2 and EE2 were the dominant environmental
estrogens in the samples. Interfering compounds or ER
antagonists present in samples (predominantly in F3) may
have acted to mask or dampen the potency of the known ER
agonists in the MVLN bioassay. All observed responses in
the MVLN bioassay were either less than, or approximately
equal to, responses predicted based on the measured
concentrations and relative potencies of known ER agonists.
NP and OP generally contributed less than 1% of the total
EEQs. Furthermore, sample fractions containing NP and OP
did not elicit significant activity. For most samples, fractions
containing E2 and EE2 elicited responses slightly greater than
the responses of the corresponding whole extracts. Thus,
among the ER agonists detected in the samples E2 and EE2
appear to be responsible for the bulk of the activity. The fact
that observed responses were generally lower than predicted
suggests the presence of interfering compounds. MVLN
responses for whole extracts were only slightly less than those
for F3 samples. This suggests that the interfering compounds
may have been present in F3. Because there were few
instances where MVLN responses were greater than those
predicted based on concentrations of EEQs present in the
extracts, the known composition can account for the
magnitude of response observed. It is unlikely that there
were additional ER agonists of significant concentration that
were not identified.
There are insufficient data to explain differences in
bioactivity among the locations investigated. It should be
noted that samples from the Trenton Channel of the Detroit
River in Michigan received less effluent from municipal
WWTPs relative to the volume of the receiving water
compared to the other sites. Also, the population served by
the WWTPs varied greatly among sites. Further investigations
would be necessary to determine the actual loading of
bioactive compounds as a function of population density.
Without knowing the available fractions and bioaccumulation
potential of the various compounds and dose-response
relationships for target species, it is not possible to predict
the potential effects of the observed concentrations of ER
agonists on biota.

Supporting Information Available
Figures of RP-HPLC fine fractionation and fine fractionations
of BV WWTP and Black Lagoon, Trenton Channel F2 extract
and BV WWTP F3 and Black Lagoon, Trenton Channel F3
extract. This material is available free of charge via the Internet
at http://pubs.acs.org.
Literature Cited
(1) Kendall, R. J.; Dickerson, R. L.; Giesy, J. P.; Suk, W. A. Principles
and Processes for Evaluating Endocrine Disruptors in Wildlife;
SETAC Press: Pensacola, FL, 1998.
(2) Ankley, G.; Mihaich, E.; Stahl, R.; Tillitt, D.; Colborn, T.;
McMaster, S.; Miller, R.; Bantle, J.; Campbell, P.; Denslow, N.;
Dickerson, R.; Folmar, L.; Fry, M.; Giesy, J.; Gray, E.; Guiney, P.;
Hutchinson, T.; Kennedy, S.; Kramer, V.; LeBlanc, G.; Mayes,
M.; Nimrod, A.; Patino, R.; Peterson, R.; Purdy, R.; Ringer, R.;
Thomas, P.; Touart, L.; Van Der Kraak, G.; Zacharewski, T.
Environ. Toxicol. Chem. 1998, 17, 68-87.
(3) U.S. EPA. Special report on environmental endocrine disruption:
an effects assessment and analysis; EPA/630/R-96/012;
Office of Research and Development: Washington, DC, February
1997.
(4) Safe, S. H. Pharmacol. Ther. 1995, 67, 247-281.
(5) Ahlborg, U. G.; Lipworth, L.; Titus-Ernustoff, L.; Hsieh, C.-C.;
Hanberg, A.; Baron, J.; Trichopoulos, D.; Adami, H. O. Crit. Rev.
Toxicol. 1995, 25, 463-531.
(6) Sharpe, R. M.; Skakkebaek, N. E. Lancet 1993, 341, 1392-1395.
(7) Carlsen, E.; Giwercman, A.; Keiding, N.; Skakkebaek, N. E.
Environ. Health Perspect. 1995, 130, 137-139.
(8) Horwitz, K. B.; Jackson, T. A.; Rain, D. L.; Richer, J. K.; Takimoto,
G. S.; Tung, L. Mol. Endocrinol. 1996, 10, 1167-1177.
(9) Katzenellenbogen, J. A. Environ. Health Perspect. 1995, 103,99-
101.
(10) Gillesby, B. E.; Zacharewski, T. R. Environ. Toxicol. Chem. 1998,
17, 3-14.
(11) White, R.; Jobling, S.; Hoare, S. A.; Sumpter, J. P.; Parker, M. G.
Endo. 1994, 135, 175-182.
(12) Bevans, H. E.; Goodbred, S. L.; Miesner, J. F.; Watkins, S. A.;
Gross, T. S.; Denslow, N. D.; Schoeb, T. Water-Resources
Investigations Report 96-4266; 1996.
(13) Snyder, S. A.; Snyder, E.; Villeneuve, D.; Kurunthachalam, K.;
Villalobos, A.; Blankenship, A.; Giesy, J. In Analysis of Environmental
Endocrine Disruptors Keith, L. H., Jones-Lepp, T. L.,
Needham, L. L., Eds.; American Chemical Society: Washington,
DC, 2000; pp 73-95.
(14) Belfroid, A. C.; Van der Horst, A.; Vethaak, A. D.; Schafer, A. J.;
Rijs, G. B. J.; Wegener, J.; Cofino, W. P. Sci. Total Environ. 1999,
225, 101-108.
(15) Bennie, D. T. Water Qual. Res. J. Canada 1999, 34, 79-122.
(16) Naylor, C. G.; Mieure, J. P.; Adams, W. J.; Weeks, J. A.; Castaldi,
F. J.; Ogle, L. D.; Romano, R. R. JAOCS 1992, 69, 695-703.
(17) Poland, A.; Knutson, J. C. Annu. Rev. Phamacol. Toxicol. 1982,
22, 517-554.
(18) Neff, J. M. Polycyclic Aromatic Hydrocarbons in the Aquatic
Environment: Sources, Fates, and Biological Effects; Applied
Science: London, 1979.
(19) Schecter, A. Dioxins and Health; Plenum Press: New York, 1994;
p 710.
(20) Krishnan, V.; Safe, S. Toxicol. Appl. Pharmacol. 1993, 120, 55-
61.
(21) Anderson, M. J.; Miller, M. R.; Hinton, D. E. Aquat. Toxicol.
1996, 34, 327-350.
(22) Tran, D. Q.; Ide, C. F.; McLachlan, J. A.; Arnold, S. F. Biophys.
Res. Commun. 1996, 229, 102-108.
(23) Clemons, J. H.; Allan, L. M.; Marvin, C. H.; Wu, Z.; McCarry, B.
E.; Zacharewski, T. R. Environ. Sci. Technol. 1998, 32, 1853-
1860.
(24) Sanderson, J. T.; Giesy, J. P. In Encyclopedia of Environmental
Analysis and Remediation; Meyers, R. A., Ed.; John Wiley and
Sons: New York, 1998; pp 5272-5297.
(25) Villeneuve, D. L.; Blankenship, A. L.; Giesy, J. P. In Toxicant-
Receptor Interactions; Denison, M. S., Helferich, W. G., Eds.;
Taylor & Francis: 1998; pp 69-97.
(26) Snyder, S. A.; Keith, T. L.; Verbrugge, D. A.; Snyder, E. M.; Gross,
T. S.; Kannan, K.; Giesy, J. P. Environ. Sci. Technol. 1999, 33,
2814-2820.
(27) Pons, M.; Gagne, D.; Nicolas, J. C.; Mehtali, M. Biotechniques
1990, 9, 450-459.
(28) Kennedy, S. W.; Jones, S. P. Anal. Biochem. 1994, 222, 217-223.
(29) Villeneuve, D. L.; Blankenship, A. L.; Giesy, J. P. Environ. Toxicol.
Chem. 2000, 19, 2835-2843.
(30) Hilscherova, K. M.; Machala, M.; Kannan, K.; Blankenship, A.
L.; Giesy, J. P. Environ. Sci. Pollut. Res. 2001, 7, 159-171.
(31) Soto, A. M.; Chung, K. L.; Sonnenschein, C. Environ. Health
Perspect. 1994, 102, 380-383.
Received for review May 10, 2000. Revised manuscript received
May 23, 2001. Accepted July 2, 2001.
ES001254N
VOL. 35, NO. 18, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3625