”u; ..

. 1.
.1...

 

a.

3.

EV
r»

I:

mam.»

Kr.

1.1.»...
«A. ‘
.35?

3:1»:-

.p..,.. bra;

 

iumtwafigfizfia

Now 01.03.

 

means

IV RSITY Ll E

llllllllllllllllllllllll lllllllll

3 1293 01566 3747

ll“

 

 

 

This is to certify that the

dissertation entitled

Xeno-estrogens in the Aquatic Environment:
Development and Application of in vitro
and in vivo Bioassays

presented by

Vincent J. Kramer

has been accepted towards fulfillment

of the requirements for
Fisheries and Wildlife -

Doctor of Philosophy degree in Environmental Toxicology

Ma 10!” Mressor

 

 

 

Date May 23, 1996

MSU is an Affirmative Action/Er; ual Opportunity lruu'lun'on 0-12771

 

LIBRARY

Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove thie checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

PM! o n «Mo
'SUN n___m‘.’p"

M

{a w e -
l ;

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

F—l—j—j
l li—TLJ
ll l—J

 

 

 

 

 

 

 

L__J_J
F—lV—T—j

MSU leAn Affirmative WW OW Intuition

 

XENO-ESTROGENS IN THE AQUATIC ENVIRONMENT:
DEVELOPMENT AND APPLICATION OF IN VITRO AND IN VIVO
BIOASSAYS

Bv

Vincent J. Kramer

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

' Department of Fisheries and Wildlife
and
The Institute for Environmental Toxicology

1996

 

ABSTRACT
XENO-ESTROGENS IN THE AQUATIC ENVIRONMENT:
DEVELOPMENT AND APPLICATION OF IN VITRO AND IN VIVO
BIOASSAYS
Bv

Vincent J. Kramer

A well-defined, logical framework for evaluating the estrogenic and anti-
estrogenic activity of chemical substances was developed. Properties of
chemical substances that determine estrogenic or anti-estrogenic potential
were investigated at the molecular, cellular, and physiological levels of
organization using both in vitro and in viva bioassays. Hydroxylated
polychlorinated biphenyls, (PCBs), metabolites of PCBs that have been identified
in human and animal samples, were studied. Molecular modelling of the
hydroxylated PCBs was used to predict estrogen receptor binding affinity and
to investigate the molecular parameters that affected binding. Hydroxyl
substitution in the para position was an important characteristic of strongly
binding compounds, provided that the hydroxyl was not flanked by vicinally
substituted chlorines. A functional in vitro bioassay utilizing a recombinant
human breast cancer cell line was used to evaluate the estrogenic and anti-
estrogenic potency of the hydroxylated PCBs. Hydroxylated PCBs that were
previously identified as mammalian metabolites exhibited anti-estrogenicity that
was readily reversed by 17B-estradiol. However, cytotoxicity appeared to
explain most of the anti-estrogenicity. A reporter gene system was designed,

constructed and partially characterized for estrogen responsiveness and utility

as a screening assay. The estrogen responsive plasmid, pGudluc2.2, was not
useful as a screening tool because the luciferase reporter gene was expressed
at low levels and estrogen treatment caused reduction of expression, instead
of induction, perhaps due to estrogen receptor repression of the promoter
activity. Finally, a functional in vivo bioassay utilizing fathead minnows was
developed and used to test the activity of the benchmark estrogen, 17%-
estradiol. Plasma alkaline-labile phosphorous, an indicator of vitellogenin
expression, proved to be a useful biomarker of reproductive impairment. The
biochemical and physiological responses identified in this assay will be useful

in evaluating effects of waterborne xenoestrogens on fish.

Copyflghtby
VINCENT JOSEPH KRAMER

1996

To Maria, Sara Rose and Little One

ACKNOWLEDGEMENTS

I was generously supported by the Institute of Water Research through a USDA
National Needs Fellowship for the first three years of my studies, for which I
thank Dr. Frank D’ltri. Other financial support was provided by the MSU Cancer
Center and the Chemical Manufacturers Association, for which I express my
sincere gratitude. I would also like to express my heartfelt thanks to my major
professor Dr. John P. Giesy. He provided me the freedom to make (and
correct!) mistakes while guiding my development as a scientist, in the tradition
of a true mentor. The technical and intellectual support of numerous colleagues
is gratefully acknowledged and includes thanks to: William Helferich, PhD; J.
Thomas Sanderson, PhD; Kenneth Froese, PhD; Susan Pierens, PhD; Stephanie
Miles-Richardson, DVM; William Bowerman, PhD; Michael Denison, PhD; David
Verbrugge, Lori Feyk, Cathy Richter, Krista Nichols, Erin Snider, Shane Snider,
George Noguchi, John Besser and Jody Kubitz. Finally, I would like to
recognize the professional and efficient work of Alice Ellis, Pat Unger and
Maxine Schafer of the Pesticide Research Center. It was a joy to work with you

vi

TABLE OF CONTENTS

Description
LIST OF TABLES

LIST OF FIGURES

INTRODUCTION

Estrogen Agonists in the Environment
Functions of Estrogens
Importance and Mechanisms of Anti-Estrogenicity
General Methodological Approach

Molecular Modelling

In Vitro Bioassay of Estrogenicity and

An ti-Es trogenicity

In Vivo Bioassay of Wa terborne Es trogens

Significance

CHAPTER 1 HYDROXYLATED METABOLITES OF PCBs
ARE ANTI-ESTROGENIC IN A HUMAN BREAST
ADENOCARCINOMA CELL LINE (MOE-7)
Introduction
Materials and Methods
Chemicals
Human Estrogen Receptor Binding Assay
Luciferase Expression Assay
Cell Viability Assay
H4llE EROD Assay
Statistical Analyses
Results and Discussion
Human Estrogen Receptor Binding
Cell Viability
Effects of £2 and OH-CBs on Luciferase
Expression
Mechanisms of Activity and Relationship
with Chemical Structure
Partial Estrogenic Activity

vii

 

xii

do oooooocnw-a-a

dd

_3
N

16
16
18
20
22
23
23
26
26
28
28

37

37

TABLE OF CONTENTS (CONT'D)

293mm

Anti-Estrogenic Activity
lmplica tions for Human Risk Assessment
Conclusions

CHAPTER 2 SPECIFIC BINDING OF HYDROXYLATED
PCBs AND OTHER COMPOUNDS TO THE ESTROGEN
RECEPTOR: STRUCTURE-ACTIVITY RELATIONSHIPS
Introduction
Estrogen Receptor A gonists and
Antagonists
Uses and Limitations of ER Binding Assay
Objectives
Materials and Methods
Chemicals
Calf Uterine Estrogen Receptor Competitive
Binding Assay
Molecular Modelling
Results and Discussion

CHAPTER 3 DEVELOPMENT AND CHARACTERIZATION
OF AN ESTROGEN RESPONSIVE LUCIFERASE
REPORTER PLASMID, pGUDLUC2.2
Introduction
Materials and Methods
Construction of pGudlch. 2
Transfection of pGudluc2. 2 into MCF- 7
Cells
Characterization of Stab/y Transfected
Estrogen Responsive Luciferase Reporter
Gene
Results and Discussion

viii

 

39
44
46

49

49
49

51
52
53
53
53

54
56

69
69
7O
71
71

74

75

TABLE OF CONTENTS (CONT’DI

Description

CHAPTER 4 REPRODUCTIVE IMPAIRMENT AND
INDUCTION OF PLASMA ALKALINE-LABILE
PHOSPHATE, A BIOMARKER OF ESTROGEN
EXPOSURE, IN FATHEAD MINNOWS (PIMEPHALES
PROMELASI EXPOSED TO WATERBORNE 17B-
ESTRADIOL
Introduction
Materials and Methods
Chemicals
Experimental Design
Maintenance and Acclima tion of Fish
Exposure of Fish to 17B—Es tradiol
Collection of Tissues
Vite/logenin Analysis
Water Quality
Statistical Analysis
Results and Discussion
Mortality
Water Quality
Egg Production
Vite/logenin Induction as a Biomarlter of
Reproductive Impairment
Mechanism of Reproductive Impairment
Caused by 17B-Estradiol

SUMMARY AND CONCLUSIONS
APPENDIX

LIST OF REFERENCES

ix

 

81

81
82
82
83
84
86
86
87
88
89
92
92
93
93
96

98

103
105

114

LIST OF TABLES

Description

Table 1. Environmental occurrence and structures of
OHCBs and 17B-estradiol assayed with MCF-7-LUC cells.

Table 2. Summary of MCF-7-LUC cellular responses to
OHCBs.

Table 3. Potency of anti-estrogenicity/estrogenicity of
OHCBs estimated by the area of activity.

Table 4. Molecular descriptors of hydroxylated PCBs
calculated using Hyperchem®

Table 5. Prediction of IogIICSO), (Iog(nM)) using multiple
linear regression of the optimum five molecular descriptors.

Table 5. (Cont’d).

Table 6. Estimates of calf uterine estrogen receptor binding
logilC50), (Iog(nM))for miscellaneous compounds.

Table 7. Compounds of environmental concern that did not
bind to calf uterine estrogen receptor.

Table 7. (Cont'dl.

Table 8. Results of molecular modelling of hydroxylated
PCBs.

Table 8. (Cont'd.)
Table 9. Optimum five variable weighted and unweighted

models for the prediction of Iog(lC50), (Iog(nM)) for
binding to calf uterine estrogen receptor.

 

17

27

38

55

57

58

59

6O

61

63

64

66

LIST OF TABLES (CONT’DI

Description

Table 10. Fathead minnow mortality occurring during the
exposure period.

Table 11. Water quality during the exposure period,
meanISE).

Table 12. Calculation of No Observed Effect Level (NOEL)
or EC5, Lowest Observed Effect Level (LOEL) or ECZO,
Maximum Allowable Toxicant Concentration IMATC), and
ECSO using ANOVA and nonlinear regression of nominal
and measured 17B—estradiol concentrations for reduction
of egg production of fathead minnows.

xi

 

Page
92

94

97

LIST OF FIGURES

Description

Figure 1. The steroid hormone "skeleton" and the
endogenous estrogens

Figure 2. Mechanism of action of estrogens within cells.

Figure 3. Representative luciferase expression of MCF-7-
LUC cells exposed to OHCBs and E2 in the assay medium:
OHCB4: 4-OH-2’,3,3',4',5,5’-Cl6-biphenyl; OHCB10, 40H-
2',4’,6’-Cl3-biphenyl; OHCB11 4-OH-2',3’,4’,5’-Cl4-
biphenyl; E2 17B—estradiol; (A) ng luciferase/mg protein,
n=2; (AI ng luciferase/mg protein, points omitted from
calculation of Iog(EC§O) and Iog(EC20) due to cytotoxicity;
symbol and bars indicate mean :tSE.

Figure 4. Effect of 17B-estradiol (E2) co-exposure on
luciferase expression response to OHCBs: (I) EC50,
OHCB Effective Concentration causing a 50% reduction (or
induction) of luciferase activity with respect to the solvent
control response; (0) EC20, OH CB Effective Concentration
causing a 20% reduction (or induction) of luciferase
activity with respect to the solvent control response; (----)
Relevant limits of concentration of OHCBs and E2 were
determined for the calculation of the "area of activity"
such that the upper limit of OHCB concentration defined as
either the maximum soluble concentration (applied to
OHCB1,5,9,10,11,13) or the cytotoxic concentration
(0HC82,3,4,6,7,8,12); E2 limits defined by highest (1 nM)
and lowest concentrations tested (0.0009 nM); (NA) Not
active, predicted E050 or EC20 greater than upper OHCB
concentration limit; (Light Gray Shading) Area of weak
activity depicting an expected response of 20% to 50%;
(Dark Gray Shading) Area of strong activity depicting an
expected response of greater than 50%. All OHCB ECSOs
and ECZOs depict reduction (E2 antagonism) of

xii

 

Page

30

33

LIST OF FIGURES (CONT'DI

Despription

luciferase expression except for OHCBIO and OHCB11
that caused induction (E2 agonism) of luciferase
expression at 0.0009 nM E2; (A) OHCBI, 3-0H-
2,3',4,4’,5-Cl5-biphenyl; (B) 0HCBZ, 4-OH-2,3,3',4’,5-CI5-
biphenyl; (C) OHCB3, 3-OH-2',3’,4,4',5-Cl5-biphenyl; (D)
0HCB4, 4-OH-2',3,3',4',5,5’-Cl6-biphenyl; (E) OHCBE, 4-
OH-2’,3,4',5,5’-Cl5-biphenyl; (F) 0HCB6, 4-0H-3,3'4,’-Cl3-
biphenyl; (G) OHCB7, 2-0H-2',3,3',4,4’-Cl5-biphenyl; (H)
OHCBB, 4—0H-3,3',4’,5,5'-Cl5-biphenyl; (I) 0HCB9, 4-0H-
2',3,3’,4',5-Cl5-biphenyl; (J) OHCBIO, 4-0H-2’,4',6’-CI3-
biphenyl; (K) OHCB11, 4-0H-2',3’, 4',5’-CI5-biphenyl; (L)
OHCBIZ, 3-OH—2’,3’,4’,5'-Cl5-biphenyl; (M) OHCBl3,
4,4'-(0H)2-3,3’,5,5'-Cl4-biphenyl.

Figure 4. (Cont’d).

Figure 5. Relationship between OHCB inductive (E2
agonist) activity area and human estrogen receptor (hER)
binding affinity represented by IC50, the OHCB Inhibitory
Concentration causing a 50% decrease in specific binding
of 3H-E2 to hER in vitro: (0) Total induction area (response
greater than 20% with respect to the solvent controls); (A)
Strong induction area (response greater than 50% with
respect to the solvent controls). See Figure 4 for
calculation of the inductive activity area.

Figure 6. Relationship between OHCB reductive (E2
antagonist) activity area and human estrogen receptor
(hER) binding affinity represented by IC50, the OHCB
Inhibitory Concentration causing a 50% decrease in
specific binding of 3H-E2 to hER in vitro: (Number label)
Indicates OHCB identification number from Table 1; (A)
Total reductive activity area (response greater than 20%
with respect to the solvent controls); (B) Strong reductive
activity area (response greater than 50% with respect to
solvent controls). See Figure 4 for calculation of
reductive activity area. p = probability that the slope of the
regression line is equal to zero, a=0.05.

xiii

 

35

40

42

LIST OF FIGURES (CONT'DI

Dsssriptipn

Figure 7. Relationship between OHCB reductive (EZ
antagonist) activity area and cell viability index for all
0HCBs except OHCB10 and OHCBI 1 that exhibited partial
estrogen agonist activity: (—) All points included in
regression; (---) Regression omitting outliers as noted; (A)
Total reductive activity area (response greater than 20%
with respect to solvent controls); (B) Strong reductive
activity area (response greater than 50% with respect to
solvent controls). p=probabi|ity that the slope of the
regression line is equal to zero, a=0.05.

Figure 8. Comparison of the area for in vitro reductive (E2
antagonism) effects and the area for possible in viva
concentrations defined by the lower limit of average E2
plasma concentrations (nM) in humans (Becker, 1990) and
the upper limit of the predicted OHCB concentration (nM)
in an adult human plasma sample (0HCB9 for example,
Bergman, et al. 1994). Determination of the areas for
reductive in vitro effects depicted in Figure 4. The
predicted OHCB concentration is the estimated upper limit
of the range of concentrations of 0HCBs in human plasma
(10-50% of the total plasma PCB concentration, Bergman,
et al. 1994). Overlap of the in vitro reductive effects area
with the in viva possible concentrations area indicates a
condition in which the OHCB could potentially inhibit E2
activity in viva. On the other hand, the extent of non-
overlap between the two areas indicates the "margin-of-
safety" area (555) between the threshold for in vitro effects
(EC20) line and the possible in viva concentrations area.

Figure 9. Progress of multiple regression relating calf
uterine binding Iog(lC50) to molecular descriptors of
0HCBs in which the best combination of descriptors that
maximized r2 was selected for each number of descriptors.

Figure 10. Comparison of predicted vs. observed Iog(lC50)
for binding of 0HCBs to calf uterine estrogen receptor
using both weighted and unweighted five variable multiple
regression models; * =asymptotic standard error of

xiv

 

47

65

67

LIST OF FIGURES (CONT’DI

Dsspriptipn

observed Iog(lC50) inestimable by nonlinear regression and
therfore not shown; other error bars not shown because
the bar was smaller than the symbol; vertical error bars are
standard error of the individual predicted Iog(lC50).

Figure 11. Construction history of pGudluc2.2.

Figure 12. Oligonucleotide containing two EREs used in
pGudluc2.2.

Figure 13. Response to 17B-estradiol by MCF-7 cells
transiently transfected with pGudluc2.2; N=2 at each
concentration; bars indicate mean :1 SE; letters indicate
means are not significantly different by Fisher's LSD,
a=0.05; F statistic produced by ANOVA.

Figure 14. Time course of luciferase expression response
to 17B—estradiol by MCF-7 cells stably transfected with
pGudlu02.2, clone GLUC1; N =1 at each concentration.

Figure 15. Luciferase expression response of GLUCI to
17B-estradiol, diethylstilbestrol and tamoxifen; N=4 at
each concentration; bars indicate meanzt ISE; letters
indicate means are not significantly different by Fisher’s
LSD, a=0.05; F statistic produced by ANOVA.

Figure 16. Luciferase expression response of GLUC5 and
GLUC6 to 17B—estradiol; N = 2 at each concentration; bars
indicate mean :1 SE; letters indicate means are not
significantly different by Fisher’s LSD, 0: 0.05; F statistic
produced by ANOVA.

Figure 17. 17B—Estradiol measured by ELISA during
exposure period, RSD =Relative standard deviation of
duplicate measurements, Method LOD=Method limit of
detection, Recovery=recovery of spiked samples
containing 2724 ng/L E2; + =Tank 1, +(bold) =Tank 2,
A =Tank 3, A =Tank 4, O =Tank 5, O =Tank 6, v =Tank
7, v =Tank 8, O =Tank 9, o =Tank 10.

XV

 

72

72

73

76

77

78

95

LIST OF FIGURES (CONT’DI

mm

Figure 18. Reduction of egg production by exposure of
fathead minnows to waterborne 17B—estradiol (E2)
modeled by nonlinear regression after a normal distribution
function; 0 =E2 measured in each tank by ELISA, O =E2
nominal concentration.

Figure 19. Induction of plasma alkaline-labile phosphorous
(ALP), an indication of vitellogenin expression, in female
(A) and male (B) fathead minnows exposed to waterborne
17B-estradiol (ELISA measured concentrations of E2); in A,
O=measurements not included in log-linear regression,
sub-threshold mean=0.473 mg/ml P, 95% UCL=95%
upper confidence limit on the sub-threshold mean,
threshold =intersection of regression line with 95% UCL;
in B, threshold =geometric mean of Iog(EC5) and
Iog(ECZO), nonlinear regression used to model apparent
plateau.

Figure 20. Correlation between egg production and

plasma alkaline-labile phosphorous in female (A) and male
I B) fathead minnows exposed to waterborne 17B—estradiol

xvi

 

97

99

100

INTRODUCTION

Estrogen Agonists in the Environment

Estrogen agonists are compounds that mimic the effects of estrogen.
The classic definition of an estrogen is a compound that produces cornification
of the vaginal epithelium of rodents, typically mice or rats (Solmssen, 1945).
Other physiological endpoints have been used to define estrogenicity including
increased uterine weight and increased uterine glycogen content after dosage
of an estrogenic compound (Clark and Peck, 1979). Under the classic definition,
it is impossible to distinguish whether the estrogenic activity is due to the
parent compound or to its metabolites. A refined definition of estrogen,
recognizing the role of nuclear hormone receptors in gene expression, is a
compound that binds to the estrogen receptor (ER), thus activating transcription
of genes under the regulatory control of estrogen responsive elements (EREs;
Evans, 1988). Therefore, the description of a compound as an estrogen is
dependent upon the method of testing and can in the classical sense include
compounds that require metabolic activation to an estrogenic form.

Estrogen agonists encompass a range of chemical classes and structures.
The structure of the endogenous estrogens, 17B—estradiol (E2), estrone, (E1),

and estriol (E3) (Figure 1) is based upon the steroid hormone template

0

 

 

Perhydrocyclopentanophenanthrene Estrone
OH OH
CHa
OH
17l3-Estradiol Estriol

Figure 1. The steroid hormone "skeleton" and the endogenous estrogens

perhydrocyclopentanophenanthrene (Zubay, 1988). The exogenous estrogens
can be classified into two broad categories: 1) natural products, and; 2)
synthetic compounds. The chemical structural characteristics of exogenous
estrogens vary but can be summarized as: 1) containing a sterically unrestricted
phenol group, and; 2) a moderately hydrophobic substituent of three or more
carbons bonded opposite the phenolic group (Dodds and Lawson, 1938; Mueller
and Kim, 1978). The effects of phytoestrogens have been long known in the
field of animal husbandry. Bennets, et al. (1946) noted infertility in sheep
grazing on clover pasture containing the isoflavone formonetin which is
metabolized in the sheep gut to the active estrogen equal (Shutt and Braden,
1968; Lundh et al., 1990). Phytoestrogens have been implicated in

reproductive dysfunction of captive cheetah (Setchell, 1987) and in precocial

3
vitellogenin production in sturgeon (Pelissero, et al., 1991 ). Synthetic estrogens
have been produced either by design: for example, diethylstilbestrol and
ethynylestradiol (Solmssen, 1945; Anstead, 1990; Jordan, et al., 1986) or by
accident: for example, DDT and its metabolites (Kupfer and Bulger, 1980;
Nelson, et al., 1978; Jensen, et al., 1979; Heath, et al., 1969, Keith, 1966);
PCB’s and their metabolites (Gellert, 1978; Ecobichon and McKenzie, 1974;
Korach, et al., 1988; Nelson, 1974; Bitman and Cecil, 1970); Kepone (Gellert
and Wilson, 1979; Gellert, 1978; Eroschenko and Plamiter, 1980; Eroschenko,
1981); alkyl phenols (Soto, et al., 1991; Soto, et al., 1992), bisphenol A
(Krishnan, et al., 1993; Brotons, et al. 1995); phthalate plasticizers and the
food additive butylated hydroxyanisole (Jobling, et al. 1995). The potential for
estrogen mimics to affect the reproduction and survival of fish and wildlife

populations deserves careful study.

Functions of Estrogens

Estrogens, such as 17B—estradiol (E2), estrone (E1), and estriol (E3), are
a group of vertebrate steroid hormones that elicit a wide range of physiological,
developmental, and behavioral responses in both females and males. Estrogen
plays an important role: 1) in stimulating sexual maturation, uterine growth,
vaginal cornification, and mating behavior in female mammals (Selcer and
Leavitt, 1991) and mating behavior in birds (Adkins, 1978); 2) via conversion
by aromatase, in sexual differentiation of mammalian and avian central nervous

systems (Arnold and Gorski, 1984); 3) in modulating hypothalamic

4

gonadotropin secretion during the primate menstrual cycle (Knobil, 1980); 4)
in stimulating vitellogenesis in oviparous vertebrates (Selcer and Leavitt, 1991;
Redding and Patina, 1993; Emmerson and Petersen, 1976; Wallace and Jared,
1968; deVIaming et al. 1984); 5) in differentiation of gonads and gonadal ducts
in mammals (Selcer and Leavitt, 1991), birds (Fry and Toone, 1981; Johnson,
1986), fish (Adkins-Regan, 1987), amphibians (Witschi, 1971), and reptiles
(Bull, et al. 1988), and; 6) in development of secondary sexual characteristics
(Selcer and Leavitt, 1991).

Estrogens elicit cellular responses (Figure 2) by directly crossing cellular
membranes to bind to nuclear ER (Lieberman, et al. 1986), which, in turn, bind
to and activate specific regulatory segments on DNA--the EREs (Klein—Hitpass,
1986). EREs are trans-acting transcriptional enhancers (Cullen, et al. 1993).
Efficient activation of the ERE occurs after binding of the ligand-bound receptor
dimer (Gordon and Notides, 1986; Kumar, et al., 1987; Fawell, et al., 1990).
The nucleotide sequence of the ERE in the vitellogenin A2 gene 5’ flanking
region of Xenapus laevis is a 15 base pair palindrome: 5’-
AGGTCACAGTGACCT-B' (Walker, et al., 1984). However, considerable
variation exists in the nucleotide sequence of EREs (Walker, et al., 1984;
Chang, et al., 1992; Evans, 1988) that affects the efficiency of transcriptional
enhancement (Pilat, et al., 1993). A number of accessory proteins are thought
to be involved in binding of ER to the ERE. Hsp90 is released from the ER
complex after E2 binding. Dimerization of the ER complex is initiated by E2
binding. Hsp70 remains bound to the ER complex. Tyrosine phosphorylation

of ER is necessary for E2 binding (Migliaccio, et al. 1991). Six functional

5

domains of the ER have been identified: NB, the N terminal "modulator"
domain confers gene transcription specificity; C, essential for activation of
transcription and binding specificity to ERE; D, the "hinge" region; E, hormone
binding and efficient transcriptional activation (but not essential for ERE

recognition); F, the carboxyl terminal (Kumar, 1987).

Importance and Mechanisms of Anti-Estrogenicity

In the "environmental estrogen" arena, discussion of environmental anti-
estrogens is under-represented. This is unfortunate because the potential
effects of an anti-estrogen are just as significant to the developing or adult

organism as an estrogen. As a general example, consider the role estrogen

 

 

I a
O my; I. EITMI Nucleus DImeZIutIonl
ERE letreoen — —

metre plough”: ‘_— a w
P Phelvhele Vitellooenln m3
Prolectln (MINA

0 2:12:06. CethepelnD fl‘ '1 ‘
”It!
, w: I" .. ’— ffll

l ILE-E'iwj
' "9le 1| Ema”- V 0

 

 

 

 

 

P
”'4”: r «— («mu
in...
man Afffltv Eocenrylttler?
VLlflfld Ilndino - - -
em
P
- _&_p («my

Target Cell

 

 

hall-1“!

Figure 2. Mechanism of action of estrogens within cells.

6

plays in signalling the production of vitellogenin, necessary for egg production
in all egg-laying vertebrates. There could be significant consequences for
successful reproduction if reduced egg mass or egg number were caused by a
blockage of estrogen action. Anti-estrogenic substances act by several
mechanisms, not necessarily related to ER binding and activation. Classical
estrogen antagonists such as lCl 164,384 and tamoxifen bind competitively to
ER but block or reduce the effectiveness of the ligand-bound receptor to
enhance gene expression (Fawell, et al. 1990; Jordan, et al. 1986). Aryl
hydrocarbon receptor (Ath agonists, such as 2,3,7,8-TCDD and coplanar
PCBs, cause down-regulation of ER (Zacharewski, et al. 1991) and may also
interfere with ER interactions with DNA. Aromatase inhibitors, such as
aminoglutethimide, block conversion of testosterone to 17B-estradiol and are
used in the treatment of metastatic estrogen-dependent breast cancer (Farooqi,
et al. 1992).

Environmental contaminants occur as mixtures that may contain both
estrogenic and anti-estrogenic components. The effects of such mixtures of
estrogens and anti-estrogens will be determined by the relative contributions of
each type of estrogen-active compound and the nature of their interactions,
either additive or non-additive. A bioassay driven approach to assess the total
estrogenic or anti-estrogenic potency of the mixtures found in fish or wildlife
is recommended so that the interactions of mixtures at the level of the estrogen
receptor can be taken into account. Additionally, it is impossible to predict
whether a compound will exhibit estrogenic or anti-estrogenic activity solely by

measuring the affinity with which it binds to the estrogen receptor. Rather, the

7
potency of the activity, either estrogenic or anti-estrogenic, may be related to
the estrogen receptor binding affinity, though anti-estrogenicity can also be
elicited by mechanisms not involving binding to the estrogen receptor (as
described previously).

To fully understand the toxicological action of xeno-estrogens, it is
necessary to elucidate the biochemical, cellular, and physiological mechanisms
of their action. To accomplish this, a framework for evaluating the
estrogenicity and anti-estrogenicity of substances was developed. This
framework defines the estrogeniclanti-estrogenic characteristics of a substance
at several levels of toxicological organization: 1) molecular structural descriptors
calculated by desktop molecular modelling; 2) an in vitro estrogen receptor
binding affinity assay; 3) an in vitro estrogen-inducible luciferase reporter gene
expression assay; 4) an in viva fish exposure assay. The ultimate goal of such
an approach is the development of quantitative structure-activity relationships
that will allow the prediction of estrogenic or anti-estrogenic potential of either
pure chemicals (for example, in the risk assessment process for the
manufacture of new chemicals) or environmentally-derived complex mixtures
(for example, in the assessment of environmental impact of contamination).
The database required for such a goal is beyond the scope of this research
project. However, this research project is intended to serve as a guide for the

development of such a database.

8

General Methodological Approach

Molecular Made/ling

Molecular modelling allows the estimation of molecular descriptors such
as minimum energy three dimensional molecular structure and various electronic
properties. These molecular descriptors can be used to quantitatively predict
the biochemical properties of molecules after an appropriate number of
"training" molecules have been used to establish a statistical model of the
relationship. In this project, the binding of hydroxylated PCBs to the estrogen
receptor was modelled using a set of molecular descriptors calculated using a
desktop molecular modelling software package. It is anticipated that the
molecular descriptors defined in this project may be used by others to develop
models of the functional activity estrogenic and anti-estrogenic compounds in

in vitro and in viva bioassays.

In Vitra Biaassay af Estragenicity and An ti-Estragenicity

Aquatic organisms are exposed to complex mixtures of endogenous and
exogenous estrogen agonists. Estrogen antagonists are an additional
component of complex mixtures. Presently, the estrogenic or anti-estrogenic
activity of a complex mixture cannot be predicted by chemical analysis alone
because: 1) the synergistic effects of agonists and antagonists are unknown;

2) new examples of estrogen agonists and antagonists continue to be

9
discovered; 3) authentic standards for many estrogen agonists and antagonists
are unavailable. The estrogenic or anti-estrogenic activity of a complex mixture
can be measured using a functional in vitro bioassay. Two avenues of
approach to the development of a functional in vitro bioassay have been taken.
First, a well-characterized recombinant cell line, MCF-7-LUC, was obtained that
was used to evaluate the estrogenicity and anti-estrogenicity of a number of
pure compounds. Second, this research project also undertook an attempt to
develop a genetically-engineered fish cell line, GUDLUC2.2, with the following
properties: 1) sensitive--responding to a wide range of concentrations; 2)
selective—-responding only to estrogenic/anti-estrogenic compounds via defined
biochemical pathways, and; 3) rapid--responding in a time frame to facilitate
screening of large numbers of samples. The fish cell line was not developed
becasue the plasmid, pGudluc2.2, was not sufficiently estrogen-responsive to
be of practical use. The molecular mechanisms for the lack of success were
not fully elucidated. However, designs for the construction of another cell line,
GUDLUC3.0, are presented for future research. Although the screening of
complex mixtures for estrogenicity/anti-estrogenicity was not the primary aim
of this research project, the methods developed to assay single chemicals can

be readily applied to complex mixtures.

10

In Viva Biaassay af Wa terbarne Estrogens

Complementary to the development of the in vitro bioassay, an assay
designed specifically to measure the effects of estrogenic compounds on the
physiology, histopathology and reproduction of fathead minnows, (Pimephales
pramelas) was developed. One application of the laboratory assay will be to
calibrate and validate the response of fathead minnows to effluents in caged
fish studies planned by other researchers. Another application will be the
development of concentration-response relationships that can be used in
environmental risk assessment. Fish are the sentinels of aquatic ecosystem
degradation and have signalled the presence of estrogenic compounds in
affected waterways. For example, fishermen in the U.K. reported catching
feminized ruffe in waters downstream of municipal wastewater treatment
plants. Subsequent caged fish studies using rainbow trout and carp, confirmed
the feminizing effect (induction of vitellogenin synthesis) of these waters.
However, the active agent (or agents) have not been identified (UK Directorate
of Fisheries Research, 1992). The proposed research will determine the
relationship between induction of vitellogenin synthesis in males and females,
a highly specific biomarker of estrogen exposure, and reproductive
performance, an important facet in the survival of feral fish populations. Also,
the effects of waterborne estrogen exposure on reproduction will be assessed.
Histopathological examination will provide insights into the mechanisms of
action of estrogens on male and female reproduction. However, the

histopathological information is the subject of another researcher’s PhD.

11
dissertation (Miles-Richardson, 1996)and will not be reported here. The results
of the exposure of fathead minnows to 17B—estradiol will be reported as the

benchmark to which all other compounds will be compared.

Significance

Concern over environmental estrogens has risen both in the media and
in regulatory agencies. There is an urgent need for contribution of scientific
information to national debate that at times has focused on supposition and
opinion. The aquatic environment frequently becomes a sink or site of first
effects for environmental contaminants, including estrogenic contaminants.
Therefore, it is important to study the effects of estrogenic contaminants in
aquatic organisms. The fathead minnow is a widely used organism for aquatic
toxicity testing and is an appropriate model organism for examining estrogenic
endpoints. The in vitro assay, when fully validated, will provide a rapid,
sensitive, and selective means of assessing estrogenicity and anti-estrogenicity
of both pure compounds and complex mixtures. The in vitro assay can also be
used to direct Toxicity Identification and Evaluation procedures rapidly and
inexpensively. The ultimate application of the results of this proposal will be
to develop an understanding of the effects of environmental estrogens/anti-
estrogens on reproduction, development, and survival of environmentally-

exposed fish and other aquatic vertebrates.

CHAPTER 1
HYDROXYLATED METABOLITES OF PCBs ARE ANTI-ESTROGENIC IN A

HUMAN BREAST ADENOCARCINOMA CELL LINE (MCF-7I

Introduction

Although the banning of PCBs (polychlorinated biphenyls) production and
use in the industrially developed world has resulted in decreasing environmental
levels (Schmitt, et al. 1990), bioaccumulatian and biomagnification of PCBs
results in significant concentrations in upper trophic level species such as fish-
eating waterbirds and mammals such as seals, and humans who eat PCB-
cantaining seafood (Jansson, et al. 1993; DeWailly, et al. 1989, Haraguchi, et
al. 1992). The relative abundance of individual chlorinated biphenyls (CBs) in
environmental samples varies with the location and source of contamination,
the age of the sample, and trophic level of the organism. In general, fish,
wildlife, and human samples are dominated by persistent penta-CBs to hepta-
085, due to their resistance to biodegradation relative to the more-readily
metabolized mono-CBs to tetra-CBS (SundstrOm, et al. 1976) that possess
vicinal hydrogens in the 3,4 position of one of the phenyl rings (Bergman, et al.
1 992).

PCBs undergo oxidative alteration via cytochrome-requiring P450s 1A1,
1A2, 2B1, and 282 enzymes which catalyze the formation of intermediate
arene oxides that lead to hydroxyl or methyl sulphone metabolites (Bergman,
et al. 1994). Hydroxylated chlorinated biphenyls (0HCBs) and their sulfate and

12

13

glucuronide conjugates are typically excreted via feces or urine. However,
organisms may retain OHCBs because these compounds are lipophilic and bind
to plasma proteins, particularly transthyretin (Bergman, et al., 1994). 0HCBs
have been detected in excreta of wildlife (Jansson, et al. 1975) and in the blood
of seals and humans (Bergman, et al. 1994). Hydroxylated metabolites of
3,3’,4,4’—Cl4-biphenyl competitively bound to human transthyretin and has been
reported to cause marked reduction in plasma thyroxin levels (Brouwer, et al.
1990) and serum transport of Vitamin A (Brouwer, et al., 1986) in rodents.
The metabolite was identified as 4-OH-3,3',4’,5-Cl4-BP and was shown to have
a 2.5 times higher affinity for the transthyretin than thyroxine itself. This
metabolite has also been indentified in mouse blood (Klasson-Wehler, et al.
1989) and in rat and mouse fetuses after maternal exposure (Morse, et al.
1995; Darnerud, et al. 1986). Similarly, 2,3,3',4,4’-Cl5—BP forms rearranged
metabolites -- 4-OH-2,3,3',4’,5—Cl5-BP and 4-OH-2',3,3’,4',5-Cl5-BP -- in mice
and mink (KIasson-Wehler, et al. 1993). These metabolites have demonstrated
in vitro a ten times higher affinity for transthyretin than thyroxin (Lans, et al.
1992), resulting in persistent retention of these metabolites in blood of both
humans and seals exposed environmentally to PCBs (Bergman, et al. 1994).

The "disruption" of steroid hormone homeostasis and activity by
environmental contaminants has become a cause for concern (Schmidt, 1994;
Wolff, et al. 1993; Dibb, 1995). The weak estrogenic activity of commercial
PCB mixtures administered to rats was recognized early in studies of PCB
toxicity (Bitman and Cecil, 1970), however, only the less chlorinated Aroclor

mixtures (<48% Cl by weight) exhibited estrogenicity, consistent with the

14

hypothesis that lesser-chlorinated OHCB metabolites were the active
compounds (Nelson, 1974). 4-OH-2',4',6'-Cl3-BP and 4-OH-2’,3',4',5’-Cl4-BP,
which have not been identified as persistent, blood-retained metabolic products
in humans or wildlife samples, were estrogenic in a rodent uterotrophic assay
and bound competitively to estrogen receptor (Korach, et al. 1988). Also, 4-
OH-2’,4',6'-CI3-BP and 4-OH-2’,3',4',5’-Cl4-BP were estrogenic in a turtle
sexual development assay (Bergeron, et al. 1994). However, the quantitatively
significant 0HCBs retained in wildlife and human blood have not been tested
for their ability to interfere with or mimic estrogen action, and were, therefore,
the subject of this research.

Estrogen agonists are compounds that mimic the effects of estrogen.
The classic definition of an estrogen is a compound that produces cornification
of the vaginal epithelium of rodents (Solmssen, 1945). Other physiological
endpoints have been used to define estrogenicity including increased uterine
weight, uterine glycogen content, protein induction (Korach, et al. 1987) and
cellular proliferation (Soto, et al. 1992). It has been recognized that numerous
synthetic as well as naturally-occurring compounds fit the classic definition of
estrogen (McLachlan, 1993). A refined definition of estrogen, recognizing the
role of nuclear hormone receptors in gene expression, is a compound that binds
to the estrogen receptor, induces dimerization of the receptors that specifically
bind to and activate transcription of genes under the regulatory control of trans-
acting estrogen responsive elements (Kumar and Chambon, 1988). The
chemical structural characteristics of estrogens vary considerably, but can be

summarized as: 1) containing a sterically unhindered phenol group, and; 2) a

1 5
hydrophobic substituent of greater than 3 carbons bonded para to the phenolic
hydroxyl (Dodds and Lawson, 1938, Jordan, et al. 1986).

Estrogen antagonists block the action of estrogens by interfering with the
normal functioning of the estrogen receptor. Classical estrogen antagonists
such as ICI 164,384 and tamoxifen bind competitively to the estrogen receptor,
displacing the natural ligand 17B-estradiol, and blocking or reducing the
effectiveness of the ligand-bound receptor to enhance gene expression (Bondy
and Zacharewski, 1993; Fawell, et al. 1990, Jordan, et al. 1986). The potency
of classical estrogen agonists and antagonists is directly related to the binding
affinity to the estrogen receptor (Chae, et al. 1991; Korach, et al. 1987;
VanderKuur, et al. 1993a). Therefore, high affinity binding to the estrogen
receptor can be an indication of possible anti-estrogenic activity, though
receptor binding affinity alone is neither a necessary nor sufficient factor for
anti-estrogenic activity. Anti-estrogenic compounds act by several
mechanisms, not necessarily related to estrogen receptor binding and
activation. For example, aryl hydrocarbon receptor agonists, such as 2,3,7,8-
tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD) and non-ortho-chlorinated PCBs,
cause down regulation of estrogen receptor and may also interfere with DNA
binding (Zacharewski, et al. 1991). Aromatase inhibitors, such as
aminoglutethimide, block conversion of testosterone to 17B—estradiol and are
used in the treatment of metastatic estrogen-dependent brease cancer (Farooqi
and Aboul-enein, 1992). lnducers of Phase I and Phase II metabolic
detoxification enzymes, such as 2, 3,7,8-TCDD and non-ortho-chlorinated PCBS,

reduce the estrogen-dependent expression of proteins by enhancing the

16
enzymatic conversion of latent 17B—estradiol, thus eliciting an anti-estrogenic
response (Spink et al. 1992).

The objectives of this study were: 1) to determine the estrogenic or anti-
estrogenic activity of tri-to hexa- chlorinated 0HCBs (representing
environmental and laboratory derived metabolites) in an in vitro bioassay using
an estrogen-responsive luciferase reporter gene system; 2) to estimate the
potency of the 0HCBs in the presence of physiologically-relevant concentrations
of 17B—estradiol, and; 3) to relate the bioassay activity to molecular structure,

human estrogen receptor binding affinity, and cytotoxicity.

Materials and Methods

Chemicals

A total of thirteen 0HCBs were studied, nine of which were confirmed
mammalian metabolites of CBs (Table 1). OHCBI to OHCB9 were synthesized
for use in this study and as analytical standards according to previously
published methods (Bergman, et al. 1995, McOmie, et al. 1968). OHCBIO to
OHCBl3 were obtained from Ultra Scientific (North Kingston, RI). All 0HCBs
were authentic standards of greater than 95% purity. Stock solutions were

prepared in ethanol and stored tightly sealed at -20°C.

17

Table 1. Environmental occurrence and structures of 0HCBs and 178-estradiol
assayed with MOF-7-LUO cells.

ID Formula Structure Identified In Slnples P: E"; gnmjn.!

OHCB1 3-0H-2,3',4,4',5-CI.-8P No published reports P081187

Rat. grey seel. human blood (1) PC8108

 

OHCB: ‘QOH'2'3'3 '4 's'Ck-BP Mink, mouse blood/feces (3) P08118 Y”
0H083 3-OH-2',3',4,4',5-CI.-8P Mink, mouse Ieces I3) P08105 No
CI
Cl 0| 0|
0HCB4 “IN-23333435. 5'-C|.-BP "0‘3”“ not. any seal blood (1) P08158 Yes
0H085 4-OH-2',3,4',5,5‘-0l,-8P “’an Rat. gray seal blood (1) P08118 Yes
0|
0|
OH086 +OH-3.3',4'-CI,-8P ”“0“ Mouse feces 141 P0877 No
Cl HO Cl CI
Cl
OHCB? 2-OH-2'.3.3’.4.4'-CIs-8P a“ Mink. mouse feces/blood I3) P08105 No
0| 0|
HO Cl
011088 4-0H-3,3',4',5,5'-Cl,-8P No published reports PC8128?
Cl Cl
0| Cl Cl
, , , no Cl Mink, mouse feces (3)
0HCBs To.” '3’3 " 's'c"'BP Gray seal. hunen blood (1) News Y”
0| 0|
HO CI
OHCBIO 4-OH-2’,4',6'-0l,-BP Ret bloodl2l P0830 No
Cl
CI 0|
HO CI
OHCB11 4.0H-2',3',4'.5'-CI,-8P No published reports P0881?
CI
no a Cl
Cl
0H0812 3-OH-2’,3',4',5'-CI.-8P No published reports P0881?
0|
0| 0|
0H0813 4,4‘-(OH),-3,3',5,5"0l.-8P WHO" Mouse feces l4) P0877 No
c1 c1
OH
E2 178-Estradiol

  
 
 

 

:erg'nen,et .. ' : ' ' - r. ' :
P0881: 2,3,4, 5-Cl-8P; P0877 3, 3' ,,4 4’Cl ~8P; P08105: 2, 3, 3’, 4 ,4'-0l. 8P;P08118: 2. 3’, 4, 4' ,;5-CI,-8P P08128: 3, 3', 4, 4',

010, 8! 8 .

 

PCB156: 2, 3, 3' ,,4 4’ ,.5-Cl ~8P.

5-Cl.-8P;

  
 

18

Human Es tragen Recep tar Binding Assay

Competitive binding assays were conducted using a modification of the
hydroxyapatite (HAP) binding procedure of Murdoch, et al. (1990). A human
adenocarcinoma cell line, MCF7 (American Type Culture Collection HTB22,
Rockville, MD), was the source of human estrogen receptor (hER) for the
binding assays. MCF7 cells were cultured in 75 cm2 plates at 37 °C in a
humidified atmosphere of 5% 002/95% air in 10% FBS medium consisting of
Dulbecco's Modified Eagle's medium with Ham's F12 Nutrient Mixture without
phenol red and with 15 mM HEPES buffer (Sigma Chemical, D2906, St. Louis,
M0) at pH 7.4 supplemented to contain 10% (v:v) fetal bovine serum (Hyclone
Laboratories, Lazan, UT), 2 mM sodium pyruvate, 1 pg/ml bovine insulin, and
10 U/ml:10pg/ml penicillin:streptomycin. Four days before harvest, the 10%
FBS medium was changed to 5% DCCFBS medium in which the serum
supplement was treated with dextran-coated charcoal (Hyclone Laboratories)
to reduce steroid hormone concentrations. Individual lot analysis of fetal bovine
serum provided by the manufacturer indicated that dextran-coated charcoal
treatment reduced the 17B—estradiol concentration from 8 pg/ml (29 pM) to 5
pg/ml (18 pM). Twenty four hours before harvest, the 5% DCCFBS medium
was replaced with serum-free medium to minimize the level of 17B—estradiol in
the cells and to improve the recovery of unliganded hER. Cells at 85%
confluence were gently suspended by a 30 min incubation in 1 mM EDTA in Ca-

Mg free phosphate buffered saline at 37°C. Cell suspensions were combined

19

and centrifuged at 8009 for 10 min. at room temperature to pellet cells. The
cells were washed once with homogenization buffer (10 mM Tris HCI pH 7.5,
1.5 mM NazEDTA, 1 mM dithiothreitol, 1 mM sodium molybdate, 0.5 mM
aminoethylbenzenesulfonyl fluoride ( a protease inhibitor) and 10% v:v glycerol),
resuspended in 2 ml ice cold homogenization buffer, disrupted with 60 passes
Of a Dounce homogenizer on ice, and centrifuged at 8009, at 4°C for 10 min.
to pellet cellular debris. A molybdate-stabilized hER preparation was obtained
by centrifugation of the supernatant at 100.0009 at 4°C for 30 min. The hER
preparation was stored in 200 pl aliquots at -80°C until use. Scatchard plot
analysis of the hER preparation indicated a 17B-estradiol specific binding
activity of 0.22 pmol/mg protein with an apparent dissociation constant of
2.3 :1: 0.4 nM (slope :l: 1 SE, df = 10). Protein was determined using the Bradford
dye binding assay in 1 ml homogenization buffer (Bradford, 1976).

Inhibition of 10 nM 3H-17IS—estradiol (3H-E2) binding to hER was
measured by incubation at 4°C for 2 hr. of 40 pM hER in 1 ml Assay Buffer (10
mM Tris HCI, pH 7.5, 1.5 mM NaZEDTA, 1 mM dithiothreitol, and 10% v:v
glycerol). Duplicate analyses were conducted at four half-decadal
concentrations ranging from 70 to 0.1 pM for OHCBl-9,12-13 and 3 decadal
concentrations from 10 to 0.01 pM for OHCBIO-11. Non-specific binding of
3H-E2 was estimated by adding a 1000 fold excess of diethylstilbestrol to
separate duplicate tubes and was used to define the 0% binding level. Total
binding of 3H-E2 was estimated in the absence of competitor and was used to

define the 100% binding level. Final ethanol concentration in the assay was

20

1%. 3H-E2 bound to hER was extracted from the assay solution by addition of
300 pl of 0.16% (wzv) HAP slurry that was gently mixed periodically over the
15 min adsorption period. The HAP-bound 3H-E2/hER was filtered onto a 24
mm Whatman GF-C disk in a 10-place stainless steel vacuum manifold at 4°C
and washed six times with 5 ml assay buffer to remove unbound 3H-E2. The
filter and HAP were placed directly into liquid scintillation vials, to which 1 ml
ethanol and then 6 ml aqueous scintillation cocktail were added. 3H-E2 was
quantitated by 10 min counts of the full 3H window and calibrated using the
standard addition technique. The log1ol050 (log1o transformed concentration of
inhibitor causing a 50% reduction in specific binding of 3H-E2) was used as a
measure of hER binding affinity and fitted to a normal distribution that was

calculated using a non-linear regression technique (Bruce and Versteeg, 1992).

Luciferase Expression Assay

Functional estrogen agonism/antagonism of the 0HCBs was determined
using a stably-transfected MCF7 cell line (MCF7-LUC, a gift of Prof. M. Pans,
INSERM, MontPelier, France) expressing an estrogen-responsive reporter gene
construct—-pVit-tk-luc (Pans, et al. 1990). This plasmid consists of the Xenapus
laevis vitellogenin A2 promoter region (containing estrogen responsive
elements) upstream of the Herpes simplex virus thymidine kinase promoter
controlling expression of the firefly luciferase gene. MCF7-LUC cells maintained

in 10% FBS medium were transferred to 96 well culture plates at a density of

21
10,000 cells/well. After 24 hr, the medium was replaced with 125 pl 5%

DCCFBS medium. At the beginning of the exposure period, cells were at 85%
confluence. Forty eight hours after plating, cells were exposed to the 0HCBs
dissolved in 5% DCCFBS medium with a final ethanol concentration of 0.5%.
Four concentrations were tested in duplicate such that the concentrations were
20%, 10%, and 2% of the highest concentrations tested. The greatest
concentrations tested were: OHCB6 37 pM; OHCBS, OHCB9 30 pM; OHCBI ,
OHCBZ, OHCBS, OHCB5, OHCB7 29pM; OHCB4 27 pM; and OHCBIO,
OHCB1 1, OHCB12, 0H0813 25 pM. Control wells (four per plate) receiving no
solvent and solvent control wells (four per plate) receiving 0.5% ethanol were
also tested. Two plates were required to test all 0HCBs. Typical in vitro
estrogenicity assays, including the present study, have used charcoal-stripped
medium to reduce background response to estrogens (Pans, et al., 1990; Soto,
et al., 1991). To fully assess the agonistic/antagonistic properties of the
0HCBs, the response of MCF7-LUC cells at higher E2 concentrations was also
determined. Therefore, two additional complete experiments were conducted
using 5% DCCFBS medium supplemented with E2 at 0.1 nM and 1.0 nM,
similar to concentrations found in human plasma (Becker, 1990). The
concentration of E2 in unsupplemented 5% DCCFBS medium was 0.0009 nM.

Luciferase expression of the chemically-lysed cells (Promega Corp,
Madison, WI) was measured using a Dynatech ML3000 plate reading
luminometer (Dynatech Laboratories, Chantilly, VA) and calibrated to standard

solutions of firefly luciferase dissolved in lysis buffer supplemented with 5

22

mglml bovine serum albumin. Method detection limit was 0.014 pg luciferase
with a relative standard deviation of 5.1% (20 pl lysate sample volume)
Luciferase activity was normalized to soluble protein by assaying 10 pl lysate
in 1 ml concentrated Bradford reagent to overcome interference by detergent

components in the lysis buffer.

Cell Viability Assay

It has been reported that 0HCBs are uncouplers of oxidative
phosphorylation in vitro (Narasimham, et al. 1991). Therefore, the potential for
cytotoxicity was considered significant and required the monitoring of cell
viability. In addition to visual inspection of cell morphology by light microscopy
(100x magnification), a cell viability assay kit (Molecular Probes Inc. Eugene,
OR) was employed utilizing the selective uptake of fluorogenic substrate to
quantify "viable" and "non-viable" cells. Viable cell activity was measured as
non-specific esterase activity by the selective uptake and subsequent enzymatic
cleavage of calcein AM. Production of the fluorescent product was measured
using a plate scanning fluorometer (Cytofluor 2300, Perseptive Biosystems,
Framingham, MA) with excitation/emission at 485/530 nm. Non-viable cells
were measured simultaneously by uptake through damaged cell membranes and
binding to DNA of ethidium homodimer, producing a fluorescent complex with
excitation/emission at 530/645 nm. A cell viability index (Equation 1) was

calculated to normalize for differences in the total number of cells in each well

23
and also for the effects of E2 treatment which caused increased cell

proliferation:

Cell Viability In de x ___ Viable Cell fluorescence
Nan Viable Cell Fluorescence

 

(1)

where: fluorescence values were untransformed relative fluorescence

units.

H4IlE EROD lnduc tion Assay

Aryl hydrocarbon receptor agonists, including some non-ortho-substituted
PCBs, have been reported to elicit a response resembling anti-estrogenicity in
MCF-7 cells (Krishnan and Safe, 1993). Therefore, a subset of 0HCBs
(OHCBI , OHCB3, OHCB5, OHCB6, and OHOB8) exemplifying several structural
features, was selected for assay of aryl hydrocarbon receptor agonism in the
H4IIE rat hepatoma EROD (ethoxyresorufin-O-deethylase) induction assay

(Sanderson, et al. 1996).

Statistical Analyses

Statistical analyses were conducted using the Statistical Analysis System

Version 6.03 (SAS® Institute, 1988). Luciferase expression response curves

were modeled without transformation after a normal distribution fitted by a non-

24

linear regression technique (Bruce and Versteeg, 1992). For inhibition of
luciferase expression, a two parameter model (parameters in bold) was used

(Equation 2):

199(5050 - Iog(OHCB)
SLOPE

Luc = Lucmax x P

 

(2)

where: Luc = luciferase expression (ng/mg protein);
Iog(OHCB) = log1o OHCB concentration, nM;
P[ ] = normal probability density function;
Lucmax = maximum luciferase expression, mean of solvent control
response for each E2 concentration;
Iog(E050) = log1o Effective Concentration causing 50% inhibition
of luciferase expression, nM;
SLOPE = slope function (standard deviate) of normal probability
density function.
Log(E050) was used as a measure of potency such that greater values reflect
lesser potency. SLOPE was an estimate of dispersion and described the
steepness of response such that lower values describe steeper curves. To aid
convergence of the nonlinear regression, SLOPE was bounded,
0.258LOPES3.5. To estimate the OHCB threshold for effects concentration,

the Iog(E020) was calculated using a modification of Equation 2 (Equation 3):

l_o_g(ECZQ - Iog(OHCB)

Luc = Lucmax x P
SLOPE

 

+40 (3)

25

where: Iog(E020) = log10 Effective Concentration causing 20% inhibition
of luciferase expression, nM;
220 = 0.8416 = normal deviate above which 20% of the standard
normal distribution lies;
and all other variables are as previously described.
For induction of luciferase expression, an additional constant was added to

define the basal expression level (Equation 4):

Iog(OHce - logiEO50

Luc = (Lucmex- Lucmin)x P
SLOPE

+ Lucmin (4)

 

where: Lucmin = mean luciferase expression of solvent controls at 0.0009
nM E2;
and all other variables as previously described.
Log(E020) was used to estimate the luciferase induction threshold for effects

concentration of OHCB (Equation 5):

Iog(OHCB) 409(qu - 40 +Lucmin (5)

Luc = (Lucmax- Lucmin)x P
SLOPE

where: all variables are as previously described.
Calculation of hER binding potency used an inhibition model similar to the

luciferase expression model (Equation 6):

I_o_g (rose) £9 ( OHCB)
SLOPE

 

Bound = Bmax x P (6)

26

where: Bound = 3H-E2 specifically bound to hER as a percent of total
specifically bound;
Iog(l050) = log1o Inhibitory Concentration causing 50% decrease
in specifically bound 3H-EZ;

and all other variables as previously described.

Results and Discussion

Human Es tragen Recap tar Binding

All OHCBs tested in this study caused measurable reduction in specific
binding of 3H-E2 to hER in vitro (Table 2). OHCB7 and OHCB8 weakly inhibited
binding of 3H-E2 and, therefore, had Iog(l050)s extrapolated beyond their
solubility limits which were estimated empirically to be 50 pM (Iog(nM) =4.7)
by the formation of a visible precipitate in the assay medium. The Iog(l050)s
for the remaining 0HCBs ranged from 4.3 to 2.5 Iog(nM) and when compared
to E2 exhibited a relative binding inhibition (RBI) range of 0.079-5.0%. The
two most important structural measures that predicted inhibition of binding to
hER were: 1) substitution position of the hydroxyl group, and; 2) minimum
energy torsion angle of the central biphenyl bond. Steric hindrance of the
hydroxyl group did not diminish binding potency when 0HCBs with similar
hydroxyl positions were compared. For example, OHCBI1 possesses an

unhindered para-substituted hydroxyl with a binding potency of 2.5% with

27

Table 2. Summary of MCF-7-LUO cellular responses to 0HCBs.

Luciferase Expression

 

 

 

 

Log(EC20) Log(ECSO) Slope‘
(LogInM) (LOOIOM) (ASE)
(ASE) (ASE)
hER Binding 17&Estrediol (nM) 178-Estrediol (nM) 178-Estrediol (nM) Cyto-
toxicity
ID Formule Log(lCSO) Slope RBI”; 0.0009 0.1 1 0.0009 0.1 1 0.0009 0.1 1 Viability
(Log(nM) (ASE) I 96) lndexlSE)
(ASE‘)
-=II=
OHCB1 MH—2,3',4,4',5-0L-8P 3.3 1.4 1.0 3.2 3.7 4.5 3.8 4.3 4.6 -0.9 -0.7 -0.2 1.137
(0.2) (0.4) (0.4) (0.6) (-l (0.2) (0.5) (~) (0.4) (0.7) H (0.024)
0H082 4-OH2,3,3',4',5-Ck-8P 3.3 1.0 1.0 3.9 3.8 >4.5 >4.5 4.0 >4.5 -3.5 -0.2 NA‘ 1.072'
(0.2) (0.3) l-) (0.3) (0.2) H (2.2) (0.026)
0H083 MH-2',3',4,4',5-CI.-8P 3.0 2.7 2.0 -0.03 3.3 >4.5 1.7 >4.5 >4.5 -2.1 -2.2 NA 0.774"
(1.2) (4.2) I 1 .3) (1.3) (0.6) (0.8) (7.4) (0.025)

01-1084 4-0H-2',3,3'.4',5, 5'-Cl.- 2.8 3.5 3.2 -2.7 3.6 >4.4 0.3 3.8 >4.4 -3.5 -0.2 NA 1.036"
6? H I-) I-) (-l (-I H (-I (-I (0.029)
0HCB5 4-OH-2'.3.4'.5.5'-C|.-6P 3.4 0.9 0.8 2.0 4.3 > 4.7 3.1 4.5 > 4.7 -1.4 -0.2 NA 1.062

 

(0.1) (0.3) (0.5) (-l (0.2) (0.6) (0.5) (0.2) (0.061)
OHCB6 4-0H-3.3',4‘-Cl,—BP 3.6 0.7 0.3 1.3 >4.6 >4.6 2.6 >4.6 >4.6 -1.6 NA NA 0.969"
(0.1) (0.2) (1.9) (0.7) (1 .4) (0.026)
0HCB7 2-OH-2',3,3’,4,4'vCl..6P 5.7 3.5 0.004 -2.5 >4.5 4.3 0.4 >4.5 >4.5 -3.4 NA -0.7 0.626' '
(0.2) H (6.6) (5.4) (3.3) (3.9) (6.5) (0.027)
OHCB6 4-0H-3,3'.4',5,5'-Cl,-8P 5.1 0.9 0.02 ~1.0 >4.5 >4.5 1.6 >4.5 >4.5 .3.1 NA NA 0.913”
(1.6) (1 .4) (7.6) (3.1) (5.4) (0.032)

0HCB9 4-OH-2',3,3’,4',5-Cl.~8P 4.2 0.2 0.1 -0.7 >4.7 >4.7 2.2 >4.7 >4.7 -3.5 NA NA 1 .133
I-) (-I I-) l-) (-I (0.023)

0HCB10 4-0H-2'.4',6'-Cl,-8P 2.5 2.6 6.3 3.9 2.8 2.4 4.2 >4.7 >4.7 +0.4 -3.5 ~3.5 1 .106
(1 .2) (3.8) (0.4) (-I (-I (0.2) (0.3) (-I H (0.024)

0HC81 1 4-0H-2',3',4',5'-CI.-8P 2.8 2.3 3.2 4.4 2.0 2.1 4.6 >4.7 3.8 +0.2 -3.4 -2.1 1.109
(0.6) (1.8) (-I (2.3) (0.6) I-) (0.2) (0.03) (4.5 (0.6) (0.016)
OHCB12 3-Ol-I-2',3',4',5'oCl‘-8P 4.3 0.8 0.1 2.6 2.3 2.2 3.6 4.1 >4.4 -1 .2 -2.2 -3.5 1.026' '
(0.7) (1 .4) (0.7) (1 .5) H (0.3) (1.2) (0.9) (2.9) (-I (0.026)

OHCB13 4.4'-(0H),-3.3'.5,5'-Cl.-8P 3.7 0.4 0.4 3.2 1.3 «0.03 >4.7 4.3 2.9 -2.2 -3.5 -3.5 1.224
(0.3) (0.5) (0.8) I-) H (-I (~) (2.2) H I-) (0.020)

E2 17&Estrediol 1.3 1.6 100 -3.9 ~2.8 +1.2 1 .156
(0.8) (1.5) (0.4) (0.3) (0.7) (0.016)

 

mnutimable, Jecobian matrix singular.

‘RBIaO: Relative Binding Inhibition for l050 =IIC50,,/Icsomi100%.

‘Slgn of slope: + -lnduction of luciferase expression, - nlnhibitlon of luciferase expression.

‘NA: Not Active.

'p<0.05; ° 'p<0.01, Student’s t-test comparison with E2.

respect to E2. In comparison, OHCB4, with a structure similar to OHCBl 1 but
with the addition of two Cl's adjacent to the hydroxyl, has a relative binding
potency of 4.0% with respect to E2. The effect of minimum energy torsion
angle is exemplified by comparison of OHCB4 and OHCB8. OHCB8, similar to
OHCB4 except that OHCBS lacked an ortho-substituted Cl and could assume

more planar configuration than OHCB4, had a relative binding potency of

0.01%. Subsitution position of the hydroxyl group strongly affected hER

28

binding affinity. For example, meta-hydroxylated OHCB12 exhibited a Iog(lC50)
of 4.3 Iog(nM) (RBI = 0.1 %) compared to its para-hydroxylated isomer OHCBI 1

with a Iog(l050) of 2.8 Iog(nM) (RBI =3.2%).

C all Viability

The cell viability index was significantly reduced by OHCB2, OHCBS,
OHCB4, OHCB6, OHCB7, OHCB8, and OHCB1 1 at the highest concentrations
tested (Table 2). No significant changes in the cell viability index occurred at
lower OHCB concentrations (data not shown). Visible changes in cell
morphology, including vacuolization, lysis and detachment were evident at cell
viability indices less than 1.030. Visible, and therefore severe, cytotoxicity was

caused by OHCB3, OHCB6, OHCB7, OHCBB, and OHCB12.

Effects of £2 and 0HCBs an Luciferase Expression

Luciferase expression was induced 2.2 fold in MCF7-LUC cells by 0.1 nM
E2 over the basal expression in the presence of 0.0009 nM E2, but only 2.3
fold by 1 nM E2 indicating near maximum induction was achieved at 1 nM
(Figure 3; luciferase ng/mg protein at 0.0009, 0.1, and 1 nM E2; 4.09
(SE=0.31, n=32). 8.96 (SE=1.00, n=8), 9.55 (SE=0.57, n=8).
respectively). E050 for E2 induction of luciferase in the absence of OHCB was

0.0015 nM(Iog(E050) =-2.81, asymptotic standard error=0.28, n =48).

29

0HCBs tested in this study were "estrogenic" (induced luciferase
expression), "anti-estrogenic" (reduced luciferase expression) or inactive (did
not affect luciferase expression) depending upon the concentration of E2 co-
administered or present in the medium. Cytotoxicity confounded the
interpretation of anti-estrogenicity, because reductions in luciferase expression
could have been caused by reduced cell viability as well as inhibition of hER-
mediated transcription. Therefore, OHCB concentrations that caused
statistically significant reductions in the cell viability index were eliminated from
the calculation of Iog(E050) and Iog(E020) values (Figure 3). An upper
concentration limit of potential activity was defined for each OHCB as either the
limit of solubility or the concentration causing statistically significant
cytotoxicity. Predicted Iog(EC 50) and Iog(E020) values that fell above theupper
concentration limit were labelled "Not Active" (Table 2). In general, E2 co-
administration reduced the potency of the "anti-estrogenic" 0HCBs, as reflected
in . increasing threshold (Iog(E020)) and mid-range (Iog(E050)) effect
concentrations (Table 2). The "anti-estrogenic" 0HCBs that followed this
pattern were OHCB1-9 and OHCB12 (Table 2, Figure 4A-l, L). OHCBlB was
an exception in eliciting reductions in luciferase expression at lower OHCB
concentrations as E2 concentration increased (Table 2, Figure 4M). OHCBIO
and OHCB11 caused greater luciferase expression in the presence of 0.0009
nM E2 (Figure 3) but lesser luciferase expression at 0.1 nM E2 and 1 nM E2.
Thus, OHCB10 and OHCB11 were "partial estrogens" because the inductive

response at 0.0009 nM E2 was reversed at 0.1 nM E2 and 1 nM E2.

30

Figure 3. Representative luciferase expression of MCF-7-LUC cells exposed to
0HCBs and E2 in the assay medium: OHCB4: 4-OH-2’,3,3',4',5,5’-CI6-biphenyl;
OHCB10, 4-OH-2',4',6'-CI3-biphenyl; OHCBII 4-OH-2',3’,4’,5’-Cl4-biphenyl;
E2 17(3-estradiol; (A) ng luciferase/mg protein, n=2; (A) n9 luciferase/mg
protein, points omitted from calculation of Iog(E050) and Iog(E020) due to
cytotoxicity; symbol and bars indicate mean :tSE.

 

31

Azcvmoxo

 

 

 

 

 

 

 

 

 

 

 

 

40—. «OF «OF 0 «OF 00—. «Or 0 e0_. new «OF 0
s. w . w 1 2 w m .
5:. 886 H
_Nm_ -
L L .
. 4
M w s“ _ H m
s A . .3.

«I. . - . b 4 r .
4. 4. _. 4.
ECF 4 « . .
am. 2 «moxou :mOIOH A:mo:o

0‘

 

 

DD

‘5

 

 

I“

1“

“

NP

.2.

(ugeIOId 6w/6u) esaiejgan'l

32

In order to estimate the potency of the 0HCBs, an integrated measure of
activity was calculated to take into account the effect of E2 concentration on
OHCB activity. Typically, potency is estimated by the value of the E050, or
some similar effective concentration. However, because OHCB potency varied
with E2 concentration, the overall potency of the OHCB was estimated by
calculation of the area of activity defined as the area bordered by the highest
and lowest E2 concentrations on the abscissa and on the ordinate by the curves
defined by the effective concentrations (both E020 and E050) and the upper
solubility or cytotoxicity concentration limits (Figures 4A-M). The total area of
activity, defined by the Iog(E020) response curve (delineating the greater than
20% effective area), and the strong activity area, defined by the Iog(E050)
response curve (delineating the greater than 50% effective area), were
calculated (Table 3; see shaded areas on Figures 4A-M). For OHCB10 and
OHCBI 1, the concentration of E2 at which the response changed from
induction to reduction was estimated by graphical interpolation (Figures 4J-K).
It was assumed that at the "cross-over" E2 concentration, the net OHCB effect

on luciferase expression was nullified.

33

Figure 4. Effect of 178-estradiol (E2) co-exposure on luciferase expression
response to 0HCBs: (I) E050, OHCB Effective Concentration causing a 50%
reduction (or induction) of luciferase activity with respect to the solvent control
response; (0) E020, OHCB Effective Concentration causing a 20% reduction
(or induction) of luciferase activity with respect to the solvent control response;
(----) Relevant limits of concentration of 0HCBs and E2 were determined for the
calculation of the "area of activity" such that the upper limit of OHCB
concentration defined as either the maximum soluble concentration (applied to
OHCBI,5,9,10,11,13) or the cytotoxic concentration (0H082,3,4,6,7,8,12);
E2 limits defined by highest (1 nM) and lowest concentrations tested (0.0009
nM); (NA) Not active, predicted E050 or E020 greater than upper OHCB
concentration limit; (Light Gray Shading) Area of weak activity depicting an
expected response of 20% to 50%; (Dark Gray Shading) Area of strong activity
depicting an expected response of greater than 50%. All OHCB EC50s and
E0203 depict reduction (E2 antagonism) of luciferase expression except for
OHCB10 and OHCB11 that caused induction (E2 agonism) of luciferase
expression at 0.0009 nM E2; (A) OHCB1, 3-OH-2,3',4,4',5-Cl5-biphenyl; (B)
OHCBZ, 4-OH-2,3,3',4',5-Cl5-biphenyl; (C) OHCB3, 3-OH-2',3',4,4’,5-Cl5-
biphenyl; (D) OHCB4, 4-OH-2’,3,3’,4’,5,5’-0l6-biphenyl; (E) OHCBS, 4-OH-
2’,3,4’,5,5’-Cl5-biphenyl; (F) OHCBB, 4-OH-3,3'4,’-Cl3-bipheny|; (G) OHCB7, 2-
OH-2’,3,3',4,4’-0l5-biphenyl; (H) 0HCBB, 4-OH-3,3’,4’,5,5’-CI5-biphenyl; (l)
OHCBQ, 4-OH-2’,3,3’,4’,5-Cl5-biphenyl; (J) OHCBIO, 4-OH-2',4’,6'-CI3-
biphenyl; (K) OHCB11, 4-OH-2',3', 4’,5’-CI5-biphenyl; (L) OHCBIZ, 3-OH-
2’,3’,4',5'-Cl5-biphenyl; (M) OHCBI3, 4,4’-(OH)2-3,3’,5,5'-Cl4-biphenyl.

34

 

 

 

 

 

 

 

2:52.... 596

178-Eetredlol (Log(nlm

178-Eltrldlol (Log(nMD

 

 

  

 

 

 

 

 

 

 

 

-l
a

23500.: «more

178-Eetredlol (Log(nun

178-Eetredlel (Log(nMn

 

 

 

 

 

 

 

 

:25qu. 80:0

178-E etudlol (Log(nlm

178-Eetredlol (Log(nMn

 

 

 

 

 

 

 

 

 

r

 

2 1 o

.3539 5010

r
4....3

178-Eetredlel (Log(nl”

17B-Eetredlol (Log(nu»

Figure 4. (Cont’d).

35

36

 

 

 

 

 

 

 

 

 

 

A
N
I | 5 I I I I: o
w w
a m.
m M
o .u
m u
a a
s . 9.. s
n n
lllllll l. ....
.L _
_ . . . L . e
5 4. 4 3 2 1 0 1 2 3
4 . . .
22593. uwmozo
A
N
. lllllllllll o llllllll o

 

    

 

 

 

 

 

 

178-Eetredlol (Log(nlm

 

 

 

 

 

)
m, D
....rn. .4m
0 I
° '
o
u .I.
m m
d
m a
e n
E
.4...» .4”
W M 7
m 1
a
u
n
A
.3 m 3
III... llllflllllll'h.
|_ K. n
—.I....ui —......
8.432104%4 57.43210123
4 4

2:58.: 30:0 235.3. :35 225352510

37

Mechanisms of activity and rala tionship with chemical structure

Partial Estragenic Activity

OHCBIO and OHCBlI exhibited partial estrogenic activity in this
experiment, consistent with the observations of estrogenic activity in a rodent
uterotrophic assay (Korach, et al. 1988) and turtle sexual development assay
(Bergeron, et al. 1994). There appeared to be quite specific structural features
that conferred partial estrogenic potential to these 0HCBs. For example, a
para-substituted hydroxyl was essential for partial estrogenic activity. Meta-
substitution of the hydroxyl led to the anti-estrogenic activity of OHCBI 2 while
the para-substituted structural isomer, OHCBI 1, exhibited partial estrogenic
activity. Meta-substitution of the hydroxyl in OHCB12 also conferred
cytotoxicity, whereas OHCBII was not cytotoxic. Another important feature
was the presence of chlorine atoms vicinal to the para-substituted hydroxyl.
OHCB4 possessed two CI atoms vicinal to the para-hydroxyl group and was
strongly anti-estrogenic, while the structurally similar but not vicinally Cl-
substituted OHCB1 1 exhibited partial estrogenic activity. Vicinal Cl substitution
may have two effects on OHCB activity: 1) steric hindrance of a hydrogen
bonding moeity on the estrogen receptor that bonds to the ligand hydroxyl; 2)
reduction of the electron density on the hydroxyl oxygen by Cl -mediated
electron withdrawal through the aromatic ring, with concomitant decrease in
acid dissociation constant, pKa (Waller, at al. 1995). Waller, et al. (1995)

estimated pK, of vicinal Cl substituted 0HCBs in the range of 5.8 to 6.7, which

38

Table 3. Potency of anti-estrogenicity/estrogenicity of 0HCBs estimated by the
area of activity.

Area' Overlap Area"

 

 

ILogInMI’I ILogInMI’I ”3“"

Type of ”t!

)0 Formula Response Strong“ Weak' Total' Strong Weak (L “(81:02)
Activity Activity Activity Activity Activity °9 "

= a
0H081 3-OH-2,3',4,4',5-0l5-BP Reduction 1.37 1.81 3.18 0 0 6.86
OHC82 4-OH-2,3,3',4',5-Cl5-BP Reduction 0.79 0.83 1.62 0 0 7.51
0H083 3-OH-2',3',4,4',5-CI,-8P Reduction 2.84 3.53 6.37 0 o 3.76
OHCB4 4-0H-2',3,3',4',5,5’-Cl,—BPReduction 5.19 3.27 8.46 0 0 3.63
OHCB!) 4-OH-2',3,4',5,5'-CI,-8P Reduction 1.97 1.46 3.43 0 0 6.72
011086 4-0H-3,3',4'-0l3-8P Reduction 1.98 1.41 3.39 o 0 6.46
OHCB7 2-0H-2',3,3',4,4'-CI5-BP Reduction 4.16 3.05 7.21 o o 4.57
OH088 4-0H-3,3',4',5,5’-0l5-8P Reduction 2.97 2.65 5.62 o o 5.12
011089 4-OH-2',3,3’,4',5—Cl5-BP Reduction 2.53 3.01 5.54 0 0 5.28
OHCB10 4-OI-I-2',4',6'-CI,-8P Induction 0.14 0.10 0.24 0 0 9.62
Reduction 0.00 3.55 3.55 0 0 6.27
OHCBII 4-OH-2',3',4',5'-CI,-8P Induction 0.02 0.02 0.04 0 o 9.78
Reduction 0.89 4.22 5.11 0 0 5.05
0H0812 3-OH-2',3',4',5'-CI,-8P Reduction 1.27 4.94 6.21 0 0 2.74
0HCB13 4,4'-IOHI,-3,3',5,5'-CI,-8P Reduction 1.56 7.51 9.07 0 069° 1.73,0°

= m

 

'Area: Calculated from plot of OHCB Iog(E050) and Iog(E020) vs. Iog(E2) (See Figures 4A-2M).
l’Overlap Area: Overlap between area of in vitro effects and area of possible in viva concentrations of
0HCBs and E2 (See Figure 6).

‘Margin-of—Safety Area: Extent of non-overlap between area of in vitro effects and area of possible in viva
concentrations of 0HCBs and E2 (See Figure 6.)

“Strong activity: Strong activity area defined as the area bounded by loglEZ)=-3.05 to 0.00 and by
Iog(OHCB) =Iog(E050) to Iog(solubility) or Iog(cytotoxic concentration). The strong activity area
represents the combinations of logIOHCB) and Iog(EZ) that result in a response (either inhibition or
induction of luciferase expression) greater than 50% of solvent control responses.

'Weak activity area defined as the area bounded by Iog(E2) = -3.05 to 0.00 and by logIOHCB) - Iog(E020)
to upper bound, where the upper bound was either the Iog(ECSO) or the log(solubility)llog(cytotoxic
concentration). The weak activity area represents the combinations of logIOHCB) and Iog(EZ) that result
in a response ranging from threshold (20%) to mid-range (50%).

'Total activity=Str0ng activity area + Weak activity area. The total activity area represents the
combinations of Iog(OHCBI and Iog(E2) that result in a response greater than the threshold (20%).
“For E2<O.1 nM, there was no overlap area, for E2>O.1 the margin-of-safety area was zero.

would lead to 83.4 to 98.8% ionization of the hydroxyl group in assay medium
at a pH of 7.4. In contrast, 0HCBs without vicinal 0| substitution were

estimated to exhibit pKa values of 9.5 to 10, which would result in ionization

39
of 0.79 to 0.25% (Waller, et al. 1995). Exactly how a reduction in pK, could

result in anti-estrogenic activity is unknown. Total inductive potency (greater
than 20% inductive effect) and strong inductive potency (greater than 50%
inductive effect) appeared to be inversely related to hER binding Iog(l050)
(Figure 5), such that lower hER binding Iog(l050) (hence greater hER binding
affinity) was related with greater inductive potency. Such a relationship was
expected for a receptor-mediated mechanism of action (VanderKuur, et al.
1993a). However, the relationship between the anti-estrogenic potency of
OHCB10 and OHCBIl with hER binding Iog(l050) was reversed (Table 3),
possibly because the more estrogenic OHCBIO caused a less inhibitory
alteration of the E2 activated receptor complex, which could take on the form

of a heterodimer containing both E2 and OHCB10 ligands (Klinge, et al. 1992).

Anti-estrogenic Activity

If the activity of the solely anti-estrogenic 0HCBs (all but OHCBlO and
OHCB1 1) were mediated by competitive binding to the hER and displacement
of E2, then it would have been expected that anti-estrogenic potency was
related to hER binding Iog(l050). In fact, there was no significant relationship
between hER binding Iog(l050) and either total (>20% reduction) reductive
area or strong (>50% reduction) reductive area (Figures 6A and 68,
respectively). Rather, a better predictor of both total reductive potency and

strong reductive potency was the mean viability index of the highest OHCB

40

 

 

 

 

0.30
E no —I—# I :lI—TI :- cu .
a: 0.20 P °‘
:2, OHCBIO
(.0
2 A
< 0| Cl
C no 01
'3 0.10 .
0
g OHCB11
E
O
A
0.00 ‘
2.00 2.25 2.50 2.75 3.00

hER l050 (Iog(nM))

Figure 5. Relationship between OHCB inductive (E2 agonist) activity area and
human estrogen receptor (hER) binding affinity represented by l050, the OHCB
Inhibitory Concentration causing a 50% decrease in specific binding of 3H-E2
to hER in vitro: (0) Total induction area (response greater than 20% with
respect to the solvent controls); (A) Strong induction area (response greater
than 50% with respect to the solvent controls). See Figure 4 for calculation of
the inductive activity area.

concentrations tested (Figures 7A and 7B). The mean viability index explained
63% (r2=0.633, p<0.006) of the variation in strong reductive potency after
the one outlier, OHCB4, was omitted from the regression. The mean viability
index explained 46% (r2 =0.456, p<0.046) of the variation in total reductive
potency after the two outliers, OHCB4 and OHCBIB, were omitted from the
regression. The relationship between reductive potency and the viability index
was observed even after the cytotoxic OHCB concentrations were omitted from

the calculation of Iog(E050) and Iog(E020), suggesting that there were

41

sublethal effects caused by the 0HCBs that were not detected by the cell
viability assay. These sublethal effects may have also been responsible for the
observed reductions in luciferase expression, though the nature of these
sublethal effects remains unknown. It was observed, however, that the
administration of E2 reversed the reductive effects of the 0HCBs, leading to the
following possibility: E2 administration reversed the sublethal effects of the
0HCBs by a non-hER-mediated mechanism (since reductive potency was not
related to hER binding affinity). Therefore, although it may be appropriate to
describe the activity of 0HCBs 1-3,5-9,12 as "anti-estrogenic", it should be
noted that the "anti-estrogenicity" was not related to hER binding affinity. It
appears to be unlikely that the 0HCBs exhibited "anti-estrogenicity" as a result
of aryl hydrocarbon receptor agonism because all 0HCBs tested were inactive
in the H4llE EROD induction assay (data not shown).

There were two 0HCBs that caused reductions in luciferase expression
that were not explained by reductions in the cell viability index. OHCB4
exhibited strong (> 50% reduction) reductive potency at a level much greater
than predicted by its slight reduction in cell viability (Figure 7A). And, both
OHCB4 and OHCBI3 exhibited total (>20% reduction) reductive potency at
levels greater than predicted from the cell viability index. These observations
suggest that OHCB4 and 0HCB13 exerted a reductive effect on luciferase
expression by a mechanism other than sublethal cytotoxic inhibition of
luciferase expression. OHCB4 bound to hER with the same relatively strong

affinity as OHCBI1, a partial estrogen. Also, OHCB4 and OHCBII are

42

 

 

 

 

 

 

 

 

10
_ .13
A 9 A .4 Y-0.448X+3.731
"g 8 . r'-0.028, p<0.62 7
C
V 0
8 7 -
c .3 12
d 8 - .
g e
< s
5 5 '
"t"; 4 -
3 .05 .6
O 3 I- 1
E
Ts
e- 2 " 2
.3 e
1 I-
o l l l
2 3 4 5 6
hER l050 (Iog(nM))
6
“A B 04 Y-0.238x+1.501
g 5 ’ r’-0.026,p<0.64
5
O .7
.2. 4 -
8
'- 8
<
‘6
3
'8 2 05 08
a: 1 .13 12
as 9 e
S 1 - 2
A O
o L l l
2 3 4 5 6

hER l050 (Iog(nM))

Figure 6. Relationship between OHCB reductive (E2 antagonist) activity area
and human estrogen receptor (hER) binding affinity represented by IC50, the
OHCB Inhibitory Concentration causing a 50% decrease in specific binding of
3H-E2 to hER in vitro: (Number label) Indicates OHCB identification number from
Table 1; (A) Total reductive activity area (response greater than 20% with
respect to the solvent controls); (B) Strong reductive activity area (response
greater than 50% with respect to solvent controls). See Figure 4 for
calculation of reductive activity area. p=probabi|ity that the slope of the
regression line is equal to zero, a=0.05.

43

 

10 *
13
A 4 '
' All points (_)
‘ ~ 4 7 Y--2.182X+7.648
- ’ ~ r‘-0.026, p<0.64

 

Total Reduction Area (Iog(nM))”)
N O) # 0| 0 ‘1 Q 0

 

 

 

Omitting 011084813 (- .. ) s ‘
Y--7.309X+11.835 .2 ‘ 4 ‘
1 I- r‘-O.456. p<0.046 ‘ s
O l L l
0.50 0.75 1.00 1.25 1.50

Viability Index (Live RFU/Dead RFU)

 

   
  
 

6
All points (—)
B .4 Y--3.985x+6.41e
5 r’-0.273, p<0.089

Omitting OHCB4 (- -)
Y--4.37SX+6.508

  

>50% Reduction Area (Iog(nM)!)
w

 

 

 

1 r'-O.633. p<0.006 0 2
~ \
l l I ~
0
0.50 0.75 1.00 1.25 1.50

Viability Index (Live RFUIDead RFU)

Figure 7. Relationship between OHCB reductive (E2 antagonist) activity area
and cell viability index for all 0HCBs except OHCB10 and OHCB11 that
exhibited partial estrogen agonist activity: (—) All points included in regression;
(---) Regression omitting outliers as noted; (A) Total reductive activity area
(response greater than 20% with respect to solvent controls); (8) Strong
reductive activity area (response greater than 50% with respect to solvent
controls). p =probability that the slope of the regression line is equal to zero,
a=0.05.

44
structurally very similar. Taken together, these two observations suggest that
OHCB4 may be acting to inhibit luciferase expression by an hER—mediated
mechanism. OHCBI3 may also be acting via an hER-mediated mechanism,
though its reductive activity was enhanced by E2 administration-~a response
that cannot be explained by a simple competitive inhibition model of E2 binding
and activiation of hER. The mechanism of action of OHCBI3 and OHCB4
remain to be fully elucidated. However, it may be appropriate to describe their

activity as "anti-estrogenic" by an hER-mediated mechanism.

Imp/ice tions for human risk assessment

The potential for 0HCBs to exhibit estrogenic or anti-estrogenic activity
in humans can be estimated by comparison of the thresholds for effect with the
concentrations of 0HCBs observed in human tissues. Iecause E2 affects the
activity of 0HCBs, the E2 concentration in the tissue is also a necessary factor
to consider. In children of 1 mo to 4 yr of age, plasma E2 concentrations
normally range from 0.04 to 0.1 1 nM. In adult human females, the E2
concentrations vary between 0.18 to 1.5 nM and in males from 0.04 to 0.22
nM (Becker, 1990). Only two of the 0HCBs tested in this experiment have
been positively identified in human plasma--OHCBZ (4-OH-2,3,3',4',5-0l5-BP)
and OHCB9 (4-OH-2’,3,3',4',5—CI5-BP). And, OHCB2 is the only OHCB
presently quantitated (Bergman, et al. 1994) in human plasma--5.3 :t 2.9 nmol/L
plasma (meaniSD). Comparing the observed concentration of OHCBZ, 5.3

nM, with the threshold effect concentrations-E020=7900 nM at 0.0009 nM

45
E2, E020 = 6300 nM at 0.1 nM E2, and E020> 32,000 nM at 1 nM E2-- reveals
that the observed concentration of OHCB2 is 3 orders of magnitude lower than
the threshold for effects concentration at 0.0009 nM E2 and 0.1 nM E2. Thus,
OHCBZ can reasonably be expected to not elicit a detectible anti-estrogenic (or
estrogenic) effect in humans, provided that the responses in viva can be
adequately predicted by the responses in vitro. The concentration of OHCBS
in human plasma has not been quantitated. However, a range can be estimated
as 10 to 50% of the total PCB concentration (Bergman, et al. 1994). Using the
concentration of total PCBs reported by Bergman, et al. (1994), 3.6 :1:1.6pg/g
plasma lipid and assuming a 0.5% lipid fraction in plasma, the concentration of
OHCB9 is estimated to range from 5.3 to 26.5 nM. The lowest threshold for
anti-estrogenic effects for OHCBQ was observed at 0.0009 nM E2--E020 = 0.2
nM. Thus, the predicted range of OHCB9 concentrations exceeds the
threshold for effects concentration at 0.0009 nM E2. However, at E2
concentrations observed in human adults and children, the anti-estrogenic
activity of OHCB9 was completely reversed, suggesting that OHCB9 may be
expected to be inactive in adults and children (Figure 8). Using this same range
estimate for OHCB plasma concentration, it may be concluded that all 0HCBs
except OHCBI 3 would not exhibit detectible anti-estrogenic activity within the
range of E2 concentrations observed in children and adults. OHCBI 3 did
exhibit weak anti-estrogenic activity at 1 nM 52. However, OHCBI 3 has not

been identified in human plasma.

46

Conclusions

As a tool for monitoring the estrogenic/anti-estrogenic activity of
environmental contaminants, the luciferase expression assay described in this
study holds promise in that the assay: 1) detects estrogenic compounds by a
specific receptor-mediated pathway; 2) detects anti-estrogenic compounds by
antagonism of co-administered E2; 3) accounts for the effects of cytotoxicity
that may confound anti-estrogenic responses, and; 4) utilizes a high-throughput
96 well plate format coupled with high-sensitivity plate-scanning
instrumentation. Other assays of estrogenicity have been reported (Soto, et al.
1992, Bergeron, et al. 1994) that were not designed to incorporate all of these
attributes. Of particular note, any in vitro estrogenicity assay that only utilizes
medium stripped of available steroid hormones, while maximizing the sensitivity
for detection of weak estrogens, does not adequately assess the effect of
physiologically-relevant concentrations of EZ on the activity of those weak
estrogens. For example, OHCBIO and OHCBII were estrogenic when
DCCFBS medium (0.0009 nM E2) was used, but were weakly anti-estrogenic
at higher E2 concentrations. Thus, the label "estrogen" or "anti-estrogen"
should not be applied without testing at multiple levels of E2. The methods of
the present assay were based upon the assay developed by Fans, at al. (1990)
for detection of estrogenic and anti-estrogenic compounds.

The risk of "hormone disruption" (Schmidt, 1994) presented by
environmentally-relevant 0HCBs (i.e. those 0HCBs detected in human and

wildlife samples) appears to be negligible, particularly when the important

47

 

  
 
    
   

    
 

 

Expected Lower
E2 Concentration

4%---- NA
4 " ’ .,
, ‘ r.4;-Ma.rgi'ri-j-o‘i.91:511. '

A 3 ' i , Area?
5 2 - I
E l l:.:» i _ i
3 1 : ’ , In -
V I - 1 I
3 Predicted Upper I Poss1ble In Vivo I
0 0 _ OHCBQ Concentration IN Concentrations l
I I l
O I l

1

I

l

Adult E2 Range

 

 

I
I I
2 _ I 1 mo-4 yr Child E2 Range —
' I
I
I

I I I l l

 

-3 -2 -1 O
17B-Estr8diol (Log(nM))

Figure 8. Comparison of the area for in vitro reductive (E2 antagonism) effects
and the area for possible in viva concentrations defined by the lower limit of
average E2 plasma concentrations (nM) in humans (Becker, 1990) and the
upper limit of the predicted OHCB concentration (nM) in an adult human plasma
sample (OHCBQ for example, Bergman, et al. 1994). Determination of the areas
for reductive in vitro effects depicted in Figure 4. The predicted OHCB
concentration is the estimated upper limit of the range of concentrations of
0HCBs in human plasma (10-50% of the total plasma PCB concentration,
Bergman, et al. 1994). Overlap of the in vitro reductive effects area with the
in viva possible concentrations area indicates a condition in which the OHCB
could potentially inhibit E2 activity in viva. On the other hand, the extent of
non-overlap between the two areas indicates the "margin-of-safety" area (SEE)
between the threshold for in vitro effects (E020) line and the possible in viva
concentrations area.

effects of physiologically-relevant E2 concentrations are included in the
analysis. Hydroxylated PCB metabolites that are selectively retained in plasma
do not pose a risk of estrogenic activity because the chemical structural
characteristics that favor plasma retention (chlorination vicinal to a para-

substituted hydroxyl and a total of 5-7 Cl's per molecule) do not favor

48
estrogenic activity (para-substituted hydroxyl without vicinal chlorination).
Also, the OHCBs that proved to be estrogenic in this experiment are rapidly
conjugated and >excreted'in viva precisely because they contain a para-
substituted hydroxyl without vicinal chlorination. Of the OHCBs that exhibited
"anti-estrogenic" activity, only two could be described as probably acting via
an estrogen receptor mediated pathway, and these two have not been identified
in human plasma. The remaining OHCBs exhibited "anti-estrogenicity" that was
related to their effect on cell viability, and therefore, cannot be described as
exhibiting "hormone disruption" solely by an estrogen receptor mediated
mechanism. Furthermore, the "anti-estrogenic" activity of the remaining OHCBs
was eliminated by E2 concentrations normally found in both adults with higher
E2 concentrations and children with relatively lower E2 concentrations. When
the low potential for estrogenic or anti-estrogenic activity present by OHCBs is
compared with the relatively strong potential for interfence with thyroxin and
vitamin A transport (Brouwer, et al. 1990, Bergman, et al. 1994), further
research into the "hormone disruption" caused by OHCBs should probably be

focused on the latter.

CHAPTER 2
SPECIFIC BINDING OF HYDROXYLATED PCBs AND OTHER COMPOUNDS

TO THE ESTROGEN RECEPTOR: STRUCTURE-ACTIVITY RELATIONSHIPS

Introduction

Estrogen Recap tar Agonists and Antagonists

There are four broad classifications of estrogens and anti-estrogens: 1)
direct-acting agonists; 2) direct-acting antagonists; 3) indirect-acting agonists,
and; 4) indirect-acting antagonists. Direct-acting estrogen agonists
competitively bind to the estrogen receptor, displace 17B-estradiol, activate
estrogen receptor dimerization, enhance estrogen-inducible gene expression in
target tissues or cell lines (Kumar and Chambon, 1988), and cause
characteristic physiological changes in whole animals (Knobil, 1980; Figure 2).
Examples of direct-acting estrogens include diethylstilbestrol, coumestrol, and
a,p’-DDT. In contrast, direct-acting estrogen antagonists also competitively
bind to the estrogen receptor, displace 17B-estradiol, but block receptor
function causing down-regulation of estrogen-inducible gene expression (Bondy
and Zacharewski, 1993) and inhibition of characteristic physiological changes
in whole animals (Korach, et al. 1985). Examples of direct-acting estrogen
antagonists include tamoxifen and ICI 164,384. It is important to note that a
necessary condition of direct action is the competitive binding of the agonist or
antagonist to the estrogen receptor. Also, metabolic activation (the enzymatic

49

50
conversion of an inactive "parent" compound to an active "metabolite") may
produce direct-acting metabolites (that bind with greater affinity to the estrogen
receptor) from parent compounds that do not bind with great affinity to
estrogen receptor. For example, methoxychlor does not appreciably bind to
the estrogen receptor, Yet, the dihydroxy metabolite of methoxychlor binds
with great affinity to the estrogen receptor and is a potent estrogen (Bulger,
19783). Metabolic activation may be responsible for the estrogenic activity of
a diphenylsiloxane (Levier and Jankowiak, 1972). Substances may also exert
agonist or antagonist activity through mechanisms that do not include
competitive binding to the estrogen receptor. Indirect-acting estrogen agonists
do not bind to the estrogen receptor but may mimic estrogen action by blocking
androgen action (for example, the anti-androgen p,p’-DDE, (Kelce, et al. 1995))
or inhibiting estrogen metabolizing enzymes (for example, chronic ethanol-
induced liver damage promotes gynecomastia in alcoholic males (Becker, 1990).
Finally, indirect-acting estrogen antagonists do not bind appreciably to the
estrogen receptor but may induce estrogen metabolizing enzymes (Standeven,
et al. 1994; Spink, et al. 1992; Johnson, et al. 1993; Singh, 1989; Dieringer,
et al. 1979) or may block normal estrogen receptor function at the DNA. For
example, activated aryl hydrocarbon receptor interferes with ER binding to DNA
and down-regulates nuclear ER content in cells (Zacharewski, et al. 1991;

Krishnan and Safe, 1993).

51

Uses and Limitations of ER Binding Assay

Ligand binding is the first step in the chain of events leading to
expression of cellular and physiological responses controlled by steroid hormone
receptors (Evans, 1988; Kumar and Chambon, 1988). The ability of the
estrogen receptor to specifically bind a large variety of molecules is well known
and has been extensively studied, including: chlorinated organic compounds
(Welch, et al. 1969; Nelson, 1974; Kupfer, 1975; Kupfer and Bulger, 1976;
Bulger, et al. 1978a,b; Nelson, et al. 1978; Hammond, et al. 1979; Robison
and Stancel, 1982; Korach, et al. 1988; Thomas and Smith, 1993; Chae, et
al. 1991, Waller, et al. 1995), polycyclic aromatic hydrocarbons (Mankowitz
and Rydstrom, 1982; Chang and Liao, 1987; Pasqualini, et al. 1990; Hwang,
et al. 1992), steroid hormone metabolites and derivatives (van Aswegen, et al.
1989; Schutze, et al. 1993; Teutsch, et al. 1988, vanderKur, et al. 1993), non-
steroidal pharmaceuticals (Korach, et al. 1985; Miquel and Gilbert, 1988;
Korach, et al. 1987; McCague, et al. 1988; Anstead, et al. 1990), natural
products of plant, fungal and bacterial origin (Katzenellenbogen, et al. 1979;
Whitten, et al. 1992), alkyl phenols (Mueller and Kim, 1978) and other organic
compounds (Krishnan, et al. 1993). Although ER binding is essential for direct
estrogen agonism or antagonism, receptor binding affinity alone is not sufficient
to predict whether a substance will elicit estrogenic or anti-estrogenic activity.
Rather, specific molecular structural characteristics of the substance determine
the agonistic or antagonistic activity. Although broad classes of compounds

have been empirically identified as "anti-estrogenic" or "estrogenic", the mode

52
of activity (agonist or antagonist) and potency cannot presently be predicted
based solely upon molecular structure. Receptor binding affinity, however, is
directly related to potency when comparing compounds with similar modes of
activity (Chae, et al. 1991; VanderKuur, et al. 1993a). Thus, measurement of
receptor binding affinity is useful for: 1) assessing the potential for direct
action, either as an agonist or antagonist, and; 2) predicting potency when the
mode of action is known. A limitation of assessing the potential for direct
action by estrogen receptor binging affinity is that the substance may require
metabolic activation for estrogen agonism or antagonism. Metabolites can be
qualitatively assayed by performing an enzymatic conversion of the parent
compound prior to the receptor binding assay. The quantitative measurement
of receptor binding affinity is limited in this case because metabolites may not

be readily identified or quantitated.

Objectives

The objectives of this research project were to: 1) survey a wide range
of pure compounds representing primarily chlorinated aromatic environmental
contaminants, and; 2) develop structure-binding relationships for a single class
of chlorinated aromatic compounds, the hydroxylated polychlorinated biphenyls

(OHCBfl.

53

Materials and Methods

Chemicals

Most hydroxylated PCBs (0HCBs) were obtained from Ultra Scientific
(North Kingston, RI). However, some OHCBs were obtained by collaboration
with Prof. A. Bergman (University of Stockholm, Sweden). Chlorinated by-
products of pulp and paper production were obtained by collaboration with Dr.
J. Koistinen (University of Jyvaskyla, Finland). Other compounds were
obtained from Sigma Chemical (St. Louis, MO) and Aldrich Fine Chemicals
(Milwaukee, WI). Most compounds were dissolved in ethanol and stored at -20
°C. However, some highly insoluble compounds (generally those that
contained fused ring systems or were co-planar biphenyls) were dissolved in 1-
propanol and stored at -20 °C. 2,4,6,7-[3H]-17B-estradiol was obtained from
Dupont/NEN at a specific activity of 90-110 Ci/mmol. All buffers and salts

were of reagent grade and obtained from Sigma Chemical.

Calf Uterine Es tragen Recep tar Competitive Binding Assay

Estrogen receptor was obtained from calf uterus because this tissue
contains abundant receptor with low concentrations of bound estrogen. One
calf uterus (approximately 500 g fresh weight) was sufficient to supply receptor
for the entire research project (and another unrelated research project

conducted by another student) with a majority of the tissue left unused. The

54
isolation of calf uterine estrogen receptor and competitive binding assays
followed a standard operating procedure (Appendix 1) that was based upon
methods adapted from Murdoch, et al. (1990) and Clark and Peck (1979).
Competitive binding curves were modelled after a normal distribution (Equation

6, Bruce and Versteeg, 1992).

Molecular Made/ling

Molecular descriptors were calculated using Hyperchem® desktop
modelling software (Hypercube Software, Waterloo, Ontario) for the OHCBs
only. The global minimum energy conformation was identified by an initial
search of the central torsion bond angle at restrained 30 degree increments
using the AMI semi-empirical method with an energy gradient criterion of 0.01
rms (roor mean square). Because the conformational energy is symmetrical
about 180 degrees, the initial minimum energy torsion angle was selected from
the 0-180 degree range for final unrestrained optimization to a gradient cutoff
of 0.001 rms. The final minimum energy torsion angle was estimated using the
fully optimized, relaxed molecule. The sine tranformed torsion angle (to achieve
Iinearization) was used in subsequent statistical analysis. Fourteen molecular
descriptors were calculated and included in a multiple regression analysis (Table
4; SAS Institute, 1988). The minimum energy structure of 17B-estradiol was
also modelled but its molecular descriptors were not included in subsequent
statistical analysis. A minimum set of molecular descriptors was selected that

optimally predicted the Iog(lC50) for specific binding to the calf uterine estrogen

55

Table 4. Molecular descriptors of hydroxylated PCBs calculated using

 

 

 

 

 

 

 

Hyperchem®
Descriptor Definition Units
E 2
Electronic Variables
EGAP Energy gap: Energy difference between the highest occupied electron volts
molecular orbital (HOMO) and the lowest unoccupied molecular
orbital (LUMO), a measure of the reactivity of the molecule.
OCHARGE Oxygen partial charge: Partial charge on the oxygen of the electronic charge
hydroxyl group, affects strength of hydrogen bonding and pKa. e'
DIPOLE Molecular dipole moment: Dipole moment calculated on entire Debyes
molecule.
ZDIPOLE Dipole moment along 2 axis: Dipole calculated on long axis of Debyes
molecule, affected by position of hydroxyl group.
Spatial Variables
SASA Solvent accessible surface area: Surface area accessible to A2
solvent molecule with mean radius of 1.4 A. Angstrom2
SAVOL Solvent accessible volume: Volume accessible to solvent A3
molecule with mean radius of 1.4 A.
VDWSA Van der Waals surface area: Surface area at the van der Waals A2
radius for each atom in molecule.
VDWVOL Van der Waals volume: Volume at the van der Waals radius for A3
each atom in molecule.
XYZVOL XYZ volume: Volume of the box defined by the dimensions of the A3
molecule in the X,Y, and 2 directions.
STERIC Steric volume: Van der Waals volume of substituents vicinal to A3
the hydroxyl group, measures steric hindrance affecting hydroxyl
group.
BULK Bulk volume: Van der Waals volume of non-hydroxylated ring, A3
measures volume of "bulky" constituent of molecule.
BULKPER 8qu volume percent: Bulk volume expressed as a percentage of 96
total van der Waals volume of molecule.
Sauctural Variables
SINTORS Sine of the torsion angle: Torsion angle of central biphenyl bond, SinIdegreesl
sine transformed for Iinearization, affected by degree of orflta
substituted chlorination.
OHDIST OH distance: Distance of OH group from molecular center of A

mass, measures "extremity" of OH group, related to substitution
position, para > meta > artha.

56
receptor. The optimum set of descriptors was selected by maximizing the
coefficient of determination (r2) for each subset of descriptors, such that the
best one variable model, best two variable model, best three variable model,
etc. was selected. The progress of this selection process was plotted and
thepoint at which no appreciable additional increase (less than 10% of the total)
in predictive power with increasing descriptors was selected as the optimum
model. Two models were evaluated in this way, an unweighted model and a
weighted model. The weighted model incorporated information on the reliability
of the Iog(lC50) measurements. In general, the accuracy of the Iog(lC50)
determination decreased as the absolute value of the Iog(lC50) increased,
because greater Iog(lC50)'s were estimated using increasingly incomplete
competitive binding curves. Therefore, the weight given to Iog(lC50)

observations was proportional to the inverse of the Iog(lCSO) (Equation 7).

Weight = ———1—— (7)

Iog(lC50)

Results and Discussion

A total of 53 chemicals were tested for the ability to bind specifically to
calf uterine ER (Tables 5-7). All but one hydroxylated PCB, (OHCBB) bound
specifically to calf uterine ER with measurable affinity (Table 5). Three DDT
isomers and derivatives were assayed, two of which bound weakly to the ER
(Table 6). Structure affected binding of the DDT compounds such that binding

affinity Iog(lC50) followed the pattern: a,p’-DDT<a,p’-DDE<a,p’-DDD (Not

Table 5. Prediction of Iog(lC50), (Iog(nM)) using multiple linear regression of the

57

optimum five molecular descriptors.

 

Predicted Logucsoi
(SE)
ID Formula Structure Observed Log(lCSO) Weighted TIanelghted
HO CI 61
, , _ - c: Q Q cu 4.25 4.51 5.20
011cm 30H 2.3 .4.4 .5 Cl, 3? H H (039, .042,
Cl
(:1 Cl CI
4.13 4.54 5.25
- ' ' - - HO Cl
once: 4-0H 2.3.3 .4 .5 cu. BP (0.101 10.441 (0.45)
Cl
140 0| 0|
. . . . . - _ c1 or 3.86 5.11 5.69
oncaa 30112.3 ,4,4,5c1.aP (0.34) (0.42) (0.46)
CI
Cl Cl Cl
. . . . Mo on 7.43 4.95 5.34
0HCB4 4—0H-2.3.3.4.5.5-C|.—BP (u, (0.38) (0.38)
on cl
Cl Cl
. , , no or 5.50 5.04 5.29
0HCBs 4"OH'2 0304 05:6 'CIs-BP (n) (0.33, (0.36)
CI CI
Cl Cl
' . "o c. 3.85 4.20 4.22
OHC86 4-OH-3.3 .4 -CI,-BP (0.08) (0.42) (O_‘8)
Cl HO Cl Cl
, , , c: O C: 4.50 5.51 5.45
OHCB7 2-0H-2.3.3,4.4-CI.-BP (0.18) (0.43) (0.48)
CI Cl
oncas 4-0H-3.3,4,5,s-c1,-aP H "A 10.43) (0.471
or Cl
Cl Cl Cl
_ I I I - - Ho CI 8.“ 5.73 6.27
oncas 4011233 .4 .5 CI. 8? (-4 10.451 (0.45)
CI c1
. . . .. Q a ..
ouce1o 4011-2 .4 .5 cure? " (0.14) (0.391 10.501
Cl
Cl CI
. . . . Ho Q Q or 2.15 2.70 2.55
onca11 eon-2 .3 ,4 .5 cu,» H (0.151 10.401 (0.53)
CI
HO Cl Cl
. . . . CI 3.41 2.75 2.85
once12 3-OH-2,3,4,5-CI.-BP 10.21) (0.341 (0.45)
CI
Cl Cl
. . . HO on 4.94 5.29 6.53
once13 4.4 40H),-3.3 .s.s -CI.-8P 10.221 10.441 (0.45)
CI Cl
OH Cl
O Q °' 511 457 455
once14 2-or+2,3.4.5,5-c1.-eP (0.90, (0,44, (0,49)

 

9
9
5'2

 

Table 5. (Cont’d).

 

 

 

Predicted Log(ICSOl
(SE)
to Formula Structure OWEELET (ICSO) Iggoggt! "Egg!
CI OH CI
4.59 4.24 5.20
01-15515 2-0n-2'.3.3'.4',5'-c en
|. ”a (0.50) (0.57) (0.42)
CI CI
c1 Cl
3.53 3.59 5.25
H 1 ' r '5 l. " P
o 55 5 4-0H22 5 5,5 Ho“ (0.05) (0.55) (0.45)
CI
HO
4.42 4.05 5.59
0n5517 son-5-51-55
(0.15) (0.40) (0.45)
CI
Cl OH
5.73 5.75 5.34
0H5515 2- n- - 1-59
0 35 (-1 (0.50) (0.35)
CI
4.20 4.55 5.29
0115519 eon-3-51-55 H C) Q (0.13, (0,29, (0,35)
OH
5.55 4.55 4.22
OHCB” Nil-5'0"” (0.31) (0.25) (0.45)
CI
on cu
. ._ _ 5.12 4.51 5.45
0115521 2.0112 .5 51,59 ”25, (0,35, (0.45)
no Cl 5)
. ._ - 4.25 4.20 5.02
OHCBZZ 3-OH-2 .5 CI, 3? (00‘) (O-‘I, (047)
CI
Cl OH
4.94 5.47 5.27
01-1523 zen-3.5-5559 (0.27) (0.25) (0.45)
CI
Cl
5.05 4.77 2.55
0115524 eon-3,551,512 ”a” (0.24) (0.44) (0.50)
CI
HO
. 4.44 4.57 2.55
0115525 son-4,4 51,59 c, 01 (0.13) (0.37) (0.53)
H0
5.11 5.17 2.55
a- l_ - Cl
01-15525 3.3 (OHI,-4,4 51,55 0'" (0.52) (0.45) (0.45)
OH
Cl OH Cl
. .- _ 4.59 5.35 5.53
0n5527 20112 .3.5 51.55 (0.14) (0.37) (0.45)
CI
on 5)
. . _ 4.25 4.57 4.55
0115525 2-0n-2 .5.5 51, 5? (0,05) (0.33) (0.49)
CI Cl
r’ 0.515 0.549
555 3.50 20.51

 

5: Asymptotic standard error, - indicates ASE inestimable by nonlinear regression.

59

Table 6. Estimates of calf uterine estrogen receptor binding Iog(lC50).
(Iog(nM))for miscellaneous compounds.

 

ID Formula Structure LogAuC50) Reason '0'
on
5)),
_ 0.837 . .
(:2 17531155101 2::635 (0.059) "mm "9"“!
no
Cl

o,p’-dichl0ro-diphenyl- H 4.93 ,

D1 1,1,1-trichloroethane Q CI (0.38l Reported estrogenic (U
CCI

C
02 a,p -d)5hloro—d1phenyl- Q Q 5‘3)" Matabolitg of 01

1 .1 -dichloroethene

51 I
Cl Cl
Cl
'd' hi 5' h l H NA.
03 a,p _ 15‘ are- no any- CI Metabolite of D1
1 , ‘l -d15hloroethane C no
Cl

CI
. . 6.77
P0830 2,4,6-d15hloro—biphenyl Parent PCB of OHCBIO

C) ("I

OH
A2 4-chloro-2-lsODrODYI-5- c: 6‘99 Structure
I

methyl-phenol l—l

-— — ====Il==
a: Symptotlc standard error, - Indicates ASE Tnesfimable 5y nonfinear regression.

5: Not Active.
1: Welch. et al. 1969.

60

Table 7. Compounds of environmental concern that did not bind to calf uterine
estrogen receptor.

 

Cl
Similar to OHCBtO

fl.

PC875 2,4,4',6-tetrachl0ro-biphenyl

9

9
9

PCBSO 3,3',5.5’-tetrachloro—biphenyl Similar to OHCB13

g

Q
9

Q
9

PC2881 2.3.4.5-tetrachl0ro-biphenyl Parent PCB of OHCB11

2
9

9

9
9
9

2.2'.3.4,5.5'-h5xachl0ro-

PC8141 Similar to OHCB14

%
C9

biphenyl
0| 0|
J1 9 hloro-reten Identified in bleached kraft mill
C 5 effluent, BKME (1)
Cl

1-(2,5~dim5thylphenyI)-2-i3¢
J2 chloro—4-lsopropylphenyll-
ethane

Identified in BKME (1)

a
.Q

1~(2 .5-dim5thylphenyll-2-(3-
J3 chloro-4-8ochloro
isopropylphenyll-ethane

Identified in BKME (1)

Q
@215

A1 5-isapropyl-2-methyI-phenol Structure
HO OH
A3 2.7-naphthalenedi0l Structure
OH
O ©©
A4 2 ,4-dichloro-1 -naphthol Structure
Cl
..~
A5 2ochloro-naphthalene Structure

Table 7. (Cont'd.)

H MIL—[.12 IW

A6 2.5-dichloro—m-terphenyl Structure
Cl
C)
A7 2,5-dichloro-p-terphenyl Q Q Q Structure
Cl

A8 2.5-dichlor0—o-terpl15nyl c, Q Structure

OH
Cl Cl
A9 2.6-dichloro-phenol ©r Identified In BKME (2)
OH
Cl
A10 2.4-dichl0ro-phenol Identified in BKME (2)
Cl
OH
Cl
A11 2,4,5-trichl0ro—phenol c. ldentlfled in BKME (2)
Cl
Cl
Cl
A12 "2'3""°"‘°"'°'° ©© Identified in BKME (2)
naphthalene Cl
Cl
OH
A13 trislkchlorophenyII-methanol © Identified In marine mammals and birds (3)

 

62
active). Also exhibiting measurable affinity were 17B—estradiol, (E2) as a
positive control and reference for the other compounds, 2,4,6-trichlorobiphenyl,
(PC330) and 4-chloro-2-is0propyl-5-methyl-phen0l (A2). Compounds that did
not bind to ER (Table 7) comprised individual PCB congeners, chlorinated
polyaromatic hydrocarbons, chlorinated bibenzyls, chlorinated phenols and other
compounds found in the aquatic environment.

Quantitative structure-binding relationships were developed using
multiple linear regression of either weighted or unweighted observations. The
fourteen molecular descriptors for the 0HCBs (and E2 for comparison) were
calculated (Table 8) and subjected to multiple linear regression using the
maximum r2 selection method for inclusion of regressors. A five regressor
model was selected that optimally explained the variation in the dependent
variable, Iog(lC50), (Figure 9). The maximum variation in Iog(lC50) explained
by the full 14 variable weighted model was 72.66% (r2=0.7266), and by the
full 14 variable unweighted model was 70.63% (r2=0.7063). The five
variables that were selected in both the weighted and unweighted models were:
EGAP, XYZVOL, STERIC, OCHARGE, and ZDIPOLE (Table 9). Each variable
described different aspects of the OHCBs molecular structure. EGAP is a
"global" electronic parameter that describes the overall reactivity of the
molecule. Inspection of the coefficient for this variable indicated that EGAP
varied directly with Iog(lC50), larger values of EGAP were related with larger
values of Iog(lC50), suggesting that more reactive molecules bind less well to
ER. However, the coefficient for EGAP was only marginally significantly

different from zero with p<0.12. All other variable coefficients were

63

Table 8. Results of molecular modelling

10 EGAP oanace 01501.1: ZDIPOLE SASA SAVOL VDWSA
(5V) (5') (Debye) (Debyel (A’) (A3) (A2)
011551 5.5521 0.2255 1.315 0.525 455.57 751.72 272.03
011552 5.5553 0.2245 1.435 0.915 470.59 754.14 273.55
011553 5.5342 0.2323 2.130 1.499 459.90 752.25 274.55
011554 5.4455 0.2255 1.050 0.502 492.50 501.09 255.45
011555 5.4714 0.2255 1.131 0.095 477.75 755.53 275.50
011555 5.3717 0.2409 1.444 0.595 732.70 554.90 242.97
011557 5.5571 0.2325 2.241 -1.159 452.72 757.33 272.51
011555 5.3020 0.2255 1.359 0.1 15 471.55 755.05 274.35
011559 5.5409 0.2257 1.473 1.404 472.95 752.35 273.55
0115510 5.7755 0.2493 1.155 0.354 425.99 559.13 245.35
011551 1 5.4335 0.2475 2.024 0.105 445.34 720.40 255.55
0115512 5.5599 0.2495 2.553 0.031 445.59 720.57 255.40
0115513 5.2920 0.2257 1.970 0.455 452.35 745.55 257.42
0115514 5.4553 0.2435 3.145 0.002 451.99 759.43 272.22
0115515 5.3532 0.2331 1.551 -1052 453.59 755.41 271.90
0115515 9.0237 0.2393 1.755 0.577 421.57 555.97 244.15
0115517 5.5422 0.2479 1.149 0.157 354.54 503.43 211.25
0115515 5.7977 0.2409 2.573 0.517 355.57 501.50 209.55
0115519 5.5731 0.2433 1.023 0.054 359.95 502.30 210.05
0115520 5.5535 0.2443 1 .532 0.342 351.45 502.40 21 1.52
0115521 5.7715 0.2471 1.709 0.125 404.25 542.05 227.95
0115522 5.7537 0.2505 1.474 0.977 409.52 545.55 225.50
0115523 5.5457 0.2359 0.515 0.035 399.25 542.51 225.54
0115524 5.5292 0.2257 1.975 -1.021 412.95 545.09 225.32
0115525 5.4594 0.2441 1 .052 0.302 405.55 545.75 225.43
0115525 5.4175 0.2440 0.549 0.450 415.51 554.91 235.95
0115527 5.7795 0.2395 1.095 0.304 425.09 551.95 243.55
0115525 5.5309 0.2455 1.511 0.215 425.24 555.59 243.51
52 9.2352 0.2524 1.214 0.073 470.30 505.50 295.25
_ i

of hydroxylated PCBs.

 

 

Table 8. (Cont'd.)

64

 

10 VDWVOL XYZVOL 9151115 BULK BULKPER 0110191 91111059
(A’) (A3) (A3) (A3) 91 (A)
E:
011551 237.95 214.23 30.55 113.52 47.7 4.10 0.5599
011552 237.52 224.45 30.94 1 13.44 47.7 4.52 0.5730
011553 237.95 177.34 17.01 127.75 53.7 4.05 0.5772
011554 251.75 151.35 31.15 141.42 55.2 4.50 0.5797
011555 235.24 155.55 31.30 127.73 53.5 4.50 0.5595
011555 209.57 101.90 17.02 113.57 54.1 4.59 0.5550
011557 237.55 157.93 133.01 127.55 53.7 2.47 0.9171
011555 235.15 102.54 31.15 127.57 53.5 4.55 0.5525
011559 237.51 154.25 31.21 127.41 53.5 4.50 0.5779
0115510 210.05 245.51 2.99 142.41 57.5 4.35 1.0000
on5511 223.73 205.95 3.01 141.73 53.3 3.44 0.5599
0115512 223.52 204.25 3.01 141.71 53.3 3.05 0.5591
0115513 230.57 103.45 31.19 120.12 52.1 4.50 0.5553
0115514 237.54 209.20 133.13 141.51 59.5 2.95 0.5475
0115515 237.74 235.50 142.27 141.53 59.5 3.23 0.9205
0115515 209.70 237.20 17.00 99.45 47.4 5.95 0.9997
0n5517 151.55 152.95 3.02 55.22 45.9 4.35 0.5525
0115515 151.31 145.50 155.32 55.27 47.0 3.44 0.7970
0115519 151.29 121.45 17.03 55.19 47.0 4.50 0.5530
0115520 151.50 135.13 75.59 55.19 45.9 3.05 0.5542
0115521 195.53 199.45 105.45 113.50 55.1 2.94 0.5270
0115522 195.55 154.57 3.00 113.57 55.0 3.15 0.5459
0115523 195.31 145.11 90.74 55.12 43.5 3.05 0.5717
0115524 195.57 140.07 31.15 55.24 43.5 4.52 0.5570
0115525 195.93 95.49 17.02 99.45 50.5 5.14 0.5459
0115525 201.95 75.55 17.01 105.75 52.4 3.34 0.5435
0115527 209.79 154.75 119.39 113.73 54.2 2.55 0.5343
0n5525 209.51 174.94 105.05 113.51 54.2 2.43 0.9021
52 257.57 245.54 3.02 213.57 52.9 7.59 .-

 

65

 

1.00
0.90 -
0.30 1 Weighted Regression
0.70 i
0.50 1
0.50 1

1

1

    

0.40
0.30
0.20 t
0.10 P

0.00 ..............
012 3 4 5 6 7 8 9101112131415

Unwelghted Regression

Coefficient of Determination (r’)

 

 

 

Number of Regressors in Model

Figure 9. Progress of multiple regression relating calf uterine binding Iog(lC50)
to molecular descriptors of 0HCBs in which the best combination of descriptors
that maximized r2 was selected for each number of descriptors.

significantly different from zero, a=0.05. ZDIPOLE described the dipole
moment of the molecule along the Z axis, which was defined as the axis aligned
with the central biphenyl bond. ZDIPOLE was affected by the substitution
position of Cl and OH groups and varied directly with log(|050) such that more
positive values of ZDIPOLE were associated with greater values of Iog(lC50).
OCHARGE described the partial charge on the oxygen of the hydroxyl group
and was strongly affected by the substitution of vicinal chlorines . Vicinal
chlorines caused electron withdrawal through the aromatic ring system from the
hydroxyl oxygen resulting in a more positive partial charge. Electron withdrawal
from the oxygen would result in greater propensity to ionize in aqueous solution
(because the proton is less tightly bound) resulting in reduced pKll (Waller, et
al. 1995). Also, reduced partial charge on the oxygen would also reduce the

hydrogen bond acceptor strength (hydrogen bond basicity) of the oxygen.

66
Though not known, the hydrogen bonding attributable to the phenolic hydroxyl
group of estrogens may contribute significantly to the ER binding energy. The

remaining two variables describe spatial characteristics of the OHCBs. STERIC

Table 9. Optimum five variable weighted and unweighted models for the
prediction of Iog(lC50), (Iog(nM)) for binding to calf uterine estrogen receptor.

 

Regressor Weighted Model Unweighted Model
Coefficient SE p Coefficient SE p

EGAP 1.8156 1.1391 0.126 1.3841 1.3788 0.327
XYZVOL -0.0148 0.0041 0.002 -0.0144 0.0048 0.007
STERIC 0.0125 0.0037 0.003 0.0121 0.0041 0.008
OCHARGE 60.030 21.443 0.011 68.155 24.813 0.012
ZDIPOLE 0.5075 0.2700 0.074 0.6832 0.2954 0.031
INTERCEPT 5.2121 9.2562 0.579 10.9236 10.5179 0.311

 

described the volume of the substituents bonded vicinally to the hydroxyl group
and varied directly with Iog(lC50). Thus, steric hindrance 0f the hydroxyl group
appeared to inhibit binding to the ER. XYZVOL described the volume of the box
that just enclosed the OHCB. XYZVOL was a volume parameter that was
affected by Cl and OH substitution pattern, torsion angle, and total number of
Cls. Thus, XYZVOL incorporated information on several aspects of the
structure of the OHCB and was most significantly related to Iog(lC50) in the
multiple regressions (weighted p<0.002; unweighted p<0.007). The

performance of the multiple regression models was of limited success as

 

 

 

 

 

 

 

 

9 9
Weighted Multiple Regression Unweighted Multlple Regression
8 1 fig,g1g_ssg.3.go 8 " r'-0.549, SSE-20.5
§ 7 t T 5'; 7 r T
s I 5 I
5 0 l 5 6 ’ T T i
g 'i 4 T—-—1 T i Z 1% A "7‘ I
1'. . It) 5 #T i
‘5? 5 + . '* . 2:- 11 133‘
O .
3: ‘ ii i 9 ‘ {i
‘3 I ‘o’ T 1
I. 3 T h 3 p . ..l
n. . lgl l l l
2 2 l
1 1
1 2 3 4 5 0 7 8 9 1 3 4 5 6 7 8 9
Observed ICSO (Iog(nM)) Observed ICSO (Iog(nM))

Figure 10. Comparison of predicted vs. observed log(lC50) for binding of
0HCBs to calf uterine estrogen receptor using both weighted and unweighted
five variable multiple regression models; *=asymptotic standard error of
observed log(lC50) inestimable by nonlinear regression and therfore not shown;
other error bars not shown because the bar was smaller than the symbol;
vertical error bars are standard error of the individual predicted log(lC50).

indicated by the relatively low coefficients of determination. However, the
model did predict log(lC50) to within 1 log unit for 24/28 0HCBs studied. This
level of resolution is useful for prediction of the general range of binding affinity
and would be useful for identifying those 0HCBs that have the potential for
strong binding. Some of the error in the predictions was also related to the
error associated with the experimental estimation of the log(lC50)'s (Figure 10).
Nearly all, 26/28, of the 0HCBs fell within two experimental estimate standard
errors of the multiple regression prediction. Those 0HCBs with relatively large
log(lC50)’s, (greater than 7 Iog(nM)) were poorly predicted by the models.
However, the experimental estimate of these log(lC50)'s was also the most
uncertain because of extrapolation well beyond the highest concentrations

tested (50 pM). Despite this limitation, the model did predict the log(lC50)’s

68
of the very weakly binding OHCBs to be greater than 5 Iog(nM), beyond the
solubility of these compounds (observed at 4.7 Iog(nM)). The use of the
weighted regression did improve the predictive power of the model by about
7%. The effect of the weighting was to de-emphasize the highly uncertain
observations greater than 7 Iog(nM) while emphasizing the more accurate
observations below 3 Iog(nM). The weighted multiple regression model
provided the best framework for predicting ER binding by OHCBs and may
prove useful in predicting the binding affinity of 0HCBs that have not yet been
assayed, including those recently identified as human and wildlife metabolites

of PCBs (Bergman, 1994).

CHAPTER 3
DEVELOPMENT AND CHARACTERIZATION OF AN ESTROGEN RESPONSIVE

LUCIFERASE REPORTER PLASMID, pGUDLUC2.2

Introduction

A wide variety of synthetic and natural chemical products possess
estrogenic (and anti-estrogenic) activity (McLachlan, 1993; Safe, 1994). The
role of estrogen mimics in human diseases such as breast cancer and testicular
abnormalities remains uncertain and controversial (Colborne and Clement, 1992;
Colborne, et al. 1996; Safe, 1995). The comprehensive screening of synthetic
chemicals for estrogenic activity has been suggested (ESCREEN, Soto, et al.
1992) utilizing a rapid in vitro bioassay based upon the proliferative response
of MCF-7 cells (a human breast adenocarcinoma cell line) to 17l3-estradiol and
other estrogens. A functional in vitro bioassay approach to screening for
estrogenic compounds provides a rapid, high-throughput technique that can
surpass a simple estrogen receptor binding assay in detecting active compounds
by incorporating all of the biochemical processes involved in the cellular
response to estrogens. A weakness of the ESCREEN bioassay is that the
precise genetic elements controlling the proliferative response of the MCF-7
cells has not been characterized, leading to the possibility that other non-
estrogen pathways can cause proliferative responses. The proliferative effects
of insulin and tumor growth factor a on MCF-7 cells are just two examples of
non-estrogen related proliferative responses (Safe, at al. 1992; Liu, et al. 1992).

69

70

To improve the specificity of response to estrogen exposure, estrogen-
inducible reporter genes have been used in a number of cell lines (White, et al.
1994; Routledge and Sumpter, 1996; VanderKuur, et al. 1993b). In this
research project, a plasmid containing a luciferase reporter gene under the
transcriptional control of tandem consensus-sequence estrogen responsive
elements (EREs) was designed and constructed to take advantage of the
specificity of reporter gene constructs and the superior sensitivity of the
luciferase light-emitting reaction. The plasmid, pGudluc2.2, was designed to
resemble a luciferase reporter gene containing-plasmid (pGudlu51.1) under the
transcriptional control of dioxin responsive elements (DREs) was successfully
constructed and utilized by others in the Aquatic Toxicology Laboratory at
Michigan State University (El-Fouly, et al. 1994). The objective of the research
project was to characterize the response of the plasmid, pGudlu52.2, in MCF-7
cells after transient and stable transfections and to evaluate the potential for
use of the stably transfected cell line for use as an estrogen and anti-estrogen

screening bioassay.

Materials and Methods

Plasmids for the construction of pGudlu52.2 and reagents for the assay
of luciferase activity were obtained from Promega Corp. (Madison, WI). The
tandem ERE sequence was synthesized in the Macromolecular Structure Facility

at Michigan State University.

71

Construction of pGudlch. 2

The plasmid pGudlu52.2 was based upon the DRE-containing plasmid
pGudluc1.1 (El-Fouly, et al. 1994). The construction of pGudlu52.2 proceeded
by insertion at the Hindlll site of the mouse mammary tumor virus promoter
(MMTV promoter, 1810 bp) upstream of the coding region of the firefly
luciferase gene found in pGLZBasic (Figure 11). For pGudlu51.1, a 482 bp
fragment of the regulatory region of the mouse cytochrome P4501A gene
containing four DRE sequences was inserted 100 bp upstream of the MMTV
promoter start site (El-Fouly, et al. 1994). For pGudlu52.2 (5858 bp), 5 1090
bp fragment between the distal Bglll site (a second Bglll site lies 29 bp
upstream of the Hindlll site) and the Hindlll site within the MMTV promoter
sequence, (preserving the MMTV promoter start site) was excised and replaced
by a 54 bp Bglll/Hindlll oligonuoleotide containing two consensus EREs (Figure

12).

Transfection of pGudlu52.2 into MCF7 cells

MCF-7 cells were chosen as the benchmark cell line for evaluation of the
effectiveness of pGudlu52.2 in conferring estrogen-inducible luciferase
expression because MCF-7 cells have been extensively characterized and
contain abundant functional ER (Eckert and Katzenellenbogen, 1982). Transient
tranfections were conducted to assess the estrogen responsiveness of the

plasmid. Initially, Polybrene® (Gibco BRL, Gaithersburg, MD) mediated

72

"In
dm Hindlll 1-3 lib
5551 bp 5 3 , 1111111111
M MTV

Promoter
pGL2Basic MMTV «up
A hr
mp' 11...... LUC 4 DRE's
pGudluc 1 .KMMTBRE
Kpnl MMTV
lll
BOMMTV Kpn|52bapom Amp, 7388179 LUC
am» L”° 2 \<ERE s
pGudluc 1.1
ERE
MMTV

Amp, 5555 bp LUC

pGudluc 2.2

Figure 11. Construction history of pGudluc2.2.

CTT CCT GGT CAG CGT GAC CGG AGT CAC ATG TTC CTG GTC AGC GTG ACC GGA GTA
C ATG GAA GGA CCA GTC GCA CTG GCC TCA GTG TAC AAG GAC CAG TCG CAC TGG CCT CAT CTA G

 

Kpnl Bng

ERE ERE

Figure 12. Oligonucleotide containing two EREs used in pGudluc2.2.

73
introduction of pGudlu52.2 was not successful. Therefore, Lipofectin® (Gibco
BRL) reagent was used instead with success according to the manufacturers
protocol. Luciferase expression of transiently—transfected MCF-7 cells was
elevated 48 hours after transfection by treatment with 10 nM 17B-estradiol

(Figure 13) prompting initiation of the stable transfections. Lipofectin® was also

 

Luciferase (pg/mg protein)

 

 

 

0 0.1 1 10
17B-Estradi0l (nM)

Figure 13. Response to 1715-estradiol by MCF-7 cells transiently transfected
with pGudlu52.2; N = 2 at each concentration; bars indicate mean :1 SE; letters
indicate means are not significantly different by Fisher’s LSD, a=0.05; F
statistic produced by ANOVA.

used in the stable transfection to deliver pGudlu52.2 and a selective resistance
marker, pSV2neo (a neomycin resistance gene; Southern and Berg, 1982) at a
molar concentration ratio of 20:1. A large marker to reporter plasmid ratio was
used to insure that resistant clones likely incorporated both plasmids.

Selection of antibiotic resistant clones was achieved by subcloning of

74

transfectants three days after plasmid incorporation at a 1:3 dilution in culture
medium (10% FBS Medium) containing 900 mglml G418 (a neomycin analog
under the trade name Geneticin®, Gibco BRL). Bacterial contamination of 100
mm Petri plates was encountered initially and eliminated by the use of 75 cm2
flasks. Resistant clones were visually identified, subcultured and grown to
confluence in 48 well culture plates. The time from transfection to confluent
cultures of clones was approximately three months. Seven clones (GLUC1-7)
were isolated from approximately 1 million MCF-7 cells transfected initially.

Four of the seven clones lost their resistance to G418 during subculturing,
leaving three clones, GLUC 1, GLUCS and GLUC6, available for characterization

of estrogen inducibility of luciferase expression.

Characterize tion of Stab/y Transfected Es tragen Responsive Luciferase

Reporter Gene

Estrogen responsiveness of the three stably transfected clones was
evaluated in the absence of selective pressure by G418, though clones were
continuously subcultured in G418-containing medium to maintain their
phenotype. It was observed that G418 treatment completely inhibited
luciferase expression, possibly by monopolizing the protein synthesis machinery
of the cells to produce the antibiotic resistance necessary for the survival of the
cells. On Day 1, clones were subcultured into G418-free 10% FBS medium in
48 well tissue culture plates at a concentration of 10,000 cells per well.

Vigorous growth and attachment was promoted by incubation overnight in 10%

75
FBS medium. On Day 2, the medium was replaced by fresh 10% FBS medium
to reduce the concentration of residual G418. On Day 3, the medium was
supplemented with 1 7B—estradiol at concentrations ranging from 0.1 to 100 nM.
A concentration of 10 nM E2 was sufficient to maximally induce MCF7LUC
cells, so the range of E2 was selected to include that concentration. In one
experiment, the time course of induction of luciferase expression was
investigated over a period of 9 days following initiation of exposure to E2. The
optimal exposure duration was selected from this experiment and applied in
further experiments that characterized the response of GLUC1 to E2, DES (a
potent non-steroidal estrogen), tamoxifen (a direct-acting anti-estrogen).
GLUCS and GLUC6 were assayed for their response to E2 only because their
pattern of response to E2 mirrored that of GLUC1. Luciferase expression was

measured as previously described (Chapter 1, Materials and Methods).

Results and Discussion

The plasmid, pGudlu52.2, transiently transfected in MCF-7 cells,
exhibited a 1.3 fold induction of luciferase activity at a concentration of 10 nM
E2 (Figure 13). It was thought that stable transfection would improve the
induction potential by improving the efficiency of gene transcription of
genomically incorporated DNA. The stable transfection produced three
successful clones that stably expressed both luciferase activity and G418
resistance. GLUC1 was more thoroughly characterized than GLUC5 and

GLUC6 after it was realized that GLUC5 and GLUC6 responded similarly to E2,

76
suggesting that a complete repeat of the characterization used for GLUC1 was
unnecessary. The time course of luciferase expression in response to E2
indicated that the maximal response was achieved by at least 72 hours after
treatment with E2 began (Figure 14). Most notably, E2 caused a concentration

dependent reduction in luciferase expression that was not expected. This

Day 3 Day 5 Day 7

 

 

 

Luciferase (pg/mg protein)
Luciferase (pg/mg protein)
Luciferase (pg/mg protein)

 

 

 

 

 

 

0 0.1 1 10 O 0.1 1 10 0 0.1 1 1O
17fl-Eatradiol (nM) 17B-Eatradlol (nM) 17B-Eetradlol (nM)

Figure 14. Time course of luciferase expression response to 17B—estradiol by
MCF-7 cells stably transfected with pGudlu52.2, clone GLUCI; N=1 at each
concentration.

reduction pattern persisted over 7 days of treatment of E2. Also notable was
the relatively low level (5-15 pg luciferase/ml protein) of luciferase expression
in the absence of E2. By comparison, MCF7LUC cells express 4000 pg
luciferase/mg protein in the absence of added E2 and can be induced to express
up to 10,000 pg luciferase/mg protein. A second experiment was conducted
on GLUC1 that examined the response to E2, DES, and tamoxifen after 48

hours exposure. Again, luciferase expression appeared to be reduced by E2

77
and DES, though only E2 produced statistically significant reductions in
luciferase expression (Figure 15). Thus, ER agonists caused concentration
related inhibition of luciferase expression by stably-tranfected pGud|u02.2.
Further evidence that the response was mediated by ER was provided by the
treatment with tamoxifen and dexamethasone. Tamoxifen treatment caused a
slight increase in luciferase expression at concentrations greater than or equal

to 82 nM, suggesting that antagonism of ER improved expression of luciferase

 

 

     

 

 

 

 

 

 

17B-Estradiol Diethylstilbestrol Tamoxifen
2 2 4
A F-13.69 F-1.22 [3.1.90
p<0.0001 p<0.34 p<0_16 A
E E 15 3 AB
3 e a
‘6 ‘9 2 AB
5 o. n. B
5 5 g ( “3
E 1 E g 2
a 2 I
5 E 5
3 1,3 g 1
o o
o 0.01 0.1 1 10 100 o 0.01 0.1 1 10 100 o 0.52 5.2 52 520
17B-Estradlol (nM) Diethylstilbestrol (nM) Tamoxifen (nM)

Figure 15. Luciferase expression response of GLUC1 to 17B—estradiol,
diethylstilbestrol and tamoxifen; N=4 at each concentration; bars indicate
meanzl:1SE; letters indicate means are not significantly different by Fisher’s
LSD, a=0.05; F statistic produced by ANOVA.

78
(Figure 15). GLUC5 and GLUC6 responded similarly to GLUC1 with mean
luciferase expression levels tending toward lower values as exposure to E2
increased (Figure 16).
The mechanism explaining reduction of luciferase expression by ER
agonists appears to directly involve ER. ER mediation of the reduction in

luciferase expression is supported by the responses to E2 and DES, both ER

GLUCS GLUC6

 

 

F-2.42
50 p<0.24

Luciferase (pglmg protein)
D
o
>
Luciferase (pg/mg proteln)

 

 

 

 

 

0 0.1 1 0 0.1 1
173-Estradiol (nM) 170-Estradlol (nM)

Figure 16. Luciferase expression response of GLUC5 and GLUCB to 1715-
estradiol; N = 2 at each concentration; bars indicate mean :l:1 SE; letters indicate
means are not significantly different by Fisher’s LSD, a=0.05; F statistic
produced by ANOVA.

agonists. Blockage of the expression of luciferase could be caused by binding
of the E2 or DES liganded ER to the ERE sequences, consequently resulting in
a blockage of the function of the MMTV promoter. Enhancer sequences
normally work to induce promoter function by forming a looped DNA structure
that stabilizes the RNA polymerase and other elongation factors at the
promoter start site (Cullen, et al. 1993). It would appear that the proper

formation of an active RNA polymerase complex at the MMTV promoter did not

79

occur in MCF-7 cells stably transfected with pGudlu52.2 and that an estrogen-
activated ER mediated this blockage. Tamoxifen reversed the ER mediated
blockage of luciferase expression. Tamoxifen binds competitively to ER but
inhibits the proper formation of an active ER dimer complex, thus inhibiting
effective binding of ER to the ERE. Therefore, ERE binding by ER appears to be
related to the blockage of luciferase expression. This evidence suggests that
the specific construction of ERE and MMTV promoter in pGudlu52.2 was not
conducive to ER mediated enhancement of luciferase expression when stably
transfected in MCF-7 cells. Perhaps the distance between the ERE and MMTV
promoter was not optimal, though DREs located at the same distance
effectively enhanced luciferase expression in a trout cell line (El-Fouly, et al.
1994). Another explanation may be that the MMTV promoter does not function
optimally in MCF-7 cells (as evidenced by the low background level of
expression) and that binding of ER to the EREs blocks binding of the DNA
polymerase complex to the MMTV promoter start site. Further mechanistic
studies are required to elucidate the mechanism of action of pGudluc2.2 in
MCF-7 cells.

Whatever the explanation for the unexpected activity of pGudlu52.2 in
MCF-7 cells, the practical conclusion is that the GLUC cell lines probably can
not be used for rapid, large-scale screening of estrogenic or anti-estrogenic
compounds because the absolute level of expression of luciferase was too low
to allow adaptation of the assay to a 96 well format. The reversed response
of the GLUC cell lines to E2 and other estrogens could be useful if the

background expression level could be raised by two to three orders of

80
magnitude, thus providing a wider dynamic range. A reversed response curve
would also allow improved reproducibility because variablility in luciferase
measurements generally increases with increasing luciferase concentration.
Also, cytotoxicity would not confound the response to anti-estrogens as found
in MCF7LUC cells because anti-estrogens in GLUC cells produce increased
luciferase expression whereas cytotoxicity reduces luciferase expression.

A new plasmid, pGudluc3, has been designed that incorporates the
strong and widely effective SV40 promoter and a modified firefly luciferase
gene that promotes transcription of the gene in vertebrate cell lines, based upon
the pGL3 luciferase reporter vectors (Promega Corp.). The ERE sequences will
be tested for their ability to enhance luciferase expression from two sites in the
plasmid, both upstream and downstream of the luciferase gene. In the
pGL3Control vector, the SV40 enhancer sequence is normally located
downstream of the luciferase gene. Therefore, in one version of pGudluc3, the
ERE sequence will be placed downstream of the luciferase gene replacing the
SV40 enhancer normally located there. Also, the number of EREs in the
sequence will be varied by concatemerization to identify the optimal number of
EREs needed for enhancement of luciferase gene transcription. These
experiments will be conducted by other researchers in the Aquatic Toxicology

Laboratory at Michigan State University.

CHAPTER 4
REPRODUCTIVE IMPAIRMENT AND INDUCTION OF PLASMA ALKALINE-
LABILE PHOSPHATE, A BIOMARKER OF ESTROGEN EXPOSURE, IN
FATHEAD MINNOWS (PIMEPHALES PROMELASI EXPOSED TO

WATERBORNE 17B—ESTRADIOL

Introduction

The aquatic environment frequently becomes a sink or site of first effects
for environmental contaminants, including estrogenic contaminants. Therefore,
it is important to study the effects of estrogenic contaminants in aquatic
organisms. The fathead minnow (Pimephales pramelas) is a widely used
organism for aquatic toxicity testing and is an appropriate model organism for
examining estrogenic endpoints, though work to date has fosuced on larger
species, including rainbow trout (Purdom, et al. 1994), carp (U.K. Directorate,
1992), channel catfish (Nimrod and Benson, 1996), goldfish (Wiegand and
Peter, 1980), Siberian sturgeon (Pelissero, et al. 1991) and flounder (Emmersen
and Petersen, 1976). An advantage to using the fathead minnow as a model
organism is that it can be readily induced to breed in captivity, facilitating the
development of assays that monitor the effect of waterborne substances on
reproduction (Donaldson, 1990; Donaldson and Scherer, 1983). Successful
reproduction is critically important to the long term survival of fish species.
Therefore, importance should be attached to the determination of the effects
of a waterborne contaminant on fish reproduction, just as signficant efforts

81

82

have been made to assess the acute and chronic lethality of contaminants.
Also, because estrogens play a critical role in modulating reproductive activity
of fish (Donaldson and Hunter, 1983; Stacey, et al. 1987), it is logical, indeed
imperative, to study the effects of waterborne xenoestrogens on reproduction.

The objectives of this research were: 1) to assess the effects of
waterborne 17B—estradiol on the reproductive function of fathead minnows as
a benchmark to which other xenoestrogens can be compared, and 2) to
correlate the effects on reproductive function with a biomarker of estrogen
exposure, plasma vitellogenin (Vtg) expression, measured as plasma alkaline-
labile phosphorous (plasma ALP). (A third objective of the overall research
project was to relate impairment in reproductive function to histopathological
changes in the gonads. However, histopathology was analyzed and reported
in another Ph.D. dissertation (Miles-Richardson, 1996) and will not be reported

here.)

Materials and Methods

Chemicals

Crystalline 17B—estradiol was obtained from Sigma Chemical (St. Louis,
MO) and stored in the dark under nitrogen. A stock solution (10 mM) in ethanol
was stored at -20 °C for use in making the aqueous working solution. All

chemicals were of reagent grade.

83

Experimental Design

Fathead minnows reproduce under specific conditions that can be
manipulated to induce reproduction. Therefore, the experimental design“ was
optimized to account for these conditions. Reproduction was defined in this
assay as the number of eggs laid per female per tank during the experimental
period. Because fathead minnows require the presence of more than one male
and female per tank to optimally induce egg laying, the reproductive
performance of individual females could not be determined. Also, reproductive
output is strongly affected by the actions of the male. Specifically, males
entice females to enter a breeding area, usually a recessed area under some
structure, and by vigorous lateral rubbing induces the female to lay adherent
eggs on the underside of the structure, simultaneously fertilizing the eggs. The
male than expels the female from the breeding area and takes up the task of
cleaning and protecting the eggs from predation until hatching. It has been
observed in the Aquatic Toxicology Laboratory that the presence of more than
one male in the breeding tank enhances the development of the secondary
sexual characteristics of the males, though crowding leads to excessive
mortality. Also, placing more than one female in the breeding tank improves
the probability that at least one of the females will have developed mature eggs
in time for the experimental breeding period. A population of three males and
three females per tank was selected as optimum based upon the considerations
mentioned previously and the observations made during pilot studies. Each

male was provided a breeding area constructed from a clay flower pot (7.5 cm

84
diameter), cut in half, and placed on the bottom of the tank to provide a shelter
(hereafter referred to as breeding tiles). Three concentrations of 17B-estradiol
were tested, 27.24 ng/L, 272.4 ng/L and 2724 ng/L (0.1 nM, 1 nM, and 10
nM, respectively), in addition to a solvent control and control receiving no
solvent. Treatments were duplicated and reproduction was reported as the tank
average over the entire period in which reproduction was monitored. Statistical
analyses included ANOVA (SAS PROC GLM, SAS Institute, 1988) with the
nominal concentrations as treatments and regression analysis (SAS PROC REG,
SAS Institute, 1988) using the measured concentrations to compare the

precision of the analyses.

Maintenance and Acclima tion of Fish

Fathead minnows were reared and maintained in the Aquatic Toxicology
Laboratory under a 16:8 lightzdark cycle in continuosly flowing, dechlorinated
tap water at 17-19 °C and fed a 1:1 (v:v) mixture of Tetramin (Tetrawerke,
Germany) and Purina Trout Chow (2-4 mm pellet size) once daily at a rate of
approximately 0.5% of fresh body weight. The dry food was supplemented
with Artemia (brine shrimp) nauplii. Freeze-dried Spirulina algae was fed to fry
up to 2.5 cm length (as this was necessary to avoid vitamin deficiency during
development leading to scoliosis and lordosis in adults).

A three week acclimation period preceded the exposure period. Three
males and three females, reproductively mature and of 12 to 18 months of age

were visually sexed and transferred to randomly assigned 19L glass aquaria

85
held at 19 °C in a circulating water-jacketed bath (Frigid Units Living Stream,
Toledo OH) receiving 3 L/hr dechlorinated tap water through a solenoid-
controlled proportional diluter apparatus (Ace Glass, Vineland, NJ). Each tank
received vigorous aeration, though escape of aerosols from the tanks was
prevented by acrylic lids. Light was delivered at the rate of 16:8 lightzdark from
two 40 watt fluorescent bulbs fitted with a cellulose acetate sheet that blocked
wavelengths of light below 290 nm to minimize UV-induced degradation of
17B—estradiol. The temperature of the water jacket was raised 2.3 °C per week
to a final temperature of 26 °C, the optimal temperature for breeding. The
determination of the sex of the fish was confounded by the fact that some
female-appearing fish later developed male characteristics by the end of the
acclimation period. The identification of male individuals was unambiguos due
to the appearance of large breeding tubercles on the snout, dark vertical
banding, a prominent dorsal fat pad and stiffening and whitening of the first ray
of the pectoral, pelvic and dorsal fins. It was necessary to recheck the identity
of the female fish each week. Any "female" fish that developed male
characteristics was removed and replaced with a female fish taken from a
separate tank held at the temperature of the acclimation tanks. The pattern of
sexual development observed in this laboratory resembles protogyny ("first
female") reported to occur in other fish (Matty, 1985), though protogyny has
not been reported to occur in fathead minnows. Fish were not allowed to breed
during the acclimation period by not placing the breeding tiles in the tanks. The
feeding regimen used during normal maintenance of the fish was continued

throughout the acclimation and exposure periods.

86

Exposure of Fish to 17B-Es tradiol

A three week exposure period commenced immediately after the
acclimation period. 1 73-estradiol (E2) was delivered at concentrations of 27.24
ng/L, 272.4 ng/L and 2724 ng/L via the proportional flow diluter accompanied
by a final ethanol solvent concentration of 0.0001 % (1 ppm ethanol by
volume). The solvent control treatment received 0.0001% ethanol also. A
control treatment received no ethanol. Fish were exposed to 17B-estradiol for
6 days before the commencement of breeding. Breeding was initiated by
placement of the breeding tiles in the tanks and monitored for two periods of
6 days each with one day of rest in between. The experiment was ended two
days early (19 days of exposure in total) due to the development of morbidity
in fish exposed to 2724 ng/L E2. Twice daily, breeding tiles were checked for
eggs, and when discovered, were counted and photographed. A fresh tile was
used to replace the tile containing eggs. Tanks were siphon-cleaned of
accumulated food and fecal material weekly. Wastewater was treated by

trickle filtration through 100 kg activated charcoal.

Collec tion of Tissues

At the end of the exposure period tissues were collected for
histopathology and analysis of plasma vitellogenin. Fish that died before the
end of the experiment were not used in the analysis of histopathology or

vitellogenin because autolysis occurred rapidly after death. After anaesthesis

87
with a lethal concentration of tricaine methanesulfonate (1 g/L), the caudal
peduncle was rapidly severed with a scalpel dipped in heparin solution (10
mglml) and blood collected from the caudal vein with a heparinized hematocrit
tube. Blood volumes collected ranged from 20-250 pl. The abdomen was
injected with Bouin’s preservative solution to speed preservation of internal
organs. Then, the entire fish was immersed in Bouin's solution for 24 hr and
transferred to ethanol for storage until histopathological analysis. A subset of
male gonads were dissected out before treatment with Bouin’s solution for
ultrastructural analysis by transmission electron microscopy (Miles-Richardson,

1996).

Vitellogenin Analysis

Blood, collected in heparinized hematocrit tubes, was centrifuged at 1 300
g for 10 min at 4°C to separate red blood cells from the plasma. The plasma
was stored at -80°C until analysis of alkaline-labile phosphorous. Plasma
alkaline-labile phosphorous has been used as a non-specific indicator of Vtg
expression (Wallace and Jared, 1968; Parker and McKeown, 1987) because
Vtg is highly phosphorylated and comprises a large majority of the
phosphorylated plasma proteins when it is induced. Vitellogenin is a high
molecular weight (130-180 kd) protein found in all egg-laying vertebrates,
synthesized and secreted by the liver, transported by the blood to developing
oocytes in the ovary where it is actively transported into oocytes and cleaved

into the two primary components of the egg yolk, phosvitin and vitellin (Chen,

88
1983; Silversand and Haux, 1989; Tyler, et al. 19885; Tyler, et al. 1988b;
Babin and Vernier, 1989; deVIaming, et al. 1984; Ho, 1991). A number of
highly specific immunoassays have been developed for measuring Vtg, though
none have been developed for use with fathead minnow vtg (So, et al. 1985;
Copeland, et al. 1986; Giorgi, et al. 1982; Benfey, et al. 1989). Alkaline-labile
phosphorous, as artha-phosphate released from Vtg and other plasma proteins,
was extracted from 10 pl plasma according to the method of Wallace and Jared
(1968) using 1/2 the recommended volumes. Ortha-phosphate was determined
by colorimetric assay of the acidified phosphomolybdate complex using a

commercially available kit (Sigma Cat. No. 670-A).

Water Quality

Four general water quality parameters were measured throughout the
exposure period, dissolved oxygen and water temperature on every other day
and pH and alkalinity weekly. Alkalinity was determined by titration of
bromcresol green indicator to a pH of 4.3 according to APHA Method 403
(APHA, 1985). 173-estradiol was measured on every other day of the
exposure period. 5.0 ml of water was volumetrically sampled from each tank
and stored in 13 x 100 mm test tubes capped with Teflon® lined lids at -80°C
until analysis. Extraction of 52 from the water was accomplished by two 2 ml
extractions with diethyl ether. After combining the ether extracts, the solvent
was evaporated by a gentle stream of nitrogen at room temperature. E2 was

redissolved in 0.5 ml buffer and quantified by ELISA (Cayman Chemical, Ann

89
Arbor, MI).

Statistical Analysis

Quality control parameters for the measurement of E2 included relative
standard deviation, limit of detection and recovery (Figure 17). Recovery could
not be quantified for Vtg analysis because characterized standards for fathead
minnow Vtg are unavailable. However, recovery of phosphate from
proteinaceous samples was estimated by Martin and Doty (1949) at
approximately 75% using the phosphmolybdate colorimetric assay.

The effect of waterborne E2 on egg production was analyzed by ANOVA
and nonlinear regression. One-way ANOVA tested the hypothesis that egg
production was not affected by E2 treatment. The Lowest Observed Effect
Level (LOEL) was determined by Dunnett’s t tests (a=0.05) as the lowest
treatment level that caused a statistically significant difference in egg
production compared to the control and solvent control treatments, since there
was no effect of solvent on egg production. The No Observed Effect Level was
determined as the highest treatment level that did not cause a statistically
significant difference in egg production. The Maximum Allowable Toxicant
Concentration (MATC) was calculated as the geometric mean of the LOEL and
NOEL and represents the expected threshold for effects. Because ANOVA
utilizes nominal concentrations to define levels, information on the actual
concentration-response curve is lost. Therefore, a nonlinear regression

technique was used to model the concentration-response after a normal

90
distribution function (Equation 8, Bruce and Versteeg, 1992) utilizing the
measured values of E2 in one case and the nominal values in a second case.

Because the nominal values of E2 for the control and solvent control

_ Iog(ECX) - log(Ea
Eggs - Eggsmalx x P — SLOPE +2 (8)

where: Eggs = Egg production, eggs laid/female;
log(E2) = log1o E2 concentration, ng/L;
P[ ] = normal probability density function;
Eggs“, = maximum egg production;
Iog(ECX) = log1o Effective Concentration causing X% inhibition
of luciferase expression, ng/L;
SLOPE = slope function (standard deviate) of normal probability
density function.
2x = normal deviate above which X% of the standard normal
distribution lies, for X=5, 20, 50, 2x: 1.6449, 0.8416, 0.
treatments were 0, they could not be defined on a log-concentration scale.
Therefore, the nonlinear regression model using nominal concentrations was
modified to include the response information from the controls (Equation 9).
Using the nonlinear regression model, the NOEL and LOEL were estimated,
respectively, as the EC5 and EC20 (after back-transformation from the log(EC5)
and Iog(EC20) estimates produced by the model). The E050 was also
estimated. Induction of vitellogenin, as indicated by plasma alkaline-labile

phosphorous (ALP), was also modelled by nonlinear

91

5995...... x P Ewen-10mm +2

If E5 0 then Eggs
SLOPE

(9)
If E2=0 then Eggs

Eggsmax

where: all variables are as previously described.
regression after a normal distribution function (Equation 10). However, this
model was applicable only to the responses of the male fish because female fish
did not exhibit a plateau of ALP with increasing E2 concentration. For female
fish, a log-linear model was fitted to the linear portion of the response curve,

from which the threshold for effects concentration could be estimated.

 

ALP = (ALPM -ALPm,n)x P '°9(Eil:("°’%50"’ -2)) +11me (10)
where: ALP = alkaline-labile phosphorous in plasma, mglml as P;
ALPM = maximum plasma ALP, mglml;
ALP”, = minimum plasma ALP, mglml;
and all other variables as defined above.
The relationship between ALP and egg production was analyzed by log-linear
regression. These models related a biomarker of estrogen exposure, ALP, with

a reproduction parameter, egg production, as a means of establishing a

predictive relationship that may be useful in monitoring environmental exposure.

92

Results and Discussion

Mortality

A total of 9 fish died during the exposure period (Table 10). Most of the
mortalities occurred in the last week of the exposure period, though one female
from the solvent control treatment died on the first day of exposure. Four
males of 12 exposed to 272.4 and 27.24 ng/L nominal E2 died during the
exposure period, apparently due to the effects of the treatment, while only two

females of 12 in this group died, which suggests a sex related difference in

Table 10. Fathead minnow mortality occurring during the exposure period.

 

 

 

 

 

 

 

 

 

 

Nominal
I17B-Estradiol)
Tank (ng/Ll Day of Exposure Sex Possible reason for death
1 2724 9 Male Toxicosis
1 2724 1 2 Male Toxicosis
1 2724 1 4 Female Toxicosis
2 2724 9 Male Toxicosis
2 2724 13 Female Low 00, aeration interrupted
3 272.4 No deaths N0 deaths
4 272.4 16 Male Toxicosis
5 27.24 No deaths No deaths
6 27.24 6 Female Male aggression, crowding
7 Solvent Control No deaths No deaths
8 Solvent Control 1 Female Male aggression, crowding
9 Control No deaths No deaths
10 Control 16 Female Male aggression, crowding

 

93

mortality caused by exposure to E2. However, because this study was not
designed to assess acute toxicity, the sample sizes were small and cannot be

used to definitively determine acute toxicity of E2.

Water Quality

General water quality parameters were uniform over all tanks throughout
the exposure period (Table 11). Dissolved 17B-estradiol was detected in all
tanks including control and solvent control tanks (Figure 17). Two possible
sources of E2 in the control and solvent control tanks was: 1) excretion of E2
and E2 metabolites by females, and; 2) dissolution of E2 from fish meal in the
food. The measured concentrations of the E2-treated tanks were 79:1:25%

(mean :l:SE) of the nominal concentrations.

Egg Production

Fathead minnow egg production was reduced in a concentration
dependant manner by exposure to waterborne 17B—estradiol over a 13 day
period of breeding (Figure 18). The reduction was statistically significant when
analyzed by ANOVA using pooled control and solvent control responses
(F =4.46, p <0.056). Dunnett’s t tests revealed that only the egg production
at the highest nominal concentration of E2, 2724 ng/L, was significantly
reduced with respect to the pooled control egg production. The estimate of

MATC using the ANOVA results was much greater than the MATC estimated

94

Table 11. Water quality during the exposure period, mean(SE).

 

 

Tank pH Alkalinity Dissolved Temperature WIS-Estradiol
(mg CaCO,/L) Oxygen (°C) (ng/L)
(mg/Ll
N = 3 N = 3 N = 19 N = 19 Measured Nominal
N = 10

1 8.17 310 7.2 26.1 1470 2724
(0.38) (3.3) (0.2) (0.04) (652)

2 8.19 302 7.3 26.1 870 2724
(0.38) (6.2) (0.04) (0.04) (281)

3 8.15 318 7.3 26.1 517 272.4
(0.35) (6.7) (0.04) (0.04) (426)

4 8.16 325 7.3 26.1 105 272.4
(0.36) (17.6) (0.05) (0.03) (15)

5 8.19 302 7.3 26.1 13.0 27.24
(0.38) (8.5) (0.05) (0.04) (1.7)

6 8.13 328 7.3 26.1 30.6 27.24
(0.35) (1 1.7) (0.06) (0.04) (9.9)

7 8.19 327 7.3 26.0 6.05 0
(0.38) (4.4) (0.05) (0.04) (2.50)

8 8.20 318 7.3 26.1 3.54 0
(0.38) (4.3) (0.04) (0.04) (0.70)

9 8.15 322 7.4 26.0 14.8 0
(0.33) (3.3) (0.04) (0.04) (4.5)

10 8.05 325 7.4 26.0 9.67 0
(0.29) (2.9) (0.08) (0.07) (3.80)

 

by nonlinear regression (Table 12). However, the use of nominal versus
measured E2 concentrations in the nonlinear models produced nearly identical
regression curves and estimates of EC50, E020 and EC5 that were not
statistically different (Figure 18). In fact, the standard error of the estimates
was greater using the measured E2 concentrations, probably because the
estimation of the maximum egg production level was less certain than that
derived from the nominal E2 concentration model. The two methods of
calculating the MATC, ANOVA vs. regression, produced widely differing values
because ANOVA is a very conservative estimation method and regression is a

generally less conservative method. Because ANOVA is strongly affected by

 

 

    

10000 1
5 asp-7.511.311, (N-116)
I “ Recovery-6321296, (rt-1o)
3 Method LCD-0.91 ng/L (5 ml sample)
\ _
3 1000 g .
c i
V I
E l .
U .- "“ T‘ 5
15 100 E
h - A
u n '
0 - z .
" = O
\\- ‘9‘
I I l l I l l l I I I l 1 l

 

 

l l l
0 1 2 3 4 5 6 7 8 9 10111213141516171619
Symbol-MeansSE, N-2
Exposure Day

Figure 17. 17B-Estradiol measured by ELISA during exposure period,
RSD=Relative standard deviation of duplicate measurements, Method
LOD =Method limit of detection, Recovery=recovery of spiked samples
containing 2724 ng/L E2; + =Tank 1, +(bold)=Tank 2, A =Tank 3, 4 =Tank
4, O =Tank 5, O =Tank 6, v =Tank 7, v =Tank 8, O =Tank 9, O =Tank 10.

variation at each level, significant differences in levels are difficult to detect,
especially with highly variable data such as egg production. The regression
model incorporates the variation over the entire response curve, thereby
reducing the effect of individual variability on the estimate of MATC. It is

important to note that there is no general consensus on what level of effect

96

should be defined as NOEL and LOEL when using a regression model. EC5 and
EC20 were selected to estimate NOEL and LOEL based upon the suggestions
of Bruce and Versteeg (1992). A 5% change (EC5) probably represents the
threshold below which effects can reasonably be expected to be biologically
unimportant and therefore was selected as the NOEL. A 20% change (EC20)
may be considered a reasonable level of effect beyond which detrimental
biological consequences could be expected and was chosen as the LOEL.
Because of the significant variability in egg production within each level, the use

of ANOVA for determining an MATC was probably the better method.

Vitellogenin lnduc tion as a Biamarker of Reproductive Impairment

Vitellogenin expression, as indicated by plasma ALP, was significantly
induced by exposure to waterborne 17B-estradiol in both female and male
fathead minnows (Figure 19). The threshold concentration of E2 for induction
of vitellogenin was similar in females (46 ng/L E2) and males (51 ng/L E2)
though females exhibited a baseline ALP level 6 times that of males (0.4726
mglml P females vs. 0.078 mglml P males). As a biomarker of reproductive
impairment by waterborne 1715-estradiol, log(ALP) was significantly correlated
with egg production. Male log(ALP) was slightly less predictive of egg
production (Figure 20B), explaining 50% of the variation (r2 = 0.50), than female
log(ALP) which explained 62% of the variation in egg production (Figure 20A).
The observation that the trend in reduction of egg production continued into the

lowest concentrations of ALP suggests that induced plasma vitellogenin could

97

Table 12. Calculation of No Observed Effect Level (NOEL) or EC5, Lowest

Observed Effect Level

(LOEL) or EC20, Maximum Allowable Toxicant

Concentration (MATC), and EC50 using ANOVA and nonlinear regression of
nominal and measured 17B—estradiol concentrations for reduction of egg

production of fathead minnows.

 

Method NOEL l EC5 LOEL / EC20 MATC EC50
(ng/L) (ng/L) (ng/L) (nglL)
ANOVA 272.4 2724 861 -
Nonlinear/Nominal 7 .4 33.6 1 5.7 1 64
Nonlinear/Measured 3.0 1 8.2 7.3 122

 

 

 

 

 

250 Nominal E2 (-.-)
Slope-0.822058, Eggsm-158225
i? _ Threshold-7.3 ngIL O MOI-mod 82 (-O -)
a 200 l Slope-03821.23, Eggsw-1662108
s _ - .: . (EstlmatetSE)
*- 150 Nnruhold-m ng/L
3 O I ‘ ~ 2 ‘ O O
5 : £050 - 122 nglL
1, . (1.119555014092124)
o; 100 ‘ 1 ‘ x \' 1:550 - 154 ngIL
- l . , (Log(E050)-2.222043)
W I l
a) I I
a) 50 r | '
Ill l I
I
I I
o a 2 LI 2 eel
1 10 100 1 000

  
   

 

 

 

17B-Estradiol (ng/L)

Figure 18. Reduction of egg production by exposure of fathead minnows to
waterborne 17B-estradiol (E2) modeled by nonlinear regression after a normal
distribution function; 9 =E2 measured in each tank by ELISA, O =E2 nominal

concentration.

98

prove to be a sensitive indicator of potential reproductive impairment in fish.

Mechanism of Reproductive Impairment Caused by 17B—Estradiol

178-estradiol is the most potent of the natural occurring estrogens.
Therefore, this study provides a valuable benchmark for determining the type
of effects expected to be caused by substances acting through the estrogen
receptor. 17B-estradiol significantly affected the physiology of both male and
female fish, as evidenced by elevated plasma ALP and reduced egg production.
Because this study was not designed to fully investigate the biochemical and
physiolgical actions of E2 underlying the reduction of egg production, the
mechanism of this action remains to be fully elucidated. However, it seems
likely that the reduction in egg production was related to effects in both males
and females. Estrogen injection of male fish caused feminization of secondary
sexual characteristics, testicular degeneration and inhibition of spermatogenesis
in several fish species (Matty, 1985). In this experiment, waterborne exposure
to 17B—estradiol caused reduction of male secondary sexual characteristics and
testicular degenerationlMiles-Richardson, 1996). Although the changes in
testicular histology probably reduced spermatogenesis and male fertility, it was
not determined whether the changes in secondary sexual characteristics could
have significantly altered male sexual behavior, which is critical for succesful
laying of eggs. Reduction in the male "libido" by estrogen exposure remains a
possible mechanism of reduction of egg production. Overt toxicity was

suggested by the mortality affecting males at the highest E2 exposure

99

 

  
   
    

 

 

 

 

 

  
    
   
   
   
 

4.00
A Y-1 .2577(log(X))-1.4611 .
r'-0.45, p<0.02
5" 3.20 -
E
\
O
E 2.40 -
a O
_l
< O
6 15° ' Threshold
g 40 ng/L £2 0
2
n, 0.00
O
0
000° ‘ t““ ‘l A Ill-An
1 10 100 1000 10000
1 7B-Estredlol (ng/L)
4.00
Slope factor-0.500101”
B PHI-233030.430 mglml, P--0.070:0.071 mglml
EetlmeletSE
’-" 3.20 - ‘ l
E
\
0 5050-251 ng/L .
E 2 4o . (Log(ECSO)-2.400:0.MI)
o. ' O
.J
< EC20-05 ng/L
. 1.50 - (Leg(£C20)-1.02030.110)
5 Threshold
5 51 ng/L
E 0-80 ' lacs-00 ng/L
(Leg(ECS)-1.470t0.220)

 

 

 

 

0.00 ' ‘ ‘ ‘ ““
1 10 100 1000 10000

17B-Estradlol (ng/L)

Figure 19. Induction of plasma alkaline-labile phosphorous (ALP), an indication
of vitellogenin expression, in female (A) and male (8) fathead minnows exposed
to waterborne 178-estradiol (ELISA measured concentrations of E2); in A,
O =measurements not included in log-linear regression, sub-threshold
mean=0.473 mglml P, 95% UCL=95% upper confidence limit on the sub-
threshold mean, threshold = intersection of regression line with 95% UCL; in B,
thresho|d=geometric mean of log(EC5) and Iog(EC20), nonlinear regression
used to model apparent plateau.

100

 

250

A Y--130.2(log(X))+83.45
r‘-0.02, p<0.0001

150 "

100 l'

Eggs Iald (per female)

 

 

 

 

0.01

Plasma ALP (mg/ml)

 

250

B Y--70.14(log(X))+08.83
r‘-0.50, p<0.0001

200 '
150 "
100 "

50)

Eggs laid (per female)

 

 

 

 

D
d P
"I

0.01 0.1

Plasma ALP (mg/ml)

Figure 20. Correlation between egg production and plasma alkaline-labile

phosphorous in female (A) and male (8) fathead minnows exposed to
waterborne 17B-estradiol.

101
concentrations. The relatively great induction of vitellogenin in males may have
been related to this mortality (Herman and Kincaid, 1988; Carragher and
Sumpter, 1991).

The female gonads were also affected by 178-estradiol exposure.
Increased follicular atresia (breakdown of mature follicles), an increased
prevalence of primary follicles and reduced mature follicles was related to E2
exposure (Miles-Richardson, 1996). The reduction of mature follicles in E2-
exposed fish is directly related to a reduction in egg laying simply because
fewer mature follicles were available for laying. ln normally developing and
spawning fish, a surge of plasma 17B-estradiol initiates vitellogenesis and egg
development well in advance of the spawning event, such that E2 levels return
to background levels during spawning (Scott and Sumpter, 1983). In this
experiment, E2 was administered after the female fish had been in the process
of developing oocytes during the acclimation period. The effect of E2
adminstration appears to have been to "reset" the egg development "clock" by
triggering atresia of mature follicles and initiating another sequence of
vitellogenesis and egg development. Mortality of females appeared to have
been unrelated to E2 exposure. The toxicity of vitellogenin postulated to occur
in male fish was probably avoided in females because the developing eggs
provided a location for the deposition of the vitellogenin, a process unavailable
to male fish. Vitellogenesis was greatly induced in females exposed to E2 but
was not at similar levels in unexposed females, even though unexposed females
laid large numbers of eggs. Thus, unexposed females were probably not

actively synthesizing vitellogenin and developing eggs at the end of the

102
exposure period. The primary effect of E2 exposure on female fathead
minnows was to alter the timing of egg production, thereby blocking spawning
during the exposure period of the experiment. In an environmental setting, the
exposure of gravid female fish to an estrogenic substance could reset the
reproductive cycle and cause the female to miss critical windows of opportunity
for successful spawning, especially in seasonally reproducing species. Also,
vitellogenin induction in male fathead minnows was correlated with E2 exposure
and may have contributed to the excessive mortality of males. The results of
this experiment firmly establish critical morphological, histological, and
biochemical indicators of waterborne estrogen exposure and establish a
relationship between these indicators with reproductive performance, an

important parameter affecting fish populations in the environment.

103
SUMMARY AND CONCLUSIONS

Estrogen mimics are a ubiquitous component of the environment
occurring in foods, plants, and fungi and in chemicals of human origin. Because
estrogens play an important role in sexual development, behavior and
reproduction, there is valid cause for concern over "disruption" of the normal
activity of estrogens by both natural and synthetic estrogen mimics. However,
this concern should be focused by issues of the dose and potency of the
particular compounds thought to be estrogen mimics. This research has sought
to develop a framework for evaluating the potential for substances to affect
estrogen action. The molecular properties governing the affinity with which
hydroxylated PCBs bind to the estrogen receptor was investigated, revealing
that moderate precision in predicting binding affinity can be achieved with a
relatively small set of general molecular parameters. However, it is important
to note that receptor binding affinity is only one aspect of estrogen mimickry.
Therefore, a functional bioassay was necessary to define the activity of the
hydroxylated. The MCF-7-LUC cell line provided a rapid and specific assay for
both estrogenic and anti-estrogenic activity. The results of the MCF-7-LUC
assays demonstrated that receptor binding, while important in indicating which
compounds had the potential to specifically interact with the estrogen receptor,
could not predict whether a particular compound exhibited estrogenic or anti-
estrogenic activity. Rather, it appears that very specific structural
characteristics differentiate the mode of activity. Although the MCF-7-LUC cell

line was useful in screening hydroxylated PCBs, alternative cell lines were

104
desired for testing on a more species—specific basis, such as the rainbow trout
cell line RTH-149. Therefore, development of an estrogen-inducible reporter
gene system was attempted for introduction in fish cell lines. Unfortunately,
the plasmid system adopted in this research proved to be incompatible with the
estrogen responsive elements that were introduced into the system, even
though a similar arrangement was successful using dioxin responsive elements.
The specific mechanism by which the estrogen responsive elements caused
reduction in luciferase expression was not elucidated because it was beyond
the scope of this work. However, it may prove to be an interesting example of
estrogen receptor induced repression of gene expression. Finally, a functional
bioassay using whole animals was developed. Fathead minnows were selected
because of their favorable breeding characterstics to serve as a model organism
for testing the effects of waterborne xeno-estrogen exposure on reproduction.
The research project completed study on the benchmark estrogen, 17(3-
estradiol, and identified a suite of responses at the biochemical, physiological,
and histopathological levels that will be useful in characterizing exposure to

estrogenic compounds in the aquatic environment.

APPENDIX

APPENDIX

Standard Operating Procedure

Estrogen Receptor Prepare tion and Competitive
Binding Assay

Version 2.1 February 23, 1995
Prepared by: Vincent J. Kramer

Aquatic Toxicology Laboratory
Department of Fisheries and Wildlife
Michigan State University
East Lansing, Michigan 48824-1222

A. Preparation of estrogen receptor from tissue or cell culture

1. Remove tissue (for example, calf uterus, fish liver etc.) from animal
and quickly freeze in liquid nitrogen unless tissue is to be processed
immediately. For cell culture, see Preparation of ER from cells in culture.

2. All procedures should be performed at 4 °C. Mince tissue to blocks
about .5 cm in width in an equal volume of homogenization buffer.

Homogenization buffer
10 mM Tris HCI pH 7.5

1.5 mM Ethylenediaminetetraacetic acid, disodium salt (EDTA)
1 mM Dithiothreitol (DTT)

10 % glycerol, v:v

1 mM Sodium Molybdate

0.1 mM Phenylmethylsulfonyl fluoride (PMSF) or

0.5 mM Aminoethylbenzenesulfonyl fluoride (AEBSF)

No te: AEBSF is more stable and soluble than PMSF.
3. Homogenize thoroughly on ice with mechanical homogenizer (for
example Tekmar Tissuemizer) , 5 min. pulsed on high setting to avoid

heating the sample.

4. Centrifuge first at 4 °C, 800 g, 20 min. to pellet debris. Centrifuge
supernatant at 4 °C, 100000 9 (42000 rpm Beckman Ti70 rotor) 30 min

105

106

to produce "cytosolic" fraction.

5. If it is suspected that the receptor preparation contains endogenous
178-estradiol (as in tissue of adult organisms) the endogenous E2 must
be removed by charcoal stripping.

5a. Prepare Dextran Coated Charcoal (DCC):
0.5 % Norit A charcoal 50 mesh w:v

0.05 % Dextran, 70 Kd w:v

in TEDG buffer.

5b. Add 1 ml DCC to 10 ml supernatant and warm to room temp
and hold at room temperature 5 min. to release E2 from receptor.

Note: Estrogen receptor is heat labile.
5c. Centrifuge at 5000 g, 10 min, 4 °C.

6. Aliquot supernatant 1 ml each into 1.5 ml microcentrifuge tube, freeze
at -80°C until used.

7. Perform Scatchard analysis (See Part B) and protein concentration
measurement characterize preparation.

B. Competitive binding assay using radio-labelled 1 73-estradiol

This procedure is adapted from Murdoch et al. 1990. It measures the inhibition
of binding of the native ligand, 178-estradiol, to the estrogen receptor by a
competitor or mixture of competitors. The ligand is radio-labelled so that the
amount of radioactivity bound to the estrogen receptor is related to the extent
of inhibition of binding by the competitor. For example, a high affinity
competitor will displace most of the radio-labelled ligand resulting low
radioactivity bound to the receptor. The amount of radioactivity bound to the
receptor is determined by adsorption of the protein (and protein-bound
radioactivity) in the assay mixture to hydroxyapatite which is then quantitated
by scintillation counting. Since the radio-labelled ligand also binds non-
specifically to other proteins in the assay mixture, a separate experiment must
be performed to quantify the non-specific binding. This is accomplished by
displacing 100% of the radio-labelled ligand from the estrogen receptor with a
high concentration of the non-steroidal estrogen diethylstilbestrol which binds
specifically to the estrogen receptor. Total binding must also be assessed by
measuring the binding of radio-labelled ligand in the absence of competitor.

Note: All steps in the following procedure must be carried out at 4 °C.
Therefore, use of a cold room is highly recommended since cooling the

107

filter apparatus and solutions is cumbersome on the bench top.

1. Prepare/label 10 glass test tubes (15 X 85 mm borosilicate culture
tubes) as follows:

1a. Tube 1, "T", total binding measurement;

1b. Tubes 2-9, "Competitor concentration, nM", concentration of
competitor in assay mixture;

1c. Tube 10, "NS", Non-specific binding measurement.
2. Prepare buffer and working solutions.

2a. 3H-178—estradiol working solution: Make 500 pl 250 nM 3H-E2.
Stock solution is approximately 10 pM 3H-E2 (DuPont-NEN,
2,4,6,7-3H-178- estradiol, 100 Ci/mmol, in ethanol). The exact
volume of stock solution used will vary as determined by the
specific activity and concentration of the stock solution which is
unique to each lot of 3H- E2. The final ethanol concentration is
~2.4%.

2b. Diethylstilbestrol working solution: 1 mM in ethanol.

2c. Competitor working solutions: Highest concentration—10 mM
or limit of solubility. Dilute subsequent solutions serially by
1/10th such that a minimum of four concentrations are tested
spanning four orders of magnitude (for example 10 mM, 1 mM,
100 pM. 10 pM). All solutions should be dissolved in 100%
ethanol. When diluted in assay test tube, the concentrations will
be 1/ 100th of the working solution concentration.

No to: If the competitor is insoluble in ethanol, 1 -propanol has been
used successfully (for example, to dissolve coplanar PCB ’s and
chlorinated naphthalenesl.

2d. TEDG assay buffer-Make 1 L for an assay of 10 test tubes.
Sterile filter and store at 4 °C.

TEDG Assay Buffer

10 mM Tris HCI, pH 7.5

1.5 mM Ethylenediaminetetraacetic acid, Na2
1 mM Dithiothreitol

10% glycerol v:v

108

2a. Hydroxyapatite (HAP) slurry: 160 mg dry HAP/ml TEDG.
Store at 4 °C.

3. Add to each test tube in order:
3a. 900 pl TEDG assay buffer

3b. 10 pl competitor (Tubes 2-9), solvent (Total binding
measurement, Tube 1), or DES (Non-specific binding
measurement, Tube 10).

3c. 4o pl 250 nM 3H-Is2

3d. 50 pl ER preparation (Final ER concentration should be about
1 nM or less. This is dependant upon specific activity of the ER
preparation).

Note: The final solvent concentration in the assay is 1 96.
4. Mix well by gentle swirling, incubate 4 °C, 2 hours, in dark.

5. Add 300 pl HAP slurry to adsorb proteins.
Mix well by hand and incubate 15 min, 4 °C, mixing every 5 min.

No te: Do not vortex solutions containing radioactive materials
because aerosoliza tion will occur posing the rislr of inhalation of
radioactivity.

6. Filter HAP out of solution using a 10 place filter manifold (Hoefer
Scientific) with stainless steel holders, 24 mm Whatman GF-C glass fiber
filters.

No to: Do not use Gelman glass fiber filters because they are too
brittle.

Note: Install an aerosol trap between the filter manifold and the
vacuum source to prevent radioactive contamination of vacuum
source.

7. Wash HAP 6 times with 5 ml TEDG.

No te: This is a critical step in the procedure. All unbound radio-
label must be rinsed from HAP to insure reproducibility.

8. After final wash, rinse down sides of filter holder with 1 ml TEDG and
allow vacuum to dry filter slightly. Add 1 drop TEDG to the center of

109

each filter to ease removal of filter disc from apparatus. Add any HAP
adhering to the filter holder to the scintillation vial.

9. Place filter with HAP in scintillation vial. Add 6 ml scintillation cocktail
(Safety-Solve for aqueous samples). Shake vigorously. Allow sample to
stand at room temperature overnight before counting to allow scintillation
fluid to completely extract 3H-E2 from HAP. Count tritium using
scintillation counter.

10. Prepare blanks and standards with HAP and filter in vial to mimic
quenching of sample. A typical standard consists of 1 pl 250 nM 3H-E2
Working Solution (expected disintegrations per minute (dpm) is about
62,000 depending upon specific activity of stock solution which is
different for every lot). Counting efficiency is typically approximately
0.35 but must be determined for every experiment.

C. Data analysis

Data is reduced and summarized using a spreadsheet program The probit
method for calculating lC50 (Bruce and Versteeg, 1992) is executed using the
statistical software package SAS (Appendix 2).

1. Subtract blank cpm from all samples.
2. Convert to pmol 3H-E2 bound for all samples.

3. Subtract non-specifically bound 3H-E2 (the reading from the "NS" tube)
from all samples to determine specifically bound 3H-E2.

4. Calculate lC-50 (concentration of competitor causing 50% reduction
in specific binding of 3H-E2) using probit method for continuous data
(Bruce and Versteeg, 1992).

5. Scatchard analysis is performed to determine Kd and Bmax for the
receptor preparation only. The typical analysis consists of measuring
specific binding at several concentrations of radio-labelled estradiol,
typically an order of magnitude above and below the expected Kd
(approximately .1 nM). The data is plotted according to the equation:

B/F = Bmax/Kd - T/Kd * B
where B=specifically bound 3H-E2, pM; F=free (unbound) 3H-E2, pM;

Bmax =maximum specifically bound 3H-E2; Kd =the dissociation constant,
pM (Clark and Peck, 1979).

110

SAS Program for Analysis of Competitive Binding Curves

/* Non-linear regression model for continuous */

/* toxicity data. Bruce and Versteeg. 1992. ET&C 10: 1485-1494*/
/* Models a cumulative normal distribution */

/* This version is for analysis of estrogen receptor competitive */

/* binding curves. */

/* c#=tritiated estradiol bound specifically, pM, for competitor c*/
/* non-specific binding already subtracted from total bound */

/* bmax=maximum tritiated estradiol bound, pM */

/* specifically, also upper limit of displacement curve */

/* log(icXX) =Iog10 of the icXX, usually XX: 50% */

/* sigma=slope function of inhibition curve */

/* NOTE: Original code used the Gauss searching algorithm. */

/* This version uses DUD to avoid problems associated with *l

/* small N and erratic der.sigma */

/* ALSO NOTE: The "nohalve" option was removed, this improved */
/* stability of search. */

data a;

input i c1 date$;

/* date=date data collected */

/* i=inhibitor concentration, nM */

if i=0 then log_i =.;
else log_i =Iog10lil:
cards;

0.0 100.0 6/12

1.0 99.76 6/12

10 93.08 6/12

100 66.32 6/12

500 44.79 6/12

1000 28.99 6/12
5000 37.74 6/12
10000 10.29 6/12
50000 3.20 6/12
0.0 104.31 6/22
1.0 115.12 6/22
10.0 97.94 6/22
100 61.66 6/22
500 55.08 6/22
1000 57.75 6/22
5000 42.29 6/22
10000 25.31 6/22
50000 20.41 6/22

title, 'Competitive binding curve calf uterine estrogen receptor';

111

data b;

set a end =last;
output;

if last then do;

c1 = .; i= .;

do log_i=-1 to 5 by .25;
output;

end;

end;

proc print data = b;
var i log_i c1 ;

proc nlin data = b method = dud maxiter = 300 smethod = cubic converge = 1e-10

I

title2 'model:c1 =bmax*PllloglC50-log_i)/sigma)’;
title3 'weighted regression er=calf uterine’;
parms log_ic50 = 2.5 sigma = 1.4 bmax = 100;
bounds sigma > 1e-10, bmax > 0;

if i ne 0 then do;

2 = (log_ic50-log_i)/sigma;

ex=z;

px = probnorm(ex);

model c1 =bmax*px;

end;

if i=0 then do;
model c1 =bmax;
end;

output out = d predicted = pred;

proc plot data =d;

plot c1 *log_i = '0'
pred *log_i = ' . ’/overlay;
titleZ;

proc print data =d;

var i log_i c1 pred;

run;

112

Preparation of EB from cells in culture

1. Four days before harvesting ER from cultured cells (for example, MCF-
7 cells), change medium from 10% FBS Medium (the normal culture
medium) to 5% DCCFBS Medium. Grow up at least twenty (20) 100
mm diameter culture plates (75 cm2) to make about 10 ml of ER
preparation.

2. One day before harvesting, switch medium to Serum-free Medium.
This causes the cells to depurate 17B-estradiol via metabolism.

10% FBS Medium

Dulbecco's Modified Eagle’s Medium/

Ham’s F12 Nutrient Mixture

without Phenol Red and sodium bicarbonate

with 15 mM HEPES buffer and L-glutamine
(Sigma 02906)

10% Fetal bovine serum (Hyclone)

2 mM Sodium pyruvate

1 pg/ml Bovine insulin

10 U/ml:10 pg/ml Penicillin:Streptomycin

5% DCCFBS Medium

Same as 10% FBS Medium except replace
10% Fetal Bovine Serum with

5% Dextran-coated charcoal stripped FBS
(available from Hyclone Laboratories)

Serum free Medium
Same as 10% FBS Medium except omit
10% Fetal Bovine Serum

3. Remove cells from plate surface by incubating cells in 1 mM EDTA in
Ca-Mg free phosphate buffered saline for 30 min at 37 °C. Cells should
lift easily off plate without scraping. Transfer cells to 15 ml conical
centrifuge tubes at room temperature. Combine cells from 20 plates.

4. Centrifuge at 800 g at room temperature 10 min. to pellet cells.

5. Discard supernatant, wash cells once with Homogenization Buffer as
in Step 4.

6. Resuspend cells in 2 ml ice-cold Homogenization Buffer and transfer
to either a Potter-Elvejhem homogenizer with Teflon Pestle or a Dounce
Homogenizer (whichever is available).

113

7. Disrupt cells with 60 passes of the pestle on ice. Rinse pestle when
completed with 1.5 ml Homogenization buffer into the homogenate.

8. Transfer homogenate to 15 ml conical centrifuge tube on ice. Rinse
homogenizer with 1.5 ml buffer and combine with homogenate.

9. Centrifuge 10 min. 800 g at 4 °C to pellet nuclei and debris.

10. Transfer supernatant ("cytosol") to 30 ml polycarbonate
ultracentrifuge tube. Wash pellet with 1 ml Homogenization buffer and
centrifuge as in Step 9. Combine supernatant with homogenate in
ultracentrifuge tube.

11. Centrifuge "cytosolic" fraction at 30 min. 42000 g at 4 °C in
Beckman L7 Ultracentrifuge or appropriate substitute.

12. Transfer supernatant to 15 ml conical centrifuge tube or an
appropriate number of 1 ml centrifuge tubes for storage at -80 °C (or
in liquid nitrogen if available).

LIST OF REFERENCES

LIST OF REFERENCES

Adkins, E.K. 1978. Sex steroids and the differentiation of avian reproductive
behavior. Am. Zool. 18:501-514.

Adkins-Regan, E. 1987. Hormones and sexual differentiation. In: 0.0. Norris
and RE Jones. (Ed.) Hormones and reproduction in fishes, amphibians,
reptiles. Plenum Press, NY. pp. 1-29.

Anstead, G.M., C.S. Peterson, K.G. Pinney, S.R. Wilson, J.A.
Katzenellenbogen. 1990. Torsionally and hydrophobically modified 2,3-
diarylindenes as estrogen receptor ligands. J. Med. Chem. 33:2726-2734.

APHA. 1985. Standard Methods for the Examination of Water and Wastewater.
16th Edition. American Public Health Association. pp. 269-273.

Arnold, A.P. and RA. Gorski. 1984. Gonadal steroid induction of structural sex
differences in the CNS. Ann. Rev. Neurgsci. 7:413-442.

Astroff, B., M. Romkes, S. Safe. 1989. Mechanism of action of 2,3,7,8-TCDD
and 6-methyl-1,3,8-trichlorodibenzofuran(MCDF) as antiestrogens in the female
rat. Qhemosghere 19:785-788.

Ausubel, F.M., R. Brent, R.E. Kingston, et al. (Eds.). 1994. Current Protocols
in Molecular Biology. John Wiley and Sons Publishing, New York. Chapters 1-
10.

Babin, P.J. and J-M. Vernier. 1989. Plasma lipoproteins in fish. J. Lipid Res.
30:467-489.

Ballschmiter, K., C. Rappe, H.R. Buser. 1989. Chemical properties, analytical
and environmental levels of PCBs, PCTs, and PCBs. In. Halogenated Biphenyls,
Terphenyls, Naphthalenes, Dibenzodioxins and Related Products. (R.D.
Kimbrough and A.A. Jensen, Eds.) Elsevier-North Holland. Amsterdam. pp. 47-
69.

114

115

Becker, K.L. 1990. Principles and Practice of Endocrinology and Metabolism.
Lippincott. Philadelphia, PA. p. 78-80, 970.

Benfey, T.J., E.M. Donaldson and T.G. Owen. 1989. An homologous
radioimmunoassay for Coho salmon (Oncorhynchus kisutch) vitellogenin, with
general applicability to other Pacific salmonids. Gen.Comg. Endogrinol. 75:78-
82.

Bennets, H.W., E.J. Underwood, F.L. Shier, 1946. A breeding problem of sheep
in the Southwest division of western Australia. J. Ag. West Australia 23:1 -1 2.

Bergeron, J.M., D. Crews, J.A. McLachlan. 1994. PCBs as environmental
estrogens: Turtle sex determination as a biomarker of environmental
contamination. Environ. Health Perspect. 102:780—781.

Bergman, A., E. Klasson-Wehler, H. Kuroki, and A. Nilsson. 1995. Synthesis
and mass spectrometry of methoxylated polychlorobiphenyls. Chemosphere
30:1921-1938.

Bergman, A., E. Klasson—Wehler, H. Kuroki. 1994. Selective retention of
hydroxylated PCB metabolites in blood. Environ. Health Perspect. 102:464-469.

Bergman, A., M. Athanasiadou, S. Bergek, K. Haraguchi, S. Jensen, and E.
KIasson-Wehler. 1992. PCB and PCB methyl sulphones in mink treated with
PCB and various PCB fractions. Ambio 21:570-576.

 

Bitman, J. and H.C. Cecil. 1970. Estrogenic activity of DDT analogs and
polychlorinated biphenyls. J. Ag. Food Chem. 18:1108-1112.

Bondy, K.L. and T.R. Zacharewski. 1993. lCl 164,384: A control for
investigating estrogen responsive genes. Nucleic Acids Re; 21:5277-5278.

Bradford, MM. 1976. A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of dye binding. Anal.
Biochem. 72:248-254.

Brotons, J.A., M.F. Olea-Serrano, M. Villalobos, V. Pedraza, and N. Olea. 1995.
Xenoestrogens released from lacquer coatins in food cans. Environ. Health
Perspect. 103: 608-612.

Brouwer, A. and K.J. VanderBerg. 1986. Binding of a metabolite of 3,4,3’,4'-
tetrachlorobiphenyl to transthyretin reduces serum vitamin A transport by
inhibiting the formation of the protein complex carrying both retinol and
thyroxin. Toxicol. Appl. Pharmacol. 85:301-312.

Brouwer, A., E. Klasson-Wehler, M. Boddam, D.C. Morse, and W.A. Traag.

116

1990. Competitive inhibition of thyroxin binding to transthyretin by mono-
hydroxy metabolites of 3,4,3’,4'-tetrachlorobiphenyl. Qhemoeehere 20:1257-
1 262.

Bruce, RD. and D.J. Versteeg. 1992. A statistical procedure for modelling
continuous toxicity data. Environ. Toxicol. Chem. 11:1485-1494.

Bulger, W.H., R.M. Muccitelli, and D. Kupfer. 1978b. Interactions of
methoxychlor, methoxychlor base-soluble contaminant, and 2,2-bis(p-
hydroxyphenyl)-1,1,1-trichloroethane with rat uterine estrogen receptor. J_.

Toxieol. Environ. Health 4:881-893.

Bulger, W.H., R.M. Muccitelli, and D. Kupfer. 1978a. Studies on the in vivo and
in vitro estrogenic activities of methoxychlor and its metabolites. Role of
hepatic mono-oxygenase in methoxylchlor activation. Biochem. Pharmaeol,
27:2417-2423.

Bull, J.J., W.H.N. Gutzke, D. Crews. 1988. Sex reversal by estradiol in three
reptilian orders. Gen. Como. Endocrinol. 70:425-428.

Carragher, J.F. and J.P. Sumpter. 1991. The mobilization of calcium from
calcified tissues of rainbow trout (Oncorhynchus mykiss) induced to synthesize
vitellogenin. Comp. Biochem. thsiol. 99A:169-172.

Chae, K., M.K. Gibson, and KS. Korach. 1991. Estrogen receptor
stereochemistry: Ligand binding orientation and influence on biological activity.
Mol. Pharmacol. 40:806-811.

Chang, C. and S. Liao. 1987. Topographic recognition of cyclic hydrocarbons
and related compounds by receptors for androgens, estrogens, and
glucocorticoids. J. Steroid Biochem. 27:123-131.

Chang, T-C., A.M. Nardulli, D. Lew, D.J. Shapiro. 1992. The role of estrogen
response elements in expression of the Xenoeus laevie vitellogenin Bl gene.
Mole. Endocrinol. 6:346-355.

Chen, T.T. 1983. Identification and characterization of estrogen-responsive
gene products in the liver of rainbow trout. Can. J. Biochem. Cell Biol. 61 :802-
810.

Colborne, T., D. Dumanoski, and J.P. Myers. 1996. Our Stolen Future: Are We
Threatening Our Fertility, Intelligence, and Survival?--A Scientific Detective
Story. Penguin Books, New York. 306 p.

Colborne, T. and C. Clement. 1992. Chemically-induced Alterations in Sexual
and Functional Development: The Wildlife/Human Connection. Princeton

117

Scientific Publishing, New Jersey. 403 p.

Copeland, P.A., J.P. Sumpter, T.K. Walker and M. Croft. 1986. Vitellogenin
levels in male and female rainbow trout (Se/mo gairdneri Richardson) at various
stages of the reproductive cycle. Comp. Biochem. Physiol. 838:487-493.

Cullen, K.E., M.P. Kladde, and M.A. Seyfred. 1993. Interaction between
transcription regulatory regions of prolactin chromatin. Scienee 261 :203-206.

Darnerud, P.O., I. Brandt, E. Klasson-Wehler, A. Bergman, R. d’Argy, L.
Dencker, and GO. Spoerber. 1986. 3,3',4,4'-tetrachloro-[“Cl-biphenyl in
pregnant mice: Enrichment of phenol and methyl sulphone metabolites in late
gestational fetuses. Xenobiotica 16:295-306.

Darnerud, P.O., E. Klasson-Wehler, D. Morse, and A. Brouwer. 1996. Binding
of the PCB congener 3,3',4,4’-tetrachlorobiphenyl (CB-77) to transthyretin and
effects on fetal thyroid hormone levels in mice. Toxicol. 106:105-114.

DeVito, M.J., T. Thomas, E. Martin, T.H. Umbreit, M.H. Gallo. 1992.
Antiestrogenic action of 2,3,7,8-tetrachlorodibenzodioxin: Tissue-specific
regulation of estrogen receptor in CD1 mice. Tox. Aeel. Pharm. 113:284-292.

deVIaming, V., R. Fitzgerald, G. Delakuntz, J.J. Cech, K. Selman, M. Barkley.
1 984. Dynamics of oocyte development and related changes in serum estradiol-
178, yolk precursor, and lipid levels in the teleostean fish, Leetoeeges armatue.

Come. Biochem. Physiol. 77A:599-610.

DeWailly, E., A. Nantel, J. Weber, F. Meyer. 1989. High levels of PCBs in
breast milk of Inuit women from arctic Quebec. Bull. Environ. Contam. Toxieel.
43:641-646.

Dibb, S. 1995. Swimming in a sea of oestrogens, chemical hormone disrupters.
The Ecologist 25:27-31.

Dieringer, C.S., C.A. Lamartiniere, C.M. Schiller and G.W. Lucier. 1979. Altered
ontogeny of hepatic steroid-metabolizing enzymes by pure polychlorinated
biphenyl congeners. Biochem. Pharmacol. 28:2511-2514.

Dodds, EC. and W. Lawson. 1938. Molecular structure in relations to
oestrogenic activity. Compounds without a phenanthrene nucleus. Pree. Reyel

Sec. London B. Biology 125:222-232.

Donaldson, E.M. 1990. Reproductive indices as measures of the effects of
environmental stressors in fish. Amer. Fish. Soc. Syme. 8:109-122.

Donaldson, E.M. and G.A. Hunter. 1983. Induced final maturation, ovulation

118

and spermiation in cultured fish. In: Fish physiology. (W.S. Hoar, D.J. Randall,
and E.M. Donaldson. Eds.) Academic Press, New York. Volume 9, Part 8, pp.
351-403.

Donaldson, E.M. and E. Scherer. 1983. Methods to test and assess effects of
chemical on reproduction in fish. In: Methods for assessing the effects of
chemicals on reproductive function (V.B. Vouk and P.J. Sheehan, Eds.) Wiley,
Sussex, England. pp. 365-404.

Eckert, R.L. and BS. Katzenellenbogen. 1982. Effects of estrogens and
antiestrogens on estrogen receptor dynamics and the induction of progesterone
receptor in MCF-7 human breast cancer cells. Cancer Res. 42:139-144.

Ecobichon, D.J. and 0.0. Mackenzie. 1974. The utertrophic activity of
commercial and isomerically pure chlorobiphenyls in the rat. Ree. Cemm. Chem.
Path. Pharm. 9:85-95.

El-Fouly, M.H., C. Richter, J.P. Giesy and MS. Denison. 1994. Production of
a novel recombinant cell line for use as a bioassay system for detection of
2,3,7,8-tetrachIorodibenzo-p-dioxin-Iike chemicals. Environ. Texicol. Chem.
13:1581-1588.

Emmersen, B.K. and I.M. Petersen. 1976. Natural occurrence and experimental
induction by estradiol-178 of a lipophosprotein (vitellogenin) in flounder
(Platichthys flesus L. Comp. Biochem. Physio); 54B:443-446.

Eroschenko, V.P. 1981 . Estrogenic activity of the insecticide chlordecone in the
reproductive tract of birds and mammals. J. Tox. Environ. Health 8:731-742.

Eroschenko, V.P. and RD. Palmiter. 1980. Estrogenicity of kepone in birds and
mammals. In: J.A. McLachlan (Ed.) Estrogens in the Environment. Elsevier-North
Holland, Amsterdam. pp. 305-325.

Evans, RM. 1988. The steroid and thyroid hormone receptor superfamily.
Science 240: 889-895.

Farooqi, Z.H. and H.Y. AbouI-enein. 1992. The influence of different
configurations of cyclohexyl moiety on the conformational flexibility of
cyclohexylaminoglutethimide, a potent aromatase inhibitor. Toxicol. Environ.
Chem. 36:189-193.

 

Fawell, S.E., R. White, S. Hoare, M Sydenham, M. Page, M.G. Parker, 1990.
Inhibition of estrogen receptor-DNA binding by the "pure" antiestrogen lCl
164,384 appears to be mediated by impaired receptor dimerization. W
Acad. Sci. USA 87:6883-6887.

119

Fawell, S.E., J.A. Lees, R. White, M.G. Parker. 1990. Characterization and
colocalization of steroid binding and dimerization activities in the mouse
estrogen receptor. Ce!) 60:953-962.

Fernandez, P., R. Burghardt, R. Smith, K. Nodland, S. Safe. 1994. High passage
T47D human breast cancer cell: altered endocrine and 2,3,7,8-
tetrachlorodibenzo-p-dioxin responsiveness. Eur. J. Pharmacol.-Environ. Tox.
Pharmacol. Section 270:53-65.

Fry, D.M. and C.K. Toone. 1981. DDT-induced feminization of gull embryos.
Science 213:922-924.

 

Gellert, R.J. 1978. Uterotrophic activity of polychlorinated biphenyls (PCBs) and

induction of precocious reproductive aging in neonatally treated female rats.
Environ. Res. 16:123-130.

Gellert, R.J. C. Wilson. 1979. Reproductive function in rats exposed prenatally
to pesticides and polychlorinated biphenyls (PCBs). Environ. Res. 18:437-443.

Gellert, R.J. 1978. Kepone, mirex, dieldrin, and aldrin: Estrogenic activity and
the induction of persistent vaginal estrus and anovulation in rats following
neonatal treatment. Environ. Re; 16:131-138.

Giorgi, F., A. Gobbetti and A. Polzonetti Magni. 1982. Variations in the
vitellogenin titre during the reproductive cycle of Rana esculenta L. Come.
Biochem.Physiol 72B:501-506.

Gordon, MS. and A.C. Notides. 1986. Computer modeling of estradiol
interactions with the estrogen receptor. J. Steroid Biochem. 25:177-181.

Goto, M., K. Sugiura, M. Hattori, T. Miyagawa, M. Okamura. 1974. Metabolism
of 2,3-dichlorobiphenyl-“C and 2,4,6-trichlorobiphenyl-“C in the rat.
Chemosehere 227-232.

Goto, M., M. Hattori, K. Sugiura. 1975. Metabolism of pentachloro- and
hexachloro biphenyls in the rat. Chemosehere 177.

Goto, M., K. Sugiura, M. Hattori, T. Miyagawa, M. Okamura. 1974. Metabolism
of 2,3, 5,6-tetrachlorobiphenyl-‘4C and 2,3,4, 5,6-pentachlorobiphenyl-"C in the
rat. Chemosehere 233-238.

Hammond, B., 8.8. Katzenellenbogen, N. Krauthammer and J. McConnell.
1979. Estrogenic activity of the insecticide chlordecone (Kepone) and
interaction with uterine estrogen receptors. Proc. Natl. Aca . Sci. U A
76:6641-6645.

120

Haraguchi, K., M. Athanasidou, A. Bergman, L. Hovander, and S. Jensen. 1 992.
PCB and PCB methyl sulphones in selected groups of seals from the Swedish
coastwaters. Ambio 21 :546-549.

 

Heath, R.G., J.W. Spann, J.F. Kreitzer. 1969. Marked DDE impairment of
mallard reproduction in controlled studies. Nature 224:47-48.

 

Herman, R.L. and H.L. Kincaid. 1988. Pathological effects of orally administered
estradiol to rainbow trout. Aguaculture 72:165-172.

Ho, S-M. 1991 . Vitellogenesis. In: Vertebrate endocrinology: Fundamentals and
biomedical implications. (P.K.T. Pang and MP. Schreibman, Eds.) Academic
Press, New York. Volume 4, Part A, pp. 91-126.

Hwang, K-J., K.E. Carlson, G.M. Anstead, and J.A. Katzenellenbogen. 1992.
Donor acceptor tetrahydochrysenes, inherently fluorescent, high-affinity ligands
for the estrogen receptor: Binding and fluorescence characteristics and
fluorometric assay of receptor. Biochem. 31 :1 1536-1 1545.

Jansson, B., S. Jensen, M. Olsson, L. Renberg, G. Sundstrbm, and R. Voy.
1975. Identification by GC-MS of phenolic metabolites of PCB and p,p’DDE
isolated from Baltic guillemot and seal. Ambio 4:93-97.

Jansson, B., R. Andersson, L. Asplund, K. Litzen, K. Nylund, U. Sellstrom, U-B.
Uvemo, C. Wahlberg, U. Wideqvist, T. Odsjb, and M. Olsson. 1993.
Chlorinated and brominated persistent organic compounds in biological samples
from the environment. Environ. Toxicol. Chem. 12:1 163-1 174.

Jarman, W.M., M. Simon, R.J. Norstrom, S.A. Burns, C.A. Bacon, B.R.T.
Simonelt and R.W. Risebrough. 1992. Global distribution of trisl4-
chlorophenyl)methano| in high trophic level birds and mammals. Enviren. Sci.
Technol. 26:1770-1774.

Jensen, 8., B. Jansson, M. Olsson. 1979. Number and identity of
anthropogenic substances known to be present in baltic seals and their possible
effects on reproduction. Ann. N. Y. Acad. Sci. pp. 436-448.

Jobling, S., T. Reynolds, R. White, M.B. Parker, and J.P. Sumpter. 1995. A
variety of environmentally persistent chemcals, including some phthalate
plasticizers, are weakly estrogenic. Environ. Health Perspect. 103:582-587.

Johnson, A.L. 1986. Reproduction in the female. ln: P.D. Starkie (Ed.) Avian
Physiology. 4th Edition. Springer-Verlag, New York. Chapter 18.

Johnson, L., E. Casillas, S. Sol, T. Collier, J. Stein and U. Varanasi. 1993.
Contaminant effects on reproductive success in selected benthic fish. Mar. Env.

121
es. 35:165-170.

Jordan, V.C., R. Kock, M.E. Lieberman. 1986. Structure-activity relationships
of nonsteroidal estrogens and anti-estrogens. In: Estrogen/Anti-estrogen action

and breast cancer therapy (V.C. Jordan, Ed.) University of Wisconsin Press,
Madison, WI pp. 19-41.

Katzenellenbogen, B.S., J.A. Katzenellenbogen, and D. Mordecai. 1979.
Zearalenones: Characterization of the estrogenic potencies and receptor
interactions of a series of fungal B-resorcylic acid lactones. Endocrinol. 105:33-
40.

Keith, J.A. 1966. Reproduction in a population of herring gulls (Lerus
argentatus) contaminated by DOT. J. Aeel. Ecol. Sueel. 3:57-70.

Kelce, W.R., C.R. Stone, S.C. Lewis, L.W. Gray, J.A. Kemppainen and E.M.
Wilson. 1995. Persistent DDT metabolite, p,p’-DDE, is a potent androgen
receptor antagonist. Nature 375:581-585.

 

Klasson-Wehler, E., A. Bergman, l. Brandt, P.O. Darnerud, C.A. Wachtmeister.
1989. 3,3’,4,4'-tetrach|orobiphenyl: Excretion and tissue retention of
hydroxylated metabolites in the mouse. Drug Metab. Diseos. 17:441-448.

Klasson-Wehler, E., Lindberg, C-J. ansson, A. Bergman. 1993. Tissue
retention and metabolism of 2,3,4,3’,4'-pentachloro biphenyl in mink and
mouse. Chemosphere 27:2397-2412.

Klein-Hitpass, L., M. Shorpp. U. Wagner, G.U. Ryffel. 1986. An estrogen-
responsive element derived from the Xenoeus vitellogenin A2 gene functions
in transfected human cell. _C_3e_ll 46:1053-1061.

Klinge, C.M., R.A. Bambara, and R. Hilf. 1992. What differentiates
antiestrogen-liganded vs estradiol-liganded estrogen receptor action? anol.
Res. 4:137-144.

Knobil, E. 1980. The neuroendocrine control of the menstrual cycle. Rec. Preg.
Hormone Res. 36:53-88.

Koistinen, J., T. Nevalainen and J. Tarhanen. 1992. Identification and level
estimation of aromatic coeluates of polychlorinated dibenzo-p-dioxins and
dibenzofurans in pulp mill products and wastes. Environ. Sei. Teeheel.
26:2499-2507.

Korach, K.S., L.A. Levy, P.J. Server. 1987. Estrogen receptor stereochemistry:
receptor binding and hormonal responses. J. Steroid Biochem. 27:281-290.

122

Korach, K.S., L.A. Levy, and P.J. Server. 1985. Stereochemical analysis of
stilbene estrogens: Receptor binding and hormone responsiveness. ln.
Estrogens in the Environment (J.A. McLachlan, Ed.) Elsevier Scientific Pub.
Amsterdam. pp. 43-65.

Korach, K.S., P. Sarver, K. Chae, J.A. McLachlan, J.D. McKinney. 1988.
Estrogen receptor binding activity of polychlorinated hydroxybiphenyls:
Conformationally restricted structural probes. Molec. Pheremaeel. 33: 1 20-126.

Krishnan, V. and 8. Safe. 1993. Polychlorinated biphenyls (PCBs) dibenzo-p-
dioxins (PCDDs) and dibenzofurans (PCDFs) as antiestrogens in MCF-7 human
breast cancer cell: Quantitative structure-activity relationships. Tox. Aeel.

Pharm. 120:55-61.

Krishnan, A.V., P. Stathis, S.F. Permuth, L. Tokes, D. Feldman. 1993.
Bisphenol-A: An estrogenic substance is released from polycarbonate flasks
during autoclaving. Endocrinol. 132:2279-2286.

Kumar, V., S. Green, G. Stark, M. Berry, J-R, Jin, P. Chambon. 1987.
Functional domains of the human estrogen receptor. Qe_ll 51 :941-951 .

Kumar, V., and P. Chambon. 1988. The estrogen receptor binds tightly to its
responsive element as a ligand induced homodimer. Qefl 55:145-156.

Kupfer, 0., W.H. Bulger. 1980. Estrogenic properties of DDT and its analogs.
In: J.A. McLachlan (Ed.) Estrogens in the Environment. Elsevier-North Holland,
Amsterdam. pp. 239-261.

Kupfer, D. and W.H. Bulger. 1976. Studies on the mechanism of estrogenic
actions of a,p ’-DDT: Interactions with the estrogen receptor. Pesticide Bioehem.
Physiol. 6:561-570.

Kupfer, D. 1975. Effects of pesticides and related compounds on steroid
metabolism and function. CRC Critical Reviews in Toxicology pp. 83-124.

Lans, M.C., Klasson-Wehler, E., M. Willemsen, E. Meussen, S. Safe, and A.
Brouwer. 1993. Structure-dependent competitive interaction of hydroxy-
polychlorobiphenyls, dibenzo-p-dioxins, and dibenzofurans with human
transthyretin. Chem.-Biol. Interact. 88:7-21.

Levier, RR. and M.E. Jankowiak. 1972. The hormonal and antifertility activity
of 2,6-cis-diphenylhexamethylcyclotetrasiloxane in the female rat. Biol. Reere,
7:260-266.

Lieberman, M.E., W.V. Welshone, J. Gorski. 1986. Nuclear localization of
unoccupied estrogen receptors. In: V.C. Jordan (Ed.) Estrogen/antiestrogen

123

action and breast cancer therapy. U.Wisconsin Press, Madison. Chapter 3, pp.
43-54.

Liu, H., T.R. Narasimham, and S. Safe. 1992. 2,3,7,8-tetrachlorodibenzo-p-
dioxin as an anti-proliferative agent in human breast cancer cells: Effects on
insulin-induced cell proliferation. Chemosehere 25:955-958.

Lundh, T.J.O., H.l. Pettersson, K.A. Martinsson. 1990. Comparative levels of

free and conjugated plant estrogens in blood plasma of sheep and cattle fed
estrogenic silage. J. Agric. Food Chem. 38:1530-1534.

Mankowitz, L. and J. Rydstrom. 1982. Interactions of polycyclic hydrocarbons
with the estradiol receptor in rat adrenal. In: E. Hietanen, M. Laitinen, O.
Hanninen. (Eds.) Cytochrome P450-Biochemistry, biophysics, and
environmental implications. Elsevier Biomedical Press. pp. 755-758.

Martin, J.B. and D.M. Doty. 1949. Determination of inorganic phosphate:
Modification fo isobutyl alcohol procedure. Anal. Chem. 21 :965-967.

Matty, A.J. Fish Endocrinology. Billing and Sons, Worcester. pp.157-161.

McCague, R., M. Jarman, O-T. Leung, A.B. Foster, G. Leclercq, and S.
Stoessel. 1988. Non-isomerisable antiestrogens related to tamoxifen. J. Steroid
Biochem. 31 (4B):545-547.

McFarland, V.A. and J.U. Clark. 1989. Environmental occurrence, abundance,
and potential toxicity fo polychlorinated biphenyl congeners: considerations for
a congener specific analysis. Environ. Health Perseect. 81:225-239.

McLachlan, J.A. 1993. Functional toxicology: A new approach to detect
biologically active xenobiotics. Environ. Health Perseect. 101:386-387.

McOmie, J.F., M.L. Watts, and DE West. 1968. Demethylation of aryl methyl
ethers by boron tribromide. Tetrahedron 24:22889-2292.

Migliaccio, A., G. Castoria, A. DeFalco, M. DiDomenico, M. Galdiero, E. Nola,
P. Chambon, F. Aurichio. 1991. In vitro phosphorylation and hormone binding
activation of the synthetic wild type human estradiol receptor. ,1. Steroid
Biochem. Molee. Biol. 38: 407-413.

Miles-Richardson, S. 1996. Reproductive toxicosis of WIS-estradiol in fathead
minnow, Pimeohales promelae. PhD. Dissertation. Michigan State University,
East Lansing.

Miquel, J-F, J. Gilbert. 1988. A chemical classification of nonsteroidal
antagonists of sex-steroid hormone action. J. Steroid Biochem. 31(48):525-

1 24
544.

Morse, D.C., E. Klasson-Wehler, M. van der Pas, A.T.H.J. de Bie, P.J. van
Bladeren, and A. Brouwer. 1995. Metabolism and biochemical effects of

3,3’,4,4’-tetrachlorobiphenyl in pregnant and fetal rats. Chem.-Biol. Interact.
95:41-56.

Mueller, G.C. and U-H. Kim. 1978. Displacement of estradiol from estrogen
receptors by simple alkyl phenols. Endocrinoloqv 102:1429-1435.

Murdoch, F.E., D.A. Meier, J.D. Furlow, K.A.A. Grunwald, and J. Gorski. 1990.
Estrogen receptor binding to a DNA response element in vitro is not dependent
upon estradiol. Biochem. 29:8377-8385.

Narasimham, T.R., H.L. Kim and S. Safe. 1991. Effects of hydroxylated
polychlorinated biphenyls on mouse liver mitochondrial oxidative
phosphorylation. J. Biochem. Toxicol. 6:229-236.

Nelson, J.A. 1974. Effects of dichlorodiphenyltrichloroethane (DDT) analogs
and polychlorinated biphenyl (PCB) mixtures on 1 78-[3Hl-estradiol binding to the
rat uterine receptor. Biochem. Pharmacol. 23:447-451.

Nelson, J.A., R.F. Struck, and R. James. 1978. Estrogenic activities of
chlorinated hydrocarbons. J. Toxicol. Environ. Health 4:325-339.

Nelson, J.A. 1974. Effects of dichlorodiphenyltrichloroethane (DDT) analogs
and polychlorinated biphenyl (PCB) mixtures on 1713-[3Hl-estradiol binding to the
rat uterine receptor. Biochem. Pharmacol. 23:447-451.

Nimrod, A.C. and W.H. Benson. 1996. Estrogenic responses to xenobiotics in
channel catfish (lctalurus punctatus). Mar. Environ. Res. in press.

Paasivirta, J., K. Heinola, T. Humpi, A. Karjalainen, J. Knuutinen, R. Paukku,
T. Piilola, K. Surma-Aho, J. Tarhanen, L. Welling and H. Vihonen. 1985.
Polychlorinated phenols, guaicols, and catechols in the environment.
Chemosehere 14:469-491 .

Parker, 0.8 and BA. McKeown. 1987. Effects of pH and/or calcium-enriched
freshwater on plasma levels of vitellogenin and Ca2+ and on bone calcium
content during exogenous vitellogenesis in rainbow trout (Sa/mo gairdnerr).

Come. Biochem. Physiol. 87A:267-273.

Pasqualini, C., A. Sarrieau, M. Dussaillant, M. Corbani, F. Bojda-Diolez, W.
Rostene, and B. Kerdelhue. 1990. Estrogen-like effects of 7,12-
dimethylbenz(a)anthracene on the female rat hypothalamo-pituitary axis. J.-
Steroid Biochem. 36:485-491.

125

Pasqualini, C., A. Sarrieau, M. Dussaillant, M. Corbani, F. Bojda-Dioley, W.
Rostene, B. Kerdelhue. 1990. Estrogen-like effects of 7,12-
dimethylbenzanthracene on the female rat hypothalamo-pituitary axis. J. Steroid
Bioehem. 36:485-491.

Pelissero, C., B. Bennetau, P. Babin, F. LeMann, J. Dunogues. 1991. The
estrogenci activity of certain phytoestrogens in the Siberian Sturgeon Acieenser
baeri. J. Steroid Biochem. Molec. Biol. 38:293-299.

Pilat, M.J., M.S. Hafner, L.G. Kral, S.C. Brooks. 1993. Differential induction of
p82 and cathepsin D mRNA’s by structurally altered estrogens. Biechem.
32:7009-7015.

Pons, M., D. Gagne, J.C. Nicolas, M. Mehtali. 1990. A new cellular model of
response to estrogens: A bioluminescent test to characterize (Anti)estrogen
molecules. Biotechniguee 9:450-456.

Purdom, C.E., P.A. Hardiman, V.J. Bye, N.C. Eno, C.R. Tyler and J.P. Sumpter.
1994. Estrogenic effects of effluents from sewage treatment works. Chem.
Ecol. 8:275-285.

Redding, J.M. and R. Patifio. 1993. Reproductive physiology. In: D.H. Evans
(Ed.) The Physiology of Fishes. CRC Marine Science Series, Boca Raton, FL.
Chapter 16, pp. 503-534.

Robison, A.K. and GM. Stancel. 1982. The estrogenic activity of DDT:
Correlation of estrogenic effect with nuclear level of estrogen receptor. Qte
_S_ei_._ 31: 2479-2484.

Routledge, E.J. and J.P. Sumpter. 1996. Estrogenic activity of surfactants and
some of their degradation products assessed using a recombinant yeast screen.
Environ. Toxicol. Chem. 15 (in press).

Safe, SH. 1994. Dietary and environmental estrogens and antiestrogens and
their possible role in human disease. Env. Sci. Pollut. Ree. 1:29-33.

Safe, S.H., L. Biegel, M. Harris, T.R. Narasimhan. H. Lui, P. Fernandez. 1992.
Inhibition of estrogen- and growth factor-induced proliferation of human brease
cancer cell lines by TCDD: Characterization and mechanistic studies.
Chemosehere 25:83-86.

Safe, SH. 1995. Do environmental estrogens play a role in development of
brease cancer in women and male reproductive problems? Human gee). Rjek
Assesement 1:17-23.

Safe, S.H., L. Safe, M. Mullin. 1985. Polychlorinated biphenyls (PCBs)-congener

126

specific analysis of commercial mixtures and a human milk extract. J- Agrie.

Foee Qhem. 33:24-29.

Safe. S.H., M. Harris, L. Biegel, T. Zacharewski. 1991. Mechanism of action of
TCDD as an antiestrogen in transformed human breast cancer and rodent cell
lines. Banbury Report 35: Biological basis for risk assessment of dioxins and
related compounds. Cold Spring Harbor Laboratory Press. pp. 367-377.

Sanderson, J.T., J.M.M.J.G. Aarts, A. Brouwer, K.L. Froese, M.S. Denison, and
J.P. Giesy. 1996. Comparison of Ah receptor-mediated luciferase and
ethoxyresorufin-O-deethylase induction in H4IIE cells: Implications for their use
as bioanalytical tools for the detection of polyhalogenated aromatic
hydrocarbons. Toxicol. Appl. Pharm. 137: in press.

SAS Institute. 1988. SAS Procedures Guide Release 6.03 Edition. Cary NC:
SAS Institute Inc. 441 pp.

Schmidt, W.A. 1994. Hormone Copycats. National Wildlife Federation Report.
Ann Arbor, MI. 59 pp.

Schmitt, C.J., J.L. Zajicek, and P.H. Peterman. 1990. National contaminant
biomonitoring programzresidues of organochlorine chemicals in US freshwater
fish 1976-1984. Arch. Environ. Contam. Toxicol. 19:748-781.

Scott, A.P. and J.P. Sumpter. 1983. A comparison of the female reproductive
cycles of autumn-spawning and winter-spawning strains of rainbow trout
(Salmo gairdneri Richardson). Gen. Come. Endocrinol. 52:79-85.

Selcer, K.W. and W.W. Leavitt. 1991. Estrogens and progestins. In: P.K.T.
Pang and M. Schreibman (Eds.) Vertebrate Endocrinology: Fundamentals and
Biomedical Implications. Academic Press, San Diego. Volume 4 Part 8, pp. 67-
1 14.

Setchell, K.D.R., S.J. Gosselin, M.B. Welsh, J.O. Johnston, W.F. Balisteri, L.W.
Kramer, B.L. Dresser, M.J. Tarr. 1987. Dietary estrogens-A probable cause of
infertility and liver disease in captive cheetahs. Gastroenterolegy 93:225-233.

Shutt, D.A., A.W.H. Braden. 1968. The significance of equol in relation to the
oestrogenic responses in sheep ingesting clover with a high formononetin

content. Austrelian J. Ag. Res. 19:545-553.

Shutze, N., G. Volmer, l. Tiemann, M. Geiger, and R. Knuppen. 1993.
Catecholestrogens are MCF-7 cell estrogen receptor agonists. J. Stereie

Bioehem. Melec. Biol. 46:781-789.

Silversand, C. and C. Haux. 1989. Isolation of turbot (Scophthalmus maximus)

127

vitellogenin by high-performance anion-exchange chromatography. zl_.

Qhrematog. 478:387-397.

Singh, H. 1989. Interaction of xenobiotics with reproductive endocrine
functions in a protogynous teleost, Monoeterus albue. Mar. Env. Res. 28:285-
289.

So, Y.P., D.R. Idler and SJ. Hwang. 1985. Plasma vitellogenin in landlocked
Atlantic salmon (Salmo salar Ouananiche): Isolation, homologous
radioimmunoassay and immunological cross-reactivity with vitellogenin from
other teleosts. Come. Biochem. Physiol. 818:63-71.

Solmssen, U.V. 1945. Synthetic estrogens and the relation between their
structure and their activity. Chem. Review 37:481-598.

Soto, A.M., H. Justicia, J.W. Wray, C. Sonnenschein. 1991. p-Nonyl phenol:
An estrogenic xenobiotic released from "modified" polystyrene. Environ. Health.
Perspect. 92:167-173.

Soto, A.M., T.M. Lin, H. Justicia, R.M. Silvia, C. Sonnennschein. 1992. An "in
culture" bioassay to assess the estrogenicity of xenobiotics (E-screen), In:
Chemically induced alterations in sexual and functional development: The
wildlife/human connection. (T. Colborn, C. Clement, Eds) Princeton Scientific
Publishing, Princeton, NJ. pp. 295-309.

Southern, P.J. and P. Berg. 1982. Transformation of mammalian cells to
antibiotic resistance with a bacterial gene under control of the SV40 early
region promoter. J. Mol. Aeel. Gen. 1:327-341.

Spink, D.C., D.W. Lincoln II, J.A. Johnson, H.W. Dickerman, J.F. Gierthy.
1992. Stimulation of WIS-estradiol metabolism in MCF-7 breast cancer cells by
2,3,7,8-tetrachlorodibenzo-p-dioxin. Chemosehere 25:87-90.

Stacey, N., P.W. Sorensen, J.G. Dulka, G.J. van der Kraak and T.J. Here.
1987. Teleost sex pheromones: recent studies on identity and function. In:
Reproductive physiology of fish. (D.R. Idler, L.W. Crim, and J.M. Walsh, Eds.)
St. John’s, Newfoundland. pp.150-153.

Standeven, A.M., D.G. Blazer Ill and T.L. Goldsworthy. 1994. Investigation of

antiestrogenic properties of unleaded gasoline in female mice. Toxicol. Applied
Pharmacol. 127:233-240.

Stob, M., R.S. Baldwin, J. Tuite, F.N. Andrews, K.G. Gillette. 1962. Isolation
of an anabolic uterotrophic compound from corn infected with Gieerelle zeae.
Natere 196:1318

128

Sundstrbm, G., O. Hutzinger, S. Safe. 1976. The metabolism of
chlorobiphenyls-A review. Chemosehere 5:267-298.

Teutsch, G., T. Ojasoo, and J.P. Raynaud. 1988. 118-substituted steroids, an
original pathway to antihormones. J. Steroid Biochem. 31(48):549-565.

Thomas, P. and J. Smith. 1993. Binding of xenobiotics to the estrogen receptor
of spotted seatrout: A screening assay for potential estrogenic effects. Marine
Environ. Res; 35: 147-151.

Tillitt, D.E., T.J. Kubiak, G.T. Ankley, J.P. Giesy. 1993. Dioxin-Iike toxic
potency in Forster’s tern eggs from Green Bay, Lake Michigan, North America.
Chemosehere 26:2079-2084.

Tyler, C.R., J.P. Sumpter, and N.R. Bromage. 1988a. In vivo ovarian uptake
and processing of vitellogenin in the rainbow trout, Salmo gairdneri. J. Exp.
Zool. 246:171-179.

Tyler, C.R., J.P. Sumpter, and N.R. Bromage. 1988b. Selectivity of protein
sequestration by vitellogenic oocytes of the rainbow trout, Salmo gairdneri. J;
Exe. Zool. 248:199-206.

U.K. Directorate of Fisheries Research. 1 992. Effects of Trace Organics on Fish.
Report No. FR/D 0008, Ministry of Agriculture, Fisheries and Food. London.

van Aswegen, C.H., R.H. Purdy and J.L. Wittliff. 1989. Binding of 2-
hydroxyestradiol and 4-hydroxyestradiol to estrogen receptors from human
breast cancers. J. Steroid. Biochem. 32:485-492.

VanderKuur, J.A., T. Wiese, and SC. Brooks. 1993a. Influence of estrogen
structure on nuclear binding and progesterone receptor induction by the
receptor complex. Biochem. 32:7002-7008.

VanderKuur, J.A., M.S. Hafner, J.K. Christman and SC. Brooks. 1993b.
Effects of Estradiol-1 713. analogues on activation of estrogen response element
regulated chlroamphenicol acetyltransferase expression. Biochem. 32:7016-
7021.

Walker, P., J-E. Germond, M. Brown-Luedi, F. Givel, W. Wahli. 1984. Sequence
homologies in the region preceding the transcription initiation site of the liver
estrogen-responsive vitellogenin and apo-VLDLII genes. Nee. Aeids Ree,
12:8611-8626.

Wallace, R.A. and D.W. Jared. 1968. Studies on amphibian yolk. Vll. Serum
phosphoprotein synthesis by vitellogenic females and estrogen treated males
of Xeneeue laevis. Can. J. Biochem. 46:953-959.

129

Waller, C.L., D.L. Minor, and JD. McKinney. 1995. Using three-dimensional
quantitative structure-activity relationships to examine estrogen recptor binding

affinities of polychlorinated hydroxybiphenyls. Environ. Heelth Pereeect.
103:702-707.

Welch, R.M., W. Levin, and A.H. Conney. 1969. Estrogenic action of DDT and
its analogs. Toxicol. Apol. Pharmacol. 14:358-367.

White, R., S. Jobling, S.A. Hoare, J.P. Sumpter and M.G. Parker. 1994.
Environmentally persistent alkylphenolic compounds are estrogenic. Endocrinel.
135:175-182.

Whitten, P.L., E. Russell, and F. Naftolin. 1992. Effects of a normal, human-
concentration, phytoestrogen diet on rat uterine growth. Steroids 57:98-106.

Wiegand, MD. and RE. Peter. 1980. Effects of sex steroids on plasma lipids
in the goldfish, Carassius auratus. Can. J. Zool. 58:967-972.

Witschi, E. 1971. Mechanism of sexual differentiation. In: M. Hamburgh and
E.J.W. Barrington (Eds.) Hormones in Development. Appleton-Century-Crofts,
NY pp. 601-618.

Wolff, M.S., P.G. Toniolo, E.W. Lee, M. Rivera, and N. Dubin. 1993. Blood
levels of organochlorine residues and risk of breast cancer. J. Natl. Ceneer. Inst.
85:648-652.

Zacharewski, T., M. Harris, and S. Safe. 1991. Evidence for the mechanism of
action of 2,3,7,8-tetrachlorodibenzo-p-dioxin mediated decrease of nuclear
estrogen receptor levels in wild-type and mutant mouse Hepa 1c1c7 cells.
Biochem. Pharmacol. 41 :1931-1939.

Zubay, G. 1988. Biochemistry. 2nd Edition. MacMillan and Co. New York. p.
168-169.

"Illllllllllllllllllllllllls