DETERMINING THE EFFECTS OF MULTIPLE STRESSORS ON FATHEAD
MINNOWS (PIMEPHALES PROMELAS) USING A FLOW-THROUGH DILUTER
SYSTEM
By
Brandon M. Armstrong

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTERS OF SCIENCE
Fisheries and Wildlife – Environmental Toxicology
2011

ABSTRACT
DETERMINING THE EFFECTS OF MULTIPLE STRESSORS ON FATHEAD
MINNOWS (PIMEPHALES PROMELAS) USING A FLOW-THROUGH DILUTER
SYSTEM
By
Brandon M. Armstrong

Aquatic organisms are exposed to a multitude of contaminants and to fully understand the
impact of multiple stressors on fish populations, we must first understand the mechanism
of action for each toxicant and how the combined effects manifest at the level of the
individual. 17α-ethinylestradiol has been known to cause adverse reproductive effects
including reduced fecundity and fertility, intersex and sex change in fish by mimicking
naturally produced estrogen at low concentrations. Ammonia has been known to cause
adverse reproductive and mortality effects in individual fish through effects or damage to
the central nervous system. A flow-through diluter system was used to test the individual
effects of environmentally relevant concentrations of these two contaminants as well as
their mixture on fathead minnow reproduction. Significant findings from this study
suggest that U.S. EPA water quality criterion for ammonia in waters where nonsalmonids are present (8.4mg/L) is above the lowest observable effect concentration for
fathead minnow reproduction. The lowest un-ionized ammonia concentration (5mg/L)
tested during this study resulted in a 29% decrease in cumulative fecundity. The mixture
of both 17α-ethinylestradiol and ammonia at their respective environmentally relevant
concentration resulted in increased mortality, but did not to show adverse effects on
reproduction. This study demonstrated the need for toxicity testing with multiple stressor
scenarios and the importance of re-evaluating current U.S. EPA water quality criteria.

Copyright by
BRANDON MICHAEL ARMSTRONG
2011

To my wife, Kristen, without whom I would most certainly be lost.

iv

ACKNOWLEDGEMENTS

1

2

Jim Lazorchak and Cheryl Murphy provided support to the development and
completion of this project between the U.S. EPA and Michigan State University. Cheryl
also provided assistance with data analysis and manuscript preparation and served as my
major adviser. Jim also provided valuable background and support to the overall
construction of this project and served as one of my committee members. My other
2

2

committee members, Mary Bremigan and James Sikarskie provided comments on the
3

3

design and analysis of this project. Mark Smith and Herman Haring brought valuable
1

insight to the construction and analysis of the diluter system. Robert Flick , David
1

1

1

1

1

Bencic , Adam Biales , Denise Gordon , Paul Wernsing and Suzanne Jackson assisted
1

1

with necropsy procedures. Michael Crane and Mitch Kostich provided data analysis of
3

3

3

the 17α-ethinylestradiol study. Katie Hammer , Melissa Wratschko , and Paul Weaver
4

aided in necropsy and weekend duties. Kathleen Jensen provided ELISA analysis of
5

plasma vitellogenin. Barry Weichman provided assistance with vitellogenin analysis of
5

the livers collected from each of the two studies. Ervin O’Bryan provided assistance
with the EE2 analytical chemistry analysis during the mixture study.
1
2
3
4
5

U.S. EPA in Cincinnati, OH
Michigan State University in East Lansing, MI
The McConnell Group c/o U.S. EPA in Cincinnati, OH
U.S. EPA in Duluth, MN
Dynamac Corporation c/o U.S. EPA in Cincinnati, OH
v

TABLE OF CONTENTS

List of Tables

ix

List of Figures

x

Chapter 1:
Introduction:
Multiple Stressors in the environment
Two contaminants in the environment: 17α-ethinylestradiol
and ammonia
Study Hypotheses
Chapter 2: Determining the NOEC of NH3 on fathead minnows (FHM)
Methods:
Toxicity of ammonia to fathead minnows
Goals of the ammonia baseline experiment
Concentrations of NH3 tested using a flow-through diluter system
Maintenance of the flow-through diluter system
Acclimation and exposure period implementation
Endpoints of the study
Vitellogenin Analysis
Results:
Fathead minnow reproductive effects from ammonia exposure
Fathead minnow fertilization effects from ammonia exposure
Analysis of ammonia in testing concentrations
Occurrence of a pump failure during the study
Fathead minnow vitellogenin response to ammonia exposure
Secondary sexual characteristic variability
Discussion:
Significant findings resulting from the ammonia baseline experiment
Comparison to Thurston et al. (1986)
Current U.S. EPA water quality criteria for ammonia
Disadvantages to using a flow-through diluter system
Artifacts of the experiment
Mobility of fathead minnows
Need for computational modeling
Conclusions

1
3
5

6
7
8
9
10
11
12
13
15
16
17
18
20
21
21
25
26
26
28
29
30

Chapter 3: Determining the mixture effects of NH3 and EE2 on FHM reproduction
Introduction:
Significance of testing with multiple stressors
31
32
Background of 17α-ethinylestradiol and its effects on fish reproduction
17α-ethinylestradiol mechanism of action and pathway
to obstructing fish reproduction
33
vi

Interaction of 17α-ethinylestradiol and ammonia in the environment
17α-ethinylestradiol study conducted by the U.S. EPA
Methods:
How study hypotheses were tested during this experiment
Reasons for choosing testing concentrations and their
environmentally relevant concentrations
Analytical analysis of 17α-ethinylestradiol samples
Results:
Mortality effect from mixture of 17α-ethinylestradiol and ammonia
Fathead minnow reproductive effects from individual contaminant
exposure and the mixture
Fathead minnow fertilization effects from individual contaminant
exposure and the mixture
Analysis of 17α-ethinylestradiol in testing concentrations
Analysis of ammonia in testing concentrations
Fathead minnow vitellogenin response to ammonia exposure in both
liver and plasma
Discussion:
Significance of the mortality effect found in the mixture of
17α-ethinylestradiol and ammonia
Previous studies that conducted experiments looking at the effects
of multiple stressors
Plausible reasons behind observed mortality effect
Vitellogenin response and possible mechanism of action
Why the flow-through diluter setup is a perfect protocol for
testing multiple stressors
Reasons why science needs to be geared toward testing with
multiple stressors
Possible aquatic multiple stressor exposures in the environment
Implications of populations to exposures of multiple stressors
Possible implications of longer-lived species
Chapter 4: Conclusions and Future Direction
Final conclusions determined by this study
Density dependence in wild populations
Extrapolating data to wild populations difficult?
Implications of mortality effect found in the mixture study
Possible ways to remove of ammonia from wastewater treatment
plant effluent
Possible ways to remove of 17α-ethinylestradiol from wastewater
treatment plant effluent
Future Directions

vii

37
38
39
40
41
42
43
43
44
45
46

49
50
51
52
53
53
54
54
55

56
57
57
59
60
60
62

Appendix A: Model organism and design of the flow-through
diluter system

65

Appendix B: Proper disposal of unwanted or expired medication

69

References

71

viii

LIST OF TABLES
Table 1

Fatpad scoring system used by the U.S. EPA.

ix

12

LIST OF FIGURES
Figure 1

FHM egg production per female per day during the 20 day
exposure to NH3

14

FHM cumulative fecundity during the 20 day exposure
to NH3

15

Figure 3

FHM fertilization rate during the 20 day exposure to NH3

16

Figure 4

Average total ammonia concentration during the 20 day
exposure to NH3

17

Average un-ionized ammonia concentration during the 20 day
exposure to NH3

18

Average female FHM liver vitellogenin concentration during
the 20 day exposure to NH3

19

Average male FHM liver vitellogenin concentration during
the 20 day exposure to NH3

20

FHM egg production per female per day during a U.S. EPA
conducted EE2 study

38

Average male FHM liver vitellogenin concentration during the
EE2 study

40

FHM egg production per female per day during the EE2 and
NH3 21 day mixture study

43

FHM fertilization rate during the EE2 and NH3 21 day mixture
study

44

EE2 analytical chemistry analysis during the EE2 and NH3 21
day mixture study

45

Average un-ionized ammonia concentration during the EE2 and
NH3 21 day mixture study

46

Average female FHM liver vitellogenin concentration during the
21 day mixture study

47

Average male FHM liver vitellogenin concentration during the
21 day mixture study

47

Figure 2

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

x

Figure 16

Figure 17

Figure 18

Figure 19

Average male FHM plasma vitellogenin concentration during
the 21 day mixture study

48

Average male FHM plasma vitellogenin concentration with
control outlier removed

48

Average female FHM plasma vitellogenin concentration
during the 21 day mixture study

49

Set up of the flow-through diluter system used during this study

68

xi

CHAPTER ONE: INTRODUCTION
Concern over the harmful effects of trace chemicals found in wastewater
treatment plant (WWTP) effluent has risen over the past two decades with the discovery
that such chemicals often act as endocrine disruptors (Parrott & Blunt 2004, Jensen et al.
2000, Lange et al. 2000). Wastewater treatment plant effluent is often comprised of a
mixture of chemicals from household and industrial sources and little is known about the
effects of individual endocrine disrupting chemicals (EDCs) on fish and wildlife. Perhaps
what is more alarming is that there is a paucity of information on how multiple EDCs
interact within individual organisms. Endocrine disrupting compounds are defined as
exogenous agents that interfere with the production, release, transport, metabolism,
binding, action, or elimination of endogenous hormones responsible for the maintenance
of homeostasis and regulation of developmental processes (Parks et al. 1999). A mixture
of chemicals can have unexplained interactive effects on the endocrine system of fish and
other aquatic organisms (Jobling & Tyler 2003). This study’s aim was to outline a
protocol for testing multiple stressors and to determine how two chemicals with different
modes of action interact within a fish to affect reproduction.
The effects of multiple stressor interactions on fish and wildlife are notoriously
difficult to assess because of several reasons. First, time lags between initial cause and
expression of cumulative effects can be long, thus making it difficult to determine a
cause-effect relationship (Adams 2005). Second, it is difficult to generate a hypothesis to
test because we often do not know if the combined effects of two chemicals will be
additive, synergistic or antagonistic. Synergistic effects occur when two chemicals are
mixed and the effects are more than the sum of their individual effects. Antagonistic

1

effects occur when the individual effects of two contaminants counteract one another
when mixed, and result in a much lower effect than predicted if the effects were additive.
Finally, several mechanisms may be involved during multiple stressor interactions
(Bonga 1997).
Information about the responses of animals to multiple natural and anthropogenic
stressors is, at the present time, insufficient for researchers to predict their combined
effects (Jenssen 2006); often the effects of just single stressors are not well-delineated.
However, the reality is that fish are exposed to multiple stressors in their natural settings
and there exists a need to understand how different chemicals interact within an organism
to assess population risk. Multiple stressors can interact with both the immune and
endocrine systems simultaneously and such interactions occur in many environments
where fish are threatened with pollutants, parasites, and other environmental stressors
(Jobling & Tyler 2003).
Endocrine disrupting compounds, a common component of WWTP effluent, are
found in many water systems throughout the United States and around the world. These
chemicals have been linked to high levels of intersex, sex change and skewed sex ratios
in fish populations (Parrott & Blunt 2005, Lange et al. 2001. Jensen et al. 2001). Sources
of EDCs have been cited as pesticides, oral contraceptives, dioxins, and other industrial
compounds (Parrott & Blunt 2005, Lange et al. 2001). The majority of these chemicals
are released into the aquatic environment via sewage systems and very few WWTPs have
proper mechanisms in place to completely remove these contaminants (Parrott & Blunt
2005).

2

Of particular interest are chemicals that mimic naturally produced estrogen; such
chemicals have the potential to disrupt the endocrine system during the critical sexual
development period (Lange et al. 2001). These estrogen imitators have been shown to
induce feminization in male fathead minnows, Pimephales promelas (FHM), by
stimulating the production of eggs and plasma vitellogenin, a lipo-phospho-glycoprotein
precursor to egg yolk that is normally produced in females (Parks et al. 1999). Estrogen
mimics also disrupt egg production in female FHM (Parrott & Blunt 2005). The process
of vitellogenesis is sensitive to EDCs (Murphy et al. 2005), and perturbations in this
process can have ecological consequences, because vitellogenesis is directly related to
fecundity and egg quality of individual fish. This study examined the effects of 17αethinylestradiol (EE2) which is a synthetic estrogen and the most commonly used active
ingredient in oral contraceptives (Lange et al. 2001) and a known EDC (Ankley et al.
2001, Lange et al. 2001, & Parrott & Blunt 2005).
Ammonia is another common pollutant in aquatic systems and has been found to
be extremely toxic to fish.

Effluent generally contains very low concentrations of

ammonia; however during WWTP malfunctions or in parts of the world that have
nonexistent or ineffective treatment systems, ammonia concentrations in waterways can
be greater than 20mg/L (Passell et al. 2007). Farm animal operations are another source
of water pollution including ammonia and the extent of pollution is entirely dependent on
the size of the production (Robbins et al. 1972). Fertilizers used during agricultural
operations also contain a significant amount of nitrogen. During runoff and flooding
events, nitrogen produced by these types of operations could be released into nearby
streams.

3

Ammonia exists in two forms, ionized (NH4+) and un-ionized (NH3) (Thurston et
al. 1986, Mayes et al. 1986). While total ammonia concentration is heavily monitored in
WWTP effluent across the United States, NH3 concentrations generally are not (Passell
et al. 2007) even though un-ionized ammonia is the larger contributor of ammonia
toxicity to fish (Thurston et al. 1986, Mayes et al. 1986, U.S. EPA 1999). As a neutral
molecule it is able to easily diffuse across the epithelial membranes of aquatic organisms
(U.S. EPA 1999). The toxicity of un-ionized ammonia is highly dependent on
temperature and pH of the water (U.S. EPA 1999). As the temperature and pH of water
increase, the toxicity of un-ionized ammonia to FHM increases as well (U.S. EPA 1999).
Increased ammonia levels within the body result in both chronic and acute effects
in fish. Ammonia targets the central nervous system of vertebrates and at high
concentrations can result in acute toxicity (Randall & Tsui 2002). Convulsions followed
by death are likely the result of this acute toxicity during exposures of high NH3
concentrations. The mechanism of action for this toxicity is believed to be from an
increased activation of the glutamate receptors for the amino acid, N-methyl-D-aspartic
acid (NMDA) (Randall & Tsui 2002). These receptors are responsible for controlling
synaptic plasticity and memory function. This increased activation of these receptors
leads to a depolarization of neurons and results in subsequent cell death within the central
nervous system.
Human urine contains a high amount of ammonia and can contain trace amounts
of EE2 (Parrott & Blunt 2005). It is possible that fish could be exposed to mixtures of
both EE2 and ammonia during release of WWTP effluent because of prevalence of birth
control usage and presence in urine, as well as the inability of WWTP to completely

4

remove these chemicals from effluent. In this study I tested the effects of both ammonia
and EE2 on fathead minnow survival and reproduction. Specifically, the ultimate goal of
this project was to determine whether ammonia will impact the exposure of FHM to EE2
and which endpoints are most sensitive to use in detecting changes in exposure and
effects. The hypotheses of this study were:
H1: Current published lowest observable effect concentrations for EE2 effects on
FHM reproduction will be reduced when coupled with the multiple stressor effect
of ammonia because having an added environmental stressor present will increase
the fish’s ventilation rate and therefore increase the uptake of EE2.
H2: Ammonia concentrations at or near U.S. EPA Water Quality Guidelines will
result in reduced vitellogenin expression in FHM males and females because
stress has been shown to decrease vitellogenin production in fish.
This thesis is split up in to separate chapters which focus on addressing these
hypotheses. Chapter two investigates how NH3 affects FHM reproduction during a 21
day flow-through diluter study. This experiment was needed because of the lack of
background information on the effects of NH3 on FHM reproduction; the most recent
study prior to this experiment was conducted in 1986 (Thurston et al. 1986).
Specifically, for this study, I needed to determine the NOEC of NH3 on FHM
reproduction using the 21 day flow-through diluter method so that I could test the
effects of both EE2 and NH3. Chapter three investigates the mixture effects of NH3
and EE2, each at their NOEC, on FHM reproduction and addresses both hypothesis
(H1 and H2).

5

CHAPTER TWO: DETERMINING THE NO OBSERVABLE EFFECT
CONCENTRATION OF NH3 ON FATHEAD MINNOW REPRODUCTION
Introduction:
Ammonia is a common pollutant in aquatic environments and sources include
agricultural run-off, biological waste decomposition and WWTP effluent (Randall & Tsui
2002, U.S. EPA 1999). The effects of ammonia on aquatic systems have been
documented world-wide, most commonly in regions with a high human population and/or
large animal farms (Randal & Tsui 2002). In fish, ammonia toxicity has been reported to
cause chronic effects such as reduced growth, reproduction and fertilization rates as well
as acute effects including convulsions and comas which result in subsequent death
(Randall & Tsui 2002, Thurston et al. 1986).
Fathead minnows were chosen as the model organism for this experiment. This
species is a part of the ecologically relevant Cyprinidae family and its reproductive life
cycle can be altered in a laboratory setting to produce clutches of 50-100 eggs every 3-5
days (Appendix A). FHM have a broad distribution across North America in both lentic
and lotic environments (Ankley and Villeneuve 2006) and are tolerant of a wide range of
water types (Ankley and Villeneuve 2006).
Ammonia toxicity has chronic effects on fish such as damage to the gills and
respiratory impairment and will result in acute effects such as mortality at sufficiently
high concentrations. Ammonia also has been known to reduce growth and inhibit
reproductive success in FHM (Thurston et al. 1986).

The no observable effect

concentration (NOEC) and lowest observable effect concentration (LOEC) of NH3 on
FHM egg production and viability were reported as 0.37mg/L and 0.91mg/L, respectively
(Thurston et al. 1986). NOEC is the greatest concentration or amount of a substance as
6

determined by an experiment or observation, that causes no alterations of morphology,
functional capacity, growth, development, or life span of target organisms distinguishable
from those observed in control organisms of the same species and strain under the same
defined conditions of exposure (McNaught & Wilkinson 1997). LOEC is the lowest
observable concentration or amount of a substance as determined by an experiment or
observation, that does not cause any alteration of morphology, functional capacity,
growth, development or life span of target organisms of the same species and strain under
the same defined conditions of the exposure (McNaught & Wilkinson 1997). FHM
survival NOEC and LOEC were 0.44mg/L and 0.91mg/L, respectively. Another metric,
the LC50 which is a statistically derived concentration of a substance in an environmental
medium expected to kill 50% of organisms in a given population under a defined set of
conditions (McNaught & Wilkinson 1997), was estimated to be 1.50mg/L for FHM
(Alexander et al. 1986). Brain lesions were noted at a concentration of 0.21mg/L but not
found at a concentration of 0.11mg/L NH3 (Thurston et al. 1986).
The purpose of this experiment was to test environmentally relevant
concentrations of NH3 and to determine the NOEC of NH3 on fish reproduction that
would be used during the mixture experiment portion of this study (Chapter 3). There
was a need to determine a NOEC on FHM reproduction using the 21 day flow-through
diluter method. Prior to this study, the most recent study reporting FHM reproductive
effects after an NH3 exposure was Thurston et al. (1986). This study was a life-cycle
study, therefore the published NOECs needed to be tested in order to determine if they
were true NOECs using my methodology.

7

I determined the NOEC on FHM reproductive endpoints by using a flow-through
diluter system to continually dose adult male and female FHM to three concentrations of
NH3. The endpoints of this experiment included fecundity, fertility, male secondary
sexual characteristics, mortality and vitellogenin expression in both plasma and liver.
Fecundity, fertility and mortality were monitored on a daily basis while vitellogenin
expression and male secondary sexual characteristic endpoints were measured at the end
of the 21 day study.
Methods:
This study used a flow-through diluter system to dose FHM to three
concentrations of ammonia and a control. The setup and design of the flow-through
diluter system is an established U.S. EPA protocol for conducting 21 day toxicity tests
and allows for a continuous toxicant dosing at predetermined concentrations throughout
the entire study (Appendix A). Total ammonia concentrations tested in this study were
5.0mg/L (0.06mg/L NH3), 15.0mg/L (0.18mg/L NH3), and 30.0mg/L (0.42mg/L NH3).
The U.S. EPA lab in Cincinnati, OH conducted a survey of 50 wastewater treatment
plants from all over the United States in 2004. The average concentration of NH3 for
these samples was 0.1mg/L and was as high as 0.76mg/L NH3 in an individual sample
(Lazorchak & Smith 2004). Thurston et al. 1986 reported the NH3 NOEC on FHM
reproduction was 0.37mg/L. Therefore, my NH3 baseline experiment tested
concentrations similar to those tested before by Thurston et al. (1986) and also bracketed
the average WWTP effluent NH3 concentration. Ammonium chloride (NH4Cl) was
chosen as the source for ammonia in this study because it did not produce any
precipitants at the pH and temperature of the control/dilution water in the presence of
fish. In a previous pilot study, I tried to use ammonium phosphate dibasic. When fish
8

were introduced to the system and fed, the pH of the tanks increased to the point where
the ammonium phosphate dibasic reacted with the control water and produced a calcium
phosphate precipitate. Although this precipitate was probably not toxic, I wanted to
eliminate any potentially confounding variables. Therefore, ammonium chloride was
chosen as the NH3 source for this study.
Water chemistries were monitored several times each week throughout the study
to ensure proper NH3 exposure. Total ammonia concentrations in individual testing
chambers were measured once each week. Also, total ammonia concentration was
measured three times each week from a composite of all 7 replicates within a treatment
group. Because temperature and pH can affect total ammonia concentrations, temperature
and pH readings for each testing chamber were recorded soon after a total ammonia
measurement was performed. I estimated NH3 concentrations by using an equation
provided

by

the

Michigan

Department

of

Environmental

Quality:

ܰ‫ 3ܪ‬ൌ ܶ‫/ܽ݅݊݋݉݉ܣ ݈ܽݐ݋‬ሺ1 ൅ 10^ሺ0.09018 ൅ ሺ2729.92/ሺTemp°C ൅ 273.15ሻሻ – pHሻሻ.

The diluter system required daily maintenance to ensure proper dosing during the
21 day exposure period. Superstocks were mixed fresh daily to provide the diluter system
with a constant source of toxicant. The 30mg/L total ammonia (0.42mg/L NH3)
superstock was made by mixing 48g of NH4Cl into 18L of laboratory line water. The
15mg/L total ammonia (0.18mg/L NH3) superstock was made up by mixing 24g of
NH4Cl into 18L of laboratory line water. The 5mg/L (0.06mg/L NH3) total ammonia
superstock was made by mixing 8g of NH4Cl into 18L of laboratory line water. The
pumps were calibrated three times each week to ensure proper mixing of toxicant and

9

target concentration delivery to the testing chambers. The tubes for the peristaltic pumps
were replaced weekly to prevent any breakage or tube malfunctioning.
This experiment consisted of a 14 day acclimation period and a 21 day exposure
period. A total of 168 six month old FHM were used during each experiment study.
Standard lengths and wet weights of all males and females were collected prior to the
study to ensure each individual met the testing criteria (Appendix B). Fish were randomly
placed into the testing chambers immediately following weight and length data
collection. Because handling induces stress and ultimately the production of cortisol, the
14 day acclimation period allowed these individuals to acclimate to the testing
environment prior to the exposure period. Any mortality occurrences during the
acclimation period were recorded and the sex of the individual was noted; deceased fish
were replaced only during the acclimation period. Reproduction was monitored during
the acclimation period to ensure each treatment was producing similar numbers of eggs
prior to exposure and to ensure no significant mortality was occurring. If any statistically
significant differences in reproduction or high mortality occurred, the study was
terminated and a new batch of fish chosen. On day 14 of the acclimation period, the
toxicant pumps were turned on and dosing began for the exposure period.
The exposure period for each experiment ran for 21 consecutive days during
which survival, fecundity, fertility and mortality data was recorded daily. Any fish that
died during the 21 day exposure period was not replaced. Bacteria convert ammonia to
nitrite (NO2) and this can be detrimental to a study (Thurston et al. 1986), therefore
testing chambers were cleaned three times each week to reduce the amount of bacteria
buildup. Extreme caution was taken during tank cleaning to prevent unnecessary stress to

10

the fish. Tanks were cleaned using a sponge to scrape down the sides of each tank. A sink
siphon was then used to remove excess food and debris. Daily fecundity data, by tank,
was recorded by counting the number of eggs laid on each spawning tile. Tiles containing
eggs were marked with respective tank number and egg count data and then placed in an
aerated egg bath of 25˚C ± 1˚C. Seventy-two hours after collection, fertilized eggs in the
eyed stage were counted.
During the study, fish were fed on a regular schedule in order to maximize
reproduction. Each testing chamber was fed three times each day. Between 8:00 and
10:00am EST, the fish were fed 20ml of concentrated newly hatched (<24H old) brine
shrimp, Artemia salina. Between 12:00 and 1:00pm, the fish were then fed with 1.0ml of
frozen adult brine shrimp (San Francisco Bay Co.). The final feeding occurred between
4:00 and 5:00 with 20ml of concentrated newly hatched brine shrimp supplied to each
chamber. The feeding regimen adopted here was found to maximize reproduction and
provide the healthiest fish during a 21 day control pilot study conducted by the U.S. EPA
in Cincinnati, OH.
The NOEC ammonia study was terminated after 21 days of exposure and the fish
were transferred for necropsy. Endpoint data collected were male secondary sexual
characteristics, fecundity, fertility, gonadosomatic index (GSI), fatpad index (FPI), liver
vitellogenin concentration and plasma steroid concentration. Prior to necropsy, the fish
were anesthetized in a 200mg/L concentration of tricaine methanesulfonate (MS-222) and
standard length and wet weights were measured. The caudal tail was severed to allow for
collections of blood from the caudal artery via a heparinized capillary tube. Following
collection, the heparinized capillary tube was centrifuged at 3000rpm for 1 minute to

11

separate the plasma from the red blood cells. After centrifugation the plasma was
transferred into a labeled vial and then snap-frozen in liquid nitrogen. Immediately
following plasma collection, the fish was euthanized by severing the spine just behind the
nape of the fish. Secondary sexual characteristics of males were then recorded. Each
tubercle was counted and fatpad size was scored using an U.S. EPA scoring system (U.S.
EPA 2002), summarized in Table 1.

Rating
None
Small
Medium
Large

Table 1. Fatpad Scoring System used by the U.S. EPA.
Description
No fatpad visible
Small fatpad evident
Fatpad is prominent, clearly above the body surface but not overhanging
Fatpad is very prominent and is starting to overhang the body surface

Each fatpad was separated using a scalpel and then weighed. This weight was then
divided by the respective fish’s body weight to calculate a fatpad index (FPI). Gonads
and livers from males and females were removed and a wet tissue weight was taken.
These weights were then divided by the body weight of the fish to calculate a
gonadosomatic index (GSI) and hepatosomatic index (HSI) for each fish.
Vitellogenin in the liver samples was measured using “real time” quantitative
polymerase chain reactions (QPCR) via a method already described (U.S. EPA 2009).
Total RNA from the liver samples was isolated using TriReagent (U.S. EPA 2005).
Relative concentrations of the total RNA were then measured using an ultraviolet
spectrophotometer. The isolated RNA was then converted into complementary DNA
(cDNA) using a reverse transcriptase. A diluted sample of cDNA was then used for the
PCR reactions along with a Hot Start DyNAmo™ SYBR® green master-mix.
Amplification of normalizing gene (18S) sequences was performed using universal 18S
12

primer pairs. Cycling was carried out using a thermocycler in cycles of 94°C, 60°C and
70°C. The fluorescent intensity of vitellogenin response from each sample was then
calculated as a ratio of vitellogenin:18S and compared against the controls. An Enzymelinked immuno-sorbent assay, ELISA, was used for the detection and quantification of
vitellogenin in the plasma samples. These samples were sent to the U.S. EPA laboratory
in Duluth, Minnesota for analysis.
Results:
FHM were sensitive to ammonia in both egg production and egg fertilization
after 20 days of exposure. The analysis only includes the first 20 days of exposure as the
control/dilution water pump failed on exposure day 21. The pump failure significantly
increased ammonia concentrations and lead to mortality in each of the treatments.
Therefore, day 21’s egg production was removed from the analysis. The NH3 baseline
experiment met EPA’s mortality criteria in the controls (≥90% survival) as total mortality
was 2.4% (97.6% survival) over the 20 day study. The only treatments that experienced
mortality in the 20 day exposure were the control and 5mg/L NH3 treatment. These
treatments each had a single male mortality; no female mortality was experienced during
the 20 day exposure period.
Over the course of the 20 day exposure period the average eggs per female per
day in the control, 5.0mg/L (0.06mg/L NH3), 15mg/L (0.18mg/L NH3) and 30mg/L total
ammonia (0.42mg/L NH3) were 16.17 ± 2.69, 11.61 ± 2.53, 8.2 ± 1.58 and 5.85 ± 1.47
eggs per female per day, respectively (Figure 1).

13

# Eggs produced per
female per day

20
16
12

*
*

8
4
0
NH3 Concentration
Control

0.06mg/L

0.18mg/L

0.42mg/L

Figure 1: 21-Day Treatment Flow-through Diluter Egg Production/Female/Series/Day.
Sample Size (N) = 28. Conducted 8/26/09. Unionized Ammonia (NH3) Baseline
Experiment. Treatments included control, 0.06mg/L, 0.18mg/L and 0.42mg/L NH3. Error
bars are ± 1 standard error. Study duration was 20 days. * = significant difference from
the control, p = 0.05.

This data set was not normally distributed so a non-parametric test, Kruskal-Wallis, was
performed. The Kruskal-Wallis test determined that the treatments were significantly
different from one another (K = 10.20, p = 0.017, df = 3). A Kruskal-Wallis Post hoc test
(kruskalmc) was then performed in Program R and determined that the two highest
concentrations (0.18 and 0.42mg/L NH3) were significantly different from the control (p
= 0.05). Because female mortality did not occur during the 20 day exposure period, I
could analyze cumulative egg production. When male mortality occurred, it was usually
only a single male from a replicate and therefore I did not have to adjust the collected
data.
After day 1 of exposure, all treatments produced fewer cumulative number of
eggs than the control indicating a rapid response of egg production to ammonia exposure
14

(Figure 2). Cumulative egg production at the end of the experiment was normally
distributed and verified by a non-significant Shapiro-Wilks test for normality (W=0.9846,

Cumulative Number of
Eggs Spawned

p = 0.9426).
10000
7500

*

5000

*
*

2500
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Time (d)
Control
0.06mg/L
0.18mg/L
0.42mg/L

Figure 2: Cumulative fecundity by concentration during the 21 day exposure period.
Sample Size (N) = 7. Conducted 8/26/09. Unionized Ammonia (NH3) Baseline
Experiment. Treatments included control, 0.06mg/L, 0.18mg/L and 0.42mg/L NH3.
Study duration was 20 days. * = significant difference from the control, p < 0.05.

Variances amongst treatments were similar as verified by a non-significant Levene’s test
for homogeneity of variance (F=0.2436, p = 0.865, df = 3). An analysis of variance
(ANOVA) was performed on cumulative egg production and found to be significantly
different among treatments (F=10.167, p = <0.01, df = 3). A Tukey Honestly Significant
Difference (Tukey HSD) test found that all three concentrations were significantly
different from the control in terms of cumulative egg production (p = <0.05 for each
treatment).
This experiment also met the aforementioned fecundity and fertility rate criteria
(Ankley et al. 2001). Fertility rates during the 21 day exposure period in the control,
5.0mg/L, 15mg/L and 30mg/L total ammonia were 89.66%, 94.39%, 92.99%, and
80.32%, respectively (Figure 3).
15

100%

Fertilization Rate

*

75%
50%

25%

0%
Control

NH3 Concentration
0.06mg/L
0.18mg/L

0.42mg/L

Figure 3: 21-Day Treatment Flow-through Diluter Concentration Fertilization Rate.
Sample Size (N) = 7. Conducted 8/26/09. Unionized Ammonia (NH3) Baseline
Experiment. Treatments included control, 0.06mg/L, 0.18mg/L and 0.42mg/L NH3.
Error bars are ± 1 standard error. Study duration was 20 days. * = significant
difference from the control, p = 0.046.
The fertility data were not normally distributed amongst treatments, however each
individual treatment was normally distributed as verified by individual Shapiro-Wilks
tests (W = 0.9744 p = 0.9283 for control, W = 0.8908 p = 0.2789 for 5mg/L NH3, W =
0.9398 p = 0.6369 for 15mg/L NH3 and W = 0.9927 p = 0.9948 for 30mg/L NH3. The
ANOVA that tested for differences in egg fertilization between treatments was significant
(F = 6.141, p = .003, df = 3). The 30mg/L total ammonia group was statistically different
from the control as verified by a Tukey HSD test (p = 0.046).
Ten ammonia sampling events occurred throughout the 21 day study in each of
the treatments. The control and blank tanks did not have significant concentrations of
ammonia (Figure 4). The blank tanks were required to monitor the water quality of the
control and dilution water without the presence of fish. Therefore if mortality in the

16

Unionized Ammonia (mg/L)

30
25
20
15
10
5
0

Total Ammonia Concentration
Control
5mg/L
15mg/L
30mg/L
Figure 4: Average total ammonia concentrations from each test concentration.
Conducted 8/26/09. Unionized Ammonia Baseline Experiment. Treatments
included control, 5 mg/L, 15mg/L and 30mg/L total ammonia. Error bars are ± 1
standard error. Study duration was 20 days.
control tanks had occurred, I could check the blank tank water quality data to determine if
water quality was an issue. The final ammonia concentration in the tanks matched their
desired concentrations. The 5mg/L total ammonia treatment tanks averaged 4.0mg/L total
ammonia (range: 2.74 to 4.74mg/L). The 15mg/L total ammonia treatment tanks
averaged 12.1mg/L total ammonia (range: 10.40 to 14.22mg/L). The 30mg/L total
ammonia treatment tanks averaged 26.8mg/L total ammonia (range: 24.65 to 31.2mg/L).
Mean measured concentrations throughout the study in the treatment tanks appear
sufficiently similar to expected concentrations, however the coefficient of variance (CV)
was well outside acceptable parameters (<20%). The dilution water pump failed on the
last day of the study and as a result the ammonia concentrations spiked significantly on
the last day because only the toxicant pump was dosing the undiluted superstock to the
tank replicates. Removal of data collected on this last day resulted in CV’s of ammonia
concentrations that were within an acceptable range (<20%). I adjusted the un-ionized

17

ammonia concentration for temperature and pH using the Michigan Department of
Environmental Quality’s formula (Figure 5).

Unionized Ammonia (mg/L)

0.5
0.4
0.3
0.2
0.1
0.0
NH3 Concentration
Control

0.06mg/L

0.18mg/L

0.42mg/L

Figure 5: Average unionized ammonia (NH3) concentrations from each test
concentration. Conducted 8/26/09. NH3 Baseline Experiment. Treatments included
control, 0.06mg/L, 0.18mg/L and 0.42mg/L NH3. Error bars are ± 1 standard error.
Study duration was 20 days.
Other chemical parameters including dissolved oxygen (ANOVA, F = 0.003, p = 1.00, df
= 3), pH (ANOVA, F = 2.940, p = 0.065, df = 3) and temperature (ANOVA, F = 0.144, p
= 0.932, df = 3) were determined for the test solutions and showed no differences
between treatments. Conductivity was significantly different between treatments
(ANOVA, F = 19.109, p = < 0.01, df = 3) which was to be expected considering
ammonium chloride was being added to each NH3 treatment.
Liver vitellogenin concentrations were analyzed on both male and female FHM.
Exposure to ammonia did not induce vitellogenin production in male fish (Figure 6).
While the control male livers exhibited a slight vitellogenin response, this was considered
non-significant as the females within the tank were most likely releasing some natural
estrogens which may have caused the males to display low levels of vitellogenin.
18

Vitellogenin/18S

1.E-03
8.E-04
6.E-04
4.E-04
2.E-04
0.E+00
Control

NH3 Concentration
0.06mg/L
0.18mg/L

0.42mg/L

Figure 6: Average male vitellogenin liver concentrations. Sample Size (N) = 56.
Conducted 8/26/09. Unionized Ammonia (NH3) Baseline Experiment. Treatments
included control, 0.06mg/L, 0.18mg/L and 0.42mg/L NH3. Error bars are ± 1
standard error. Study duration was 20 days.
A log transformation was conducted on the data to account for non-normal distributions.
A Shapiro-Wilks test for normality was conducted on the log transformed data
(W=0.9612, p = 0.1196). Variances amongst treatments were similar as verified by a nonsignificant Levene’s test for homogeneity of variance (F=2.76, p = 0.0537, df = 3).
Concentration of vitellogenin in male livers suggests a reduction in vitellogenin, however
the reduction was not significant (F = 1.34, p = 0.27, df = 3). Female vitellogenin
expression also was not significantly different from the controls (K = 2.44, p = 0.49, df =
3) (Figure 7). The data was not normally distributed so therefore the Kruskal-Wallis nonparametric test was used in the analysis. A non-parametric test was conducted on the
female liver vitellogenin concentration after several transformations failed to produce a
normalized data set.

19

25

Vitellogenin/18S

20
15
10
5
0
Control

NH3 Concentration
0.06mg/L
0.18mg/L

0.42mg/L

Figure 7: Average female vitellogenin liver concentrations. Sample Size (N) = 112.
Conducted 8/26/09. Unionized Ammonia (NH3) Baseline Experiment. Treatments
included control, 0.06mg/L, 0.18mg/L and 0.42mg/L NH3. Error bars are ± 1
standard error. Study duration was 20 days.
Fatpad, gonadosomatic and hepatosomatic index data were collected during this
experiment, however the results were extremely variable and bias was introduced
depending on which technician was removing the tissues. For example, one fatpad was
marked as “Large” but weighed less than a fatpad that was marked “Small”. I felt that
these differences were sufficient means to disregard these data in my analysis. Similarly,
because the evaluation of other secondary sex characteristics was subjective and variable,
I decided to disregard tubercle count data. For future studies, it is advised that one
technician concentrate solely on one tissue throughout the entire necropsy to reduce the
variability in tissue collection and tubercle count data.
Discussion:
The goal of my study was to determine a NOEC for NH3 on fish reproduction and
my results were surprising when compared to what has been published previously. The
lowest NH3 concentration that I tested was 5mg/L total ammonia (0.06mg/L NH3),
20

which caused a significant fecundity effect (29% reduction) in cumulative eggs after a 20
day exposure. This concentration is far below any NOEC reported previously. Thurston
et al. (1986) reported 0.37mg/L NH3 as the no effect level on FHM egg production. The
observed reproductive effect at such very low concentrations is significant as the current
average NH3 concentration discharging from a US WWTP effluent is estimated to be
0.1mg/L (Lazorchak & Smith 2004), twice that of the 0.06mg/L NH3 concentration
found to cause an effect on fecundity during this study.
The differences in results between this study and those reported by Thurston et al.
(1986) were likely due to differences in methodology. Thurston et al. (1986)
implemented a life cycle study where three to five day old FHM larvae were tested in
concentrations of NH3 beyond the age of maturation. My study was only a 21 day study
that exposed adult FHM to similar concentrations of NH3. Prior to my study, I speculated
that the 21 day study would result in much higher NOECs than the life cycle study as the
length of NH3 exposure was much shorter. Since my study ended up determining
considerably lower NOECs for FHM reproduction, the differences in methodology needs
to be documented.
The control reproduction was extremely low in the study conducted by Thurston
et al. (1986). The tanks averaged only 1.29 eggs per female-day. This low reproduction
did not meet the 15 eggs per female per day criteria used during my study (Ankley et al.
2001). The controls also produced fewer eggs than any NH3 concentration except for the
0.91mg/L NH3 concentration that induced 100% mortality. It is unknown as to why the
controls produced so few eggs during the life cycle study of Thurston et al. (1986) but

21

could be related to having a different male to female ratio than the one implemented
during my 21 day exposure.
Thurston et al. (1986) also used a 30L tank volume compared to my 9.5L volume.
The difference in tank volume was due to differences in the amount of fish being exposed
in each replicate. I wanted to maximize the number of replicates for each tank
concentration; therefore I tested 7 replicates per concentration each with a 2 male to 4
female ratio. This ratio was known to optimize control reproduction from an earlier 21
day flow through diluter pilot study. Thurston et al. (1986) only tested one replicate per
treatment, each containing 50 three to five day old larvae. After 60 days of exposure the
50 fish within each tank were randomly thinned to 15 fish each containing no more than 4
FHM males. My 21 day test tested a more stringent male to female ratio than Thurston et
al. (1986). The number of males in each treatment ranged from 3 to 5 and females ranged
from 3 to 11 in Thurston et al. (1986). The differences in the number of males and
females between treatments could have impacted reproduction during the life cycle study
reported by Thurston et al. (1986).
Flow rates to the tanks were also considerably different between studies; Thurston
et al. (1986) implemented a 1L flow every 4-5 minutes where my study used a 40ml/min
flow. This difference was unlikely to influence reproduction as there were no reports of
water quality issues for either study. Temperature and dissolved oxygen concentrations in
the control tanks were similar between each study; Thurston et al. (1986) used a control
water with a temperature at 24.0°C (+/- 0.08°C) and dissolved oxygen content of
6.49mg/L (+/- 0.41) where my study used a control water at 25.1°C (+/- 0.30°C) and
dissolved oxygen content of 6.26mg/L (+/- 0.84). The pH of the control water was

22

different between each study; Thurston et al. (1986) control water had a pH of 8.05 (+/0.04) and in my study was 7.43 (+/- 0.06). It is unknown as to whether or not the
difference in pH resulted in differences in reproduction between the two studies.
Thurston et al. (1986) also tested a fluctuating photoperiod where the amount of daylight
shortened and lengthened throughout the life cycle study (10.0 – 15.75 hours of daylight)
and my study kept the photoperiod constant (16 hours of daylight).
The fish in my 21 day study were fed only newly hatched brine shrimp and frozen
adult brine shrimp. Thurston et al. (1986) fed both newly hatched brine shrimp and
commercial trout food (Silvercup salmon). I decided not to use trout food in my study
because it builds up easily and fouls the water; I wished to reduce the frequency of tank
cleanings to minimize disturbance. Differences in food quality could have impacted
reproduction, however this is unlikely as the fish tested in my study and those tested by
Thurston et al. (1986) were not reported to show signs of emaciation. In light of all these
differences, I would recommend that the study conducted by Thurston et al. (1986) be
repeated by the U.S. EPA in Cincinnati, Ohio in order to determine if the results
documented in that study can be replicated. I also recommend that the study be repeated
and controlled for consistent photoperiod and male to female ratio in each exposed
chamber and also implemented using a minimum control reproduction criteria specified
by the U.S. EPA that is similar to the 15 eggs per female per day criteria used in my
study.
The cumulative egg production appeared to be the most sensitive endpoint, but
other endpoints indicated potential for quantifying NH3 exposure and effect. During the
last week of exposure, days 14 to 21, FHM in all tested NH3 concentrations appeared to

23

have a reduction in reproduction. The two highest concentrations, 0.18mg/L and
0.42mg/L NH3, produced very few eggs during the last five days of exposure. If this
study were to be carried out for an additional length of time, I think changes in the
average eggs produced per female per day after exposure to NH3 endpoint would have
shown similar reductions as cumulative egg production as the treatments appeared to
produce fewer eggs over time in comparison to the control.
The determination of a NOEC was difficult because all NH3 concentrations
caused a significant fecundity effect. Therefore, I assumed that 50% of the LOEC
(0.06mg/L NH3) was a good estimate of NOEC (0.03mg/L NH3) on FHM fecundity.
This assumption is justified because the 0.06mg/L NH3 concentration caused a
significant effect in only one of the two fecundity endpoints.
The lowest concentration tested in this study, 0.06mg/L NH3, is lower than U.S.
EPA water quality guidelines for adult fish exposed to NH3. These current water quality
criteria for fish are based on toxicity tests using rested or non-stressed fish (Randall &
Tsui 2002). These criteria may be overestimating the NOEC because the criteria do not
account for swimming fish which generally have elevated internal ammonia levels
compared to resting (Randall & Tsui 2002) nor stressed fish which have an increased
level of cortisol. Cortisol is the primary steroid produced upon stimulation by an
environmental stressor (Giesy et al. 2003). Stress, which can be induced from a variety of
biological and chemical agents, and the induction of cortisol have been known to increase
ammonia toxicity in some fish species (Randall & Tsui 2002). Furthermore, the
formation of cortisol can also decrease the production of vitellogenin in fish (Giesy et al.
2003). Additional studies are needed to obtain a full understanding of how ammonia

24

toxicity relates to cortisol production in FHM and at which rate the induction of
vitellogenin is reduced.
Currently, the U.S. EPA (1999) has developed its recommended water quality
criteria based on whether or not salmon or early life stage fish are present within a body
of water. There are two criterions that a freshwater source must meet: an acute criteria
concentration (CMC) and a chronic criteria concentration (CCC). The literature states
that freshwater aquatic organisms will not be affected adversely if the one hour average
concentration of total ammonia does not exceed, more than once every three years on the
average, the CMC. In addition, they will not be affected unacceptably if the 30 day
average of total ammonia does not exceed, more than once every three years on the
average, the CCC. These criteria are temperature and pH based, so for these purposes I
standardized all NH3 values to a pH and temperature similar to the dilution water tested
in this study, 7.3 and 25°C, respectively. If salmon are present, the CMC is 5.62mg/L
total ammonia and if they are not present the CMC is 8.4mg/L (0.063 mg/L and 0.094
mg/L NH3, respectively). If early life stage fish are present the CCC is 1.23mg/L and if
they are not present the CCC is 1.8 mg/L total ammonia (0.014 mg/L and 0.02 mg/L
NH3).
The reduction in reproductive endpoints in terms of both fecundity and
fertilization success can be detrimental to fish populations. The current CMC guidelines
for adult, non-salmonid fish such as the fathead minnow is currently set at 0.094 mg/L
NH3. This concentration is well above my observed LOEC concentration of 0.06 mg/L
NH3. This tested concentration considerably lowered reproductive output (29%

25

reduction) when compared to the control. The production of fewer eggs could result in
decreased population abundance for fish exposed to these levels of NH3.
In this study I found a few disadvantages of using a flow-through dilutor system.
Such disadvantages include the challenge of maintaining a constant rate of NH3, the
dependence on functional and operational equipment, and a reliance on chemistry data
for nominal concentrations. In a natural environment, NH3 concentrations change due to
rain events and fluctuations in WWTP effluent concentrations. However, because this test
was only looking at the effects of NH3, I did not need to mimic rain events or
fluctuations in toxicant concentrations. Pump failure could have resulted in an alteration
of tested concentrations and indeed, pump failure occurred at the end of the NH3 baseline
experiment on Day 21. The laboratory line control and dilution water pump failed which
meant that only the superstock was dispensed into the testing chambers. This caused a
significant spike in NH3 concentrations, therefore only the first 20 days of the study were
used to analyze reproductive endpoints. Despite issues with the flow-through diluter
system, this study still produced several significant results.
Since the fish were housed in relatively small aquaria, it is possible that artifacts
and unknown stressors could have influenced the results of this study. Territorial defense,
mate selection, foraging for food and overall male dominance were probable throughout
the study. However, much consideration was taken to limit any competition stress during
the study. Three spawning tiles were placed in to each tank in order to provide each male
a spawning site to court female; the remaining spawning tile was to serve as a hiding
place for the remaining females. Each day the replicates were fed at three different times
in order to reduce the aggressiveness of the fish over food availability. Other stressors

26

including tank cleaning and technician presence were also minimized to reduce
influencing endpoint data. Cleaning was conducted in order to remove excess food and
fish waste. Since the fish biomass within each tank differed by less than 10% (12.89g +/0.6g) the amount of waste being produced by the fish was consistent between replicates.
Therefore the amount of time it took to clean each tank was approximately the same
between replicates and this stressor should have had minimal influence on reproduction.
Also, when the fish were not being fed or tanks being cleaned, a black liner was pulled
down over the diluter system to minimize the presence of technicians walking around the
room. Therefore it these experimental artifacts were not likely to strongly influence the
results of this study.
Fish can be exposed to elevated NH3 concentrations both internally and
externally and both exposures can have effects on reproduction. Increased exposure to
environmental NH3 reduces the ability of the fish to reduce its internal NH3
concentration. Many fish species have the ability to detoxify internal ammonia by
converting it to glutamine, glutamate or urea (Miller 2001). As the exposure to external
ammonia is prolonged, this detoxification mechanism is weakened and ultimately the fish
experiences ammonia toxicity.
Ammonia concentration in the water is likely to be directly proportionate to the
distance downstream of a WWTP discharge. River water most certainly will dilute the
concentration of ammonia in the effluent and as the distance downstream from the release
point increases, ammonia will become far more dilute. This study showed that very low
concentrations of ammonia, 0.06ppm under very specific environmental conditions
known to optimize reproduction (light intensity, temperature, dissolved oxygen content,

27

etc.) can cause adverse reproductive effects in fish as an individual contaminant The
adverse reproductive effects could increase if other environmental stressors were present.
Mobility of FHM within streams has been described and it is generally quite
restricted and only experienced during periods of high discharge (Schlosser 1995). FHM
movement tends to be relatively short in duration (less than 2 days) which suggests that
FHM dwelling near WWTP effluent will not have the ability to escape its exposure.
Seasonal effects may be observed as well because during warmer months, WWTP
effluent concentrations in the river will increase due to lower groundwater inputs and
reduced rain events. This increase in concentrations could be critical to species such as
the FHM as it occurs during their spawning season. FHM spawn when temperatures
reach above 16°C, extending from May and into the middle of August (Starrett 1951).
Hodges (2007) found that during periods of low flow and stream drying, movement of
minnow species into deep pools was observed. These pools create a refuge for small fish;
however can lead to exposures to a variety of environmental stressors. In the real world,
fish are exposed not only to ammonia, but also to mixtures of other chemicals and a
variety of environmental stressors such as fluctuating temperatures and episodes of low
dissolved oxygen. It is unknown how these naturally occurring environmental stressors
will interact with additional inputs of ammonia by WWTP effluent.
This study required extensive time and resources to complete and to fully explore
the effects of toxicants on fish reproduction; such methodology should be complemented
by less labor intensive computational methods. As an example, to conduct this study a
total of 35 testing days were needed to fully understand the toxicity of NH3 on FHM
reproduction. This study also involved many resources, including expensive equipment,

28

coordination with multiple agencies, and the assistance of more than 10 technicians
during the 6 hour long necropsy. Following each necropsy, several hours were required to
analyze samples for vitellogenin concentration. In addition, prior to the initiation of this
experiment, several pilot studies were run using the flow-through diluter system. These
pilot studies were aimed at maximizing control reproduction and consisted of different
feeding regimens, types of food, flow rates, male to female ratios, and light intensity.
This project was extremely cost intensive and therefore it is recommended that a
computational modeling approach should be applied to help predict the effects of
multiple stressors.
Computational models may play a significant role in relating the findings of
studies like these to impacts on wild fish populations. Miller et al. (2007) suggested that
data derived from laboratory studies which analyzed reproductive and survival endpoints
on an individual could be related to the population level as long as a clear biological
linkage could be established. Mechanistic studies regarding ammonia toxicity have not
been thoroughly researched (Wood 2001). In order for computational models to use
fecundity data derived from this study, future studies need to focus on determining a clear
mode of action for NH3 on fish reproduction. The mode of action of NH3 is most likely
linked to increased stress of the fish which leads to increased ventilation, plasma
concentrations of glucose and cortisol, and oxygen consumption (Wood 2001). Miller et
al. (2007) found that when FHM were exposed to various androgens, fecundity was
directly proportional to vitellogenin concentration. As female FHM vitellogenin was
reduced, the fecundity was also reduced. The reduction in vitellogenin was then used to

29

predict population trajectories. This research suggests that models can be used to relate
results of laboratory studies, such as the one conducted here, to populations level effects.
In conclusion, the results of this study suggest that U.S. EPA water quality criteria
be re-evaluated to include recent research focused on both individual contaminants as
well as multiple stressor interactions. Future research should focus on exposing fish to
environmentally relevant concentrations of NH3 while the fish are being exposed to other
environmental stressors in order to determine if a similar reproductive NOEC exists.
Once these questions have been answered, computational models could then be created to
predict population trajectories in order to determine if wild fish populations are at risk.

30

CHAPTER THREE: DETERMINING THE MIXTURE EFFECTS OF NH3 AND
EE2 ON FATHEAD MINNOW REPRODUCTION
Introduction:
Very few studies have tested the effects of multiple stressors on fish reproduction;
most research has been focused on the effects of single contaminants. The U.S. EPA has
used these single contaminant exposure studies to set current water quality criteria.
However, in real world situations, fish are thought to be exposed to a variety of
contaminants; the interactive effects of these contaminants on reproduction are unknown.
For example, WWTP effluent is known to contain multiple contaminants, such as NH3
and EE2 (Nash et al. 2004, Parrott & Blunt 2004, Lange et al. 2000 & U.S. EPA 1999).
Both NH3 and EE2 have been known to reduce FHM reproduction (Chapter 2, Nash et
al. 2004, Parrott & Blunt 2004, Pawlowski et al. 2003, Lange et al. 2000 & Thurston et
al. 1986), but there is little information on how these two chemicals interact within an
individual fish. This study was designed to determine the simultaneous effects of both
NH3 and EE2 on fish reproduction. As more data become available on multiple stressor
relationships, it is my view that the U.S. EPA water quality criteria will need to be
readjusted to incorporate multiple stressor scenarios.
To date, there are no records of any veterinary use of EE2, which suggests that
this chemical is exclusive to human oral contraceptives and hormone replacement therapy
(King County 2007). EE2 is excreted through urine as free EE2 or is conjugated to
increase solubility (e.g. EE2-glucuronide) and undergoes limited degradation in WWTP
(Parrott & Blunt 2005). Bacteria from WWTP metabolize EE2-glucuronide and then
release free EE2 into the water body, thereby, increasing the concentration of free EE2
released from a sewage treatment plant (Parrott & Blunt 2005). EE2 concentrations may
31

also increase over time, causing concerns because of its non-degrading property
(Wolfand 2007).
Surveys of WWTP around the world have detected EE2 concentrations greater
than 1ng/L (Parrott & Blunt 2005) which is significant because previous studies suggest
that an EE2 dose of less than or equal to 1ng/L can decrease FHM male sexual
characteristics, reduce egg fertilization and also reduce reproductive success. Fertilization
success and male secondary sexual characteristics were significantly different in EE2exposed males (<1ng/L) from the controls during a life cycle exposure (Parrott & Blunt
2005). Pawlowski et al. (2004) also found that male secondary sexual characteristics
were significantly different from the controls at 1ng/L EE2 in a 3 week exposure. The
mean number of spawned eggs per breeding pair had a lowest observable effect
concentration (LOEC) of 0.1ng/L EE2. A LOEC is the lowest concentration found by an
experiment or observation which causes an alteration of morphology, functional capacity,
growth, development, or life span of a target organism distinguishable from control
organisms of the same species and strain under defined conditions of the exposure
(McNaught & Wilkinson 1997). Nash et al. (2004) demonstrated that when male and
female zebra fish, Danio rerio, were exposed to concentrations of 0.5ng/L EE2 over a 40
day period, males showed a significant increase in vitellogenin expression, but there were
no observed differences in female vitellogenin expression. Additional studies report that
vitellogenin expression in male FHM was significantly greater than that of control in fish
exposed to concentrations <1.0ng/L of EE2 (Pawlowski et al. 2004).
When exposed to EE2 concentrations higher than 1ng/L, male FHM experienced
alterations of the testis and experienced intersex development. No testicular tissue was

32

found in male FHM exposed to concentrations of EE2 higher than 4ng/L after exposure
for 289 days (Grist et al. 2003). The decrease in the testicular tissue of male fish may be
due to a decrease of plasma sex steroids (Parrott & Blunt 2005) which also occurs when
exposed to EE2. A LOEC for the gonadosomatic index (GSI), which is a ratio of the
gonad weight to the fish’s total weight, for male and female P. promelas was 10 and
100ng/L EE2, respectively (Pawlowski et al. 2004). Parrott & Blunt (2005) reported
female FHM experienced significantly different GSI when exposed to EE2
concentrations between 3.5 and 23ng/L. Kidd et al. (2007) conducted a whole lake
experiment and found when male FHM were exposed to 5-6ng/L EE2 concentrations
they produced significant amounts of vitellogenin and showed signs of intersex through
oocyte development in the testis. After 7 years of exposure, the population of FHM
within the lake was near extinction, thus implying that estrogenic exposures can impact
the sustainability of wild populations.
The uptake mechanism of EE2 of fish is still being studied, but some pathways
have been identified and once in the water column, EE2 is readily absorbed via branchial
absorption (i.e. through the gills). The absorption rate is entirely dependent on the fish’s
gill ventilation volume (Skillman 2006). If a fish is stressed, such as in the case of poor
water quality, the presence of ammonia, or other factors, the ventilation rate of the gills
generally increases leading to a higher absorption rate of EE2. EE2 and other xenobiotics
are diffused across the gill epithelium through lipid membranes (Di Giulio and Hinton
2008). EE2 has an octanol-water partition coefficient (Kow) value of 3.67, meaning that it
has a relatively high sorption potential (King County 2007). Kow is defined as the ratio of
a chemical substance concentration in the octanol phase to its concentration in the

33

aqueous phase. Xenobiotic uptake via branchial absorption is relatively low for toxicants
with a Kow value less than 1. Uptake rates increase fourfold between Kow value 1.0 and
3.0.
EE2, a hydrophobic compound, is relatively insoluble in water and can easily be
absorbed by fatty tissues or membranes. EE2 persists in the organism longer than native
hormones and in turn can activate more hormone receptor sites leading to a higher overall
potency than native hormones. Once it enters the fish via the gills, the EE2 is distributed
into the blood supply where it binds to plasma proteins. The bloodstream then distributes
the protein bound EE2 to the several organs such as the liver, kidney, testes/ovaries, and
brain. These tissues all contain estrogen receptors. Fish have three distinct estrogen
receptors, ERα, ERβ and ERγ (Hawkins et al. 2000). ERα is the principle estrogen
receptor in the liver and is the primary control of estrogen-responsive genes in fish
(Skillman et al. 2006). This estrogen receptor is also expressed in several other tissues
including the gonads (ovaries or testes) and brain (Islinger et al. 2003). EE2 will bind to
these receptor sites and mimic the functions of the naturally produced estrogen. Once the
site is activated, the proliferation of certain cell types (i.e. production of vitellogenin)
begins (King County 2007, Osborne et al. 2007).
Once the hormone EE2 enters the fish via the gills, the bloodstream distributes the
toxicant to the kidney and liver. Weber et al. (2004) reported that kidney tubule cells
were undergoing necrosis leading to nephrotoxicity when fish were exposed to EE2. The
glomeruli within the kidney were enlarged and undergoing a significant extent of cell
death during the EE2 exposure. Damage to the kidney could result in poor excretory and
osmoregulatory functioning (Weber et al. 2004) leading to a lower elimination rate of

34

xenobiotics. It was unknown as to whether or not the nephrotoxicity was a direct
endpoint of exposure to EE2 or an elevated concentration of vitellogenin. Another
finding by Weber et al. (2004) was that the fish were also undergoing hepatotoxicity, i.e.
damage to the liver via necrosis. Since hepatocytes have the ability to regenerate, the
constant replacement of these cells could lead to liver carcinogenesis and a lowered
energy supply for reproduction (Weber et al. 2004).
In addition to disrupting the functions of the kidney and liver, the gonads of fish
are also affected by exposure to levels of EE2. The production of and plasma level
concentration of testosterone and its metabolite, 11-ketotestosterone, is reduced by
exposure to EE2 (Wolfand 2007, Filby et al. 2007). 11-ketotestosterone is associated with
the development of secondary sexual characteristics and essential reproductive behaviors
(Wolfand 2007). EE2 inhibits the genes responsible for coding the enzymes (CYP11A1,
CYP17, 3β-HSD, and cytochrome P450 11B) involved in the production of androgens
(Filby et al. 2007). The ratio of androgens to estrogens produces a ‘male’ versus a
‘female’ hormonal environment and if disrupted can lead to altered sex ratios (Filby et al.
2007).
Another effect of EE2 exposure in fish is the induction of vitellogenin in male
fish. Vitellogenin is normally synthesized in the liver of female fish (Jobling & Tyler
2003) and transferred to the ovaries by circulation to eventually form yolk (Jobling &
Tyler 2003, Islinger et al. 2003). Vitellogenin production is regulated by the native
steroid 17β-estradiol, E2. After production, vitellogenin is biotransformed by
phosphorylation, glycosylation, and lipidation and the metabolites are secreted into the
bloodstream (Parks et al. 1999). Once in the bloodstream, the vitellogenin metabolites are

35

transported to the oocyte surface where they are seized by receptor-mediated endocytosis.
Enzymes then break down the vitellogenin metabolites into egg yolk proteins (phosvitin
and vitellin) and lipids as food for the developing embryo (Parks et al. 1999). When
exposed to EE2, male fish will produce vitellogenin (Jobling & Tyler 2003). Therefore, a
good indication of exposure to EE2 is concentrations of vitellogenin in the blood of male
fish (Jobling & Tyler 2003). Another advantage of using vitellogenin and other
biomarker data is that responses at lower levels of organization can be much more
specific to modes or mechanisms of action (MOA) (Miller et al. 2007).
For several fish species, glucuronidation is the major pathway for inactivating and
eliminating substances via the bile (Skillman et al. 2006). EE2 undergoes glucuronide
conjugation and secretion into bile, which is similar to EE2 metabolism in mammals. The
highest concentration of EE2 is found in the bile as glucuronide conjugates.
Glucuronidation does not appear to significantly facilitate excretion of EE2 (Skillman et
al. 2006). EE2 undergoes enterohepatic recirculation in some fish species such as trout,
where the EE2-glucuronide conjugates in the gall balder are released into the gut and
reabsorbed into the intestinal tract. This process increases the half-life of EE2 within the
organism (Schultz et al. 2001) and allows for a re-dosing of the organism. After
reabsorption into the gastrointestinal tract, the EE2-glucuronide conjugates then undergo
deconjugation and are reabsorbed into the bloodstream or excreted via feces (Schultz et
al. 2001). This recirculation allows the EE2 to become concentrated in the blood plasma,
up to nearly 780 fold relative to the exposure water (Skillman et al. 2006) Extensive
enterohepatic recirculation and mass balance studies suggest that most of the excretion of
EE2 is the unconjugated form (Skillman et al. 2006).

36

The NH3 baseline experiment (Chapter 2) revealed a couple of significant
findings regarding NH3 toxicity on FHM reproduction. All three concentrations tested in
this experiment (0.06, 0.18 and 0.42mg/L NH3) significantly reduced fecundity (Chapter
2). The most important finding from this study was that the estimated 0.03mg/L NH3
NOEC for FHM fecundity is far below previously reported values. Thurston et al. (1986)
reported the CETC for FHM egg production as 0.58mg/L. The LOEC determined by this
study (0.06mg/L NH3) is also below current U.S. EPA water quality criteria. NH3 also
appeared to reduce the induction of vitellogenin in males exposed to natural estrogens
being produced by female fish.
One mechanism by which EE2 and NH3 can interact synergistically is by
decreasing egg production in FHM. Individually, these chemicals both inhibit
reproduction in fish (Thurston et al. 1986, Parrott & Blunt 2005, Lange et al. 2001,
Jensen et al. 2001, Chapter 2). A fish’s stress response includes an increased ventilation
rate in order to increase the oxygen uptake rate of the gills (Bonga 1997). Additional
stressors such as ammonia may increase the uptake of EE2 due to this increased
ventilation rate; thus these two contaminants may act synergistically by increasing a
fish’s sensitivity to the other. Lower concentrations of EE2 may have a higher absorption
rate due to increased ventilation when in the presence of NH3.
In addition to acting synergistically with one another, NH3 and EE2 may act
antagonistically on other endpoints such as vitellogenin production. Stress can also
induce the formation of cortisol, which can decrease the production of vitellogenin in fish
(Giesy et al. 2003) and potentially act antagonistically from EE2 effects. High
concentrations of toxic substances such as ammonia have been shown to impair

37

vitellogenesis in fish (Ma et al. 2005). Due to this antagonistic relationship, vitellogenin
production in a male fish may not be initiated during an EE2 exposure when in the
presence of NH3.
The main hypothesis of this experiment was that the effects of EE2 on FHM
reproduction will experience a synergistic relationship when in the presence of ammonia.
I tested this hypothesis using the flow-through diluter system (Appendix A) by testing the
individual reproductive effects of both NH3 and EE2 at their respective estimated NOEC
as well as the mixture of the two contaminants. A follow up hypothesis is that EE2 and
NH3 will act antagonistically on vitellogenin production in FHM males, and I tested this
hypothesis by analyzing both male and female FHM plasma and liver vitellogenin
concentrations from the diluter system experiment.
Data from a previous EE2 baseline study conducted in 2009 at the U.S. EPA
facility located in Cincinnati, Ohio was used to determine the testing concentration of
EE2 to P. promelas (Figure 8).

# Eggs produced per
female per day

16
12
8
4
0
EE2 Concentration
Control
0.5ng/L
1.5ng/L
4.5ng/L
Figure 8: 21-Day Treatment Flow-through Diluter Egg
Production/Female/Series/Day. Sample Size (N) = 28. Conducted by U.S. EPA
in Cincinnati, OH. Treatments included control, 0.5ng/L, 1.5ng/L and 4.5ng/L
17α-ethinylestradiol (EE2). Error bars are ± 1 standard error. Study duration
was 21 days.
38

The EE2 study used the same diluter system and methodology as this project (Appendix
A). A 0.33 dilution factor was used in determining the EE2 concentrations tested during
the EE2 baseline study. This dilution factor was chosen because we needed to establish
an EE2 NOEC for FHM egg production using this system. The low concentration,
0.5ng/L EE2, was chosen since it was the lowest concentration that could be analytically
measured by the U.S. EPA’s chemistry laboratory in Cincinnati, OH. Other flow-through
diluter studies chose concentrations such that the lowest concentration was 10 times
lower than the highest concentration. The 0.33 dilution factor satisfied both of these
requirements. Three concentrations of EE2 were tested, 0.5ng/L, 1.5ng/L, and 4.5ng/L.
Although recent studies have shown that concentrations as low as 0.1ng/L activate the
production of vitellogenin in adult male FHM, this low concentration could not be tested
because the U.S. EPA in Cincinnati, Ohio’s chemistry methods require 0.2ng/L error.
Methods:
This experiment’s purpose was to test the hypothesis that FHM will be more
sensitive to EE2 when also exposed to NH3. Environmentally relevant concentrations of
NH3 and EE2 that were at or near the NOEC for FHM reproduction as determined from
the baseline studies were tested in order to mimic those concentrations found in an
average WWTP effluent. By testing the two toxicants at their respective NOEC, any
effect found in their mixture could be attributed to a multiple stressor relationship.
A control, 0.25ng/L EE2 concentration, 2.5mg/L total ammonia (0.03mg/L NH3),
and a mixture of both the EE2 and NH3 concentrations were tested during this
experiment. The 0.25ng/L EE2 concentration was chosen because it is near the average
concentration found in WWTP effluent and is near the LOEC on FHM reproduction and

39

also can stimulate the production of vitellogenin in male FHM. The EE2 baseline study
found that after a 21 day exposure of FHM to 0.5ng/L EE2, reproduction was
significantly reduced (Figure 8) and the induction of vitellogenin in EE2 exposed males
was significantly greater than control males (Figure 9).
5.E-04

B

*

Relative expression (vtg / 18S)

Relative expression (vtg / 18S)

A

4.E-04
3.E-04
*

2.E-04
1.E-04
0.E+00
Control

EE2 Concentration
0.5 ng/L 1.5 ng/L 4.5 ng/L

4.0E-06
3.2E-06
2.4E-06
1.6E-06
8.0E-07
0.0E+00

EE2 Concentration

Control

0.5 ng/L

Figure 9: Average male vitellogenin liver concentrations. Sample Size (N) = 56.
Conducted by U.S. EPA in Cincinnati, OH. Treatments included control, 0.5ng/L,
1.5ng/L and 4.5ng/L 17α-ethinylestradiol (EE2). Error bars are ± 1 standard error. Study
duration was 21 days. A) Control versus 0.5, 1.5 and 4.5ng/L EE2. B) Control versus
0.5ng/L EE2.

The 0.03mg/L NH3 concentration was chosen because it was the estimated NOEC for
NH3 on FHM reproduction per the NH3 baseline experiment results of this study
(Chapter 2).
Previous laboratory studies that conducted exposures in static systems reported
that EE2 breaks down over time resulting in nominal concentrations that are close to
0.0mg/L. Because the flow-through diluter system provided a constant renewal of testing
solution, the breakdown of EE2 had a minimal influence on my nominal concentrations.
This constant renewal also eliminated the need to use dimethyl sulfoxide, DMSO, as a
40

carrier solvent. This was important because the potential adverse effects of DMSO on
fish had not clearly been determined at the time of the study and could have resulted in
false data interpretations after a 21 day exposure.
As performed in the NH3 experiment (Chapter 2), water chemistries were
monitored several times each week throughout the study to ensure proper toxicant
exposure. The NH3 baseline experiment determined that collecting individual tank
ammonia measurements was unnecessary as all tank NH3 concentrations were similar to
the composite concentration. Therefore only a total ammonia concentration was
measured from a composite of all 7 replicates within a treatment group three times each
week. Temperature and pH readings for each composite were recorded soon after a total
ammonia measurement was performed. EE2 analysis was performed twice each week.
One (1) liter samples from a composite of each treatment and their respective superstock
were sent to the chemistry department at the U.S. EPA in Cincinnati, OH for analysis.
The 0.25ng/L EE2 concentration was below detection limits using the U.S. EPA method;
therefore two (2) liter samples of each testing chamber treatment were analyzed. The
superstocks were measured to ensure that the flow-through diluter pumps were accurately
and precisely mixing each treatment throughout the study.
The diluter system required daily maintenance to ensure proper dosing during the
21 day exposure period. Fresh superstocks were mixed daily in order to provide the
diluter system with a constant source of toxicant. The 2.5mg/L total ammonia (0.03mg/L
NH3) superstock was made by mixing 4g of NH4Cl into 18L of laboratory line water. A
1.7mg/L EE2 dosing stock was provided by the U.S. EPA chemistry group for making
the EE2 superstock. This superstock was made by adding 79.4µL of dosing stock to 18L

41

of laboratory line water. The pumps were calibrated three times each week to ensure
proper mixing of toxicant and delivery to the testing chambers. The tubes for the
peristaltic pumps were replaced weekly to prevent any breakage or tube malfunctioning.
Results:
This experiment met mortality criteria in the controls (Appendix B) as total
mortality was 4.8% (95.2% survival) over the 21 day study. Male mortality was 1, 4, 3,
and 6 for the control, 0.25ng/L EE2, 0.03mg/L NH3 and mixture treatments, respectively.
Female mortality was 1, 1, 2 and 6 for each of the treatments, respectively. The mortality
data were not normally distributed; therefore, a nonparametric Kruskal-Wallis test was
performed. Total mortality was found to be significantly different amongst treatments (K
= 7.972, p = 0.047, df = 3). The mixture (EE2 and NH3) treatment was significantly
higher than the control for total mortality as determined by the Kruskal-Wallis multiple
comparison Post hoc test (kruskalmc) in Program R (p = 0.05).
To account for female mortality in the treatments, egg production was
standardized by taking the total number of eggs produced within each treatment and
dividing by the total number of surviving females. Over the course of the 21 day
exposure period the average eggs produced per female per day in the control, 0.25ng/L
EE2, 2.5mg/L total ammonia and the mixture were 19.79 ± 2.82, 23.49 ± 1.89, 22.47 ±
2.48 and 21.02 ± 3.37 eggs per female per day, respectively (Figure 10).

42

# Eggs produced per
female per day

30
25
20
15
10
5
0
Control

Concentration
EE2
NH3

Mixture

Figure 10: 21-Day Treatment Flow-through Diluter Egg
Production/Female/Series/Day. Sample Size (N) = 28. Conducted 4/8/10. Unionized
Ammonia (NH3) and 17α-ethinylestradiol (EE2) Interaction Experiment. Treatments
included control, 0.25 ng/L EE2, 0.03mg/L NH3 and a mixture of both toxicants.
Error bars are ± 1 standard error. Study duration was 21 days.
Because female mortality occurred during the 21 day exposure period, the egg production
per female data set was chosen for analysis instead of cumulative egg production. Egg
production per female per day was normally distributed and verified by a non-significant
Shapiro-Wilks test for normality (W=0.9731 p = 0.076). Variances amongst treatments
were similar as verified by a non-significant Levene’s test for homogeneity of variance
(F=2.2264, p = 0.09147, df = 3). There were no differences in egg production per female
between treatments (ANOVA, F=0.363, p = 0.7799, df = 3).
Fertility rates during the 21 day exposure period in the control, 0.25ng/L EE2,
2.5mg/L total ammonia and the mixture were 89.08%, 88.87%, 88.14%, and 82.08%,
respectively (Figure 11). Fertilization rates were normally distributed and verified via a
Shapiro Wilks test for normality (W = 0.9414, p = 0.1202).

43

Fertilization Rate

100%
75%
50%
25%
0%
Control

Concentration
EE2
NH3

Mixture

Figure 11: 21-Day Treatment Flow-through Diluter Concentration Fertilization Rate.
Sample Size (N) = 7. Conducted 4/8/10. Unionized Ammonia (NH3) and 17αethinylestradiol (EE2) Interaction Experiment. Treatments included control, 0.25
ng/L EE2, 0.03mg/L NH3 and a mixture of both toxicants. Error bars are ± 1
standard error. Study duration was 21 days.

Differences in variances among treatments were negligible as verified by a nonsignificant Levene’s test (F = 0.7773, p = 0.5181, df = 3). Fertility rates were also not
significantly different amongst treatments as determined by an ANOVA (F = 0.5252, p =
0.67, df = 3).
Seven sampling events occurred throughout the 21 day exposure to verify EE2
exposure (Figure 12). The control and NH3 tanks did not have any significant trace
concentrations of EE2 and because the 0.25ng/L EE2 concentration is lower than the
chemists’ detection limits (+/- 0.2ng/L error) only the superstock concentrations were
reported. The superstock (7.5ng/L EE2) was 30x the testing concentration in the tank
(0.25ng/L EE2). The 0.25ng/L EE2 concentration and mixture concentration had an
average superstock EE2 concentration of 7.27ng/L and 8.04ng/L, respectively.

44

Measured EE2 (ng/L)

12
9
6
3
0
0

2
EE2 Measured

4
Sample
Mix Measured

6

8
Expected

Figure 12: Analytically measured 17α-ethinylestradiol (EE2) concentrations from the
superstock for the 0.25ng/L and mixture concentrations. Conducted 8/26/09. EE2 and
Unionized Ammonia (NH3) Interaction Experiment.
Sampling event #4 for the 0.25ng/L EE2 concentration was not included in the analysis
as it was significantly lower than expected and may be due to an analysis error. The
coefficient of variation values were below the required 20% test acceptability limit
(Ankley et al. 2001) as the EE2 and mixture superstock concentration CV values were
7.71% and 8.13%, respectively.
Ten sampling events for ammonia occurred throughout the 21 day study. The
control and EE2 tanks did not have significant trace concentrations of NH3 (Figure 13).
The 2.5mg/L total ammonia treatment tanks ranged from 1.96 to 2.47mg/L total ammonia
with an average of 2.26mg/L total ammonia (0.03mg/L un-ionized ammonia). The
mixture tanks ranged from 1.58 to 2.51mg/L total ammonia with an average of 2.18mg/L
total ammonia (0.03mg/L un-ionized ammonia). Mean measured concentrations
throughout the study in the treatment tanks appeared sufficiently similar to expected

45

concentrations and the coefficient of variance (CV) was within acceptable parameters

Unionized Ammonia
(mg/L)

(<20%) for both treatments containing ammonia.
0.04
0.03
0.02
0.01
0.00

Concentration
Control
EE2
NH3
Mixture
Figure 13: Average unionized ammonia (NH3) concentrations from each test
concentration. Conducted 4/8/10. NH3 and 17α-ethinylestradiol (EE2) Interaction
Experiment. Treatments included control, 0.25 ng/L EE2, 0.03mg/L NH3 and a
mixture of both toxicants. Error bars are ± 1 standard error.
Chemical parameters including dissolved oxygen (ANOVA, F = 0.949, p = 0.425, df =
3), pH (ANOVA, F = 0.417, p = 0.741, df = 3), temperature (ANOVA, F = 0.281, p =
0.839, df = 3), and conductivity (ANOVA, F = 0.210, p = 0.889, df = 3) were determined
for the old test solutions and showed no significant differences between any treatments.
Vitellogenin concentrations were analyzed on the male and female livers (Figures
14 and 15). Male vitellogenin concentration was log transformed in order to obtain
normality (Shapiro-Wilks, W=0.9759, p = 0.4939). Variances amongst treatments were
similar as verified by a non-significant Levene’s test for homogeneity of variance
(F=0.691, p = 0.5631, df = 3). Male liver vitellogenin concentration was not significantly
different between treatments (F = 1.844, p = 0.155, df = 3). Female liver vitellogenin
concentrations were not affected by exposure to either contaminant or their mixture (K =
2.09, p = 0.55, df = 3). This data was not normally distributed; therefore a Kruskal Wallis

46

test was used in the analysis. Again, no transformations were able to normalize the
female liver vitellogenin concentrations data set.

Relative expression
Vtg / 18S

8.E-03
6.E-03
4.E-03
2.E-03
0.E+00
Concentration
Control

EE2

NH3

Mixture

Figure 14: Average male vitellogenin liver concentrations. Sample Size (N) = 56.
Conducted 4/8/10. Unionized Ammonia (NH3) and 17α-ethinylestradiol (EE2)
Interaction Experiment. Treatments included control, 0.25 ng/L EE2, 0.03mg/L NH3
and a mixture of both toxicants. Error bars are ± 1 standard error. Study duration was
21 days.

Relative expression
Vtg / 18S

50
40
30
20
10
0
Control

Concentration
EE2
NH3

Mixture

Figure 15: Average female vitellogenin liver concentrations. Sample Size (N) = 112.
Conducted 4/8/10. Unionized Ammonia (NH3) and 17α-ethinylestradiol (EE2)
Interaction Experiment. Treatments included control, 0.25 ng/L EE2, 0.03mg/L NH3
and a mixture of both toxicants. Error bars are ± 1 standard error. Study duration was
21 days.
Plasma vitellogenin was negligible amongst all treatments in male FHM (Figure 16). The
data were normally distributed, therefore a Kruskal-Wallis test was used to determine the

47

non-significance (K = 5.72, p = 0.12, df = 3). One control male outlier displayed a high

Plasma Vitellogenin
(mg/ml)

amount of plasma vitellogenin and was removed from the analysis (Figure 17).
0.04
0.03
0.02
0.01
0.00
Concentration
Control

EE2

NH3

Mixture

Plasma Vitellogenin
(mg/ml)

Figure 16: Average male plasma vitellogenin concentrations. Sample Size (N) = 56.
Conducted 4/8/2010, data provided by U.S. EPA in Duluth, MN. Unionized Ammonia
(NH3) and 17α-ethinylestradiol (EE2) Interaction Experiment. Treatments included
control, 0.25 ng/L EE2, 0.03mg/L NH3 and a mixture of both toxicants. Error bars are
± 1 standard error. Study duration was 21 days.

2.0E-03
1.6E-03
1.2E-03
8.0E-04
4.0E-04
0.0E+00
Concentration
Control

EE2

NH3

Mixture

Figure 17: Average male plasma vitellogenin concentrations with control outlier
removed. Sample Size (N) = 55. Conducted 4/8/2010, data provided by U.S. EPA in
Duluth, MN. Unionized Ammonia (NH3) and 17α-ethinylestradiol (EE2) Interaction
Experiment. Treatments included control, 0.25 ng/L EE2, 0.03mg/L NH3 and a
mixture of both toxicants. Error bars are ± 1 standard error. Study duration was 21
days.
This non-response may be due to the concentrations of EE2 being below minimum
detection limits by the ELISA assay. Female plasma vitellogenin (Figure 18) was also not

48

effected by exposure to any of the treatments (K = 3.23, p = 0.36, df = 3). The data was
also non-normally distributed and analyzed using a Kruskal-Wallis test.

Plasma Vitellogenin
(mg/ml)

25
20
15
10
5
0
Concentration
Control

EE2

NH3

Mixture

Figure 18: Average female plasma vitellogenin concentrations. Sample Size (N) = 112.
Conducted 4/8/2010, data provided by U.S. EPA in Duluth, MN. Unionized Ammonia
(NH3) and 17α-ethinylestradiol (EE2) Interaction Experiment. Treatments included
control, 0.25 ng/L EE2, 0.03mg/L NH3 and a mixture of both toxicants. Error bars are ±
1 standard error. Study duration was 21 days.
Discussion:
Neither EE2 nor NH3 caused a fecundity effect on FHM reproduction
individually. This was expected as both toxicants were being tested at an estimated
NOEC. Also, the mixture of these two contaminants did not affect fecundity during the
21 day study. This finding rejects the author’s hypothesis that the mixture of these
contaminants at their respective NOEC would produce an effect on FHM reproduction.
Although a fecundity effect was not apparent, a significant mortality effect was observed
in the mixture concentration. This finding is surprising because the concentrations of both
EE2 and NH3 were far below any previous reported lethality concentration. Thurston et
al. (1986) reported a LOEC of 0.91mg/L NH3 on FHM survival. Lange et al. (2001)
determined the LOEC of EE2 on adult FHM survival was greater than 1ng/L. The tested

49

concentrations were chosen to mimic the average WWTP effluent NH3 and EE2
concentrations, so this mortality effect is quite alarming for small short-lived species such
as the fathead minnow. In natural systems, Lazorchak & Smith (2004) determined that
the average NH3 concentration in U.S. WWTP effluent was around 0.1mg/L. Parrott &
Blunt (2005) also stated that EE2 concentrations greater than 1ng/L are found in WWTP
effluent around the world. Therefore, such a high mortality rate in a mixture of only two
contaminants at their environmentally relevant concentrations is a concern. In light of
these findings, the current U.S. EPA water quality criteria should be re-evaluated, and
other stressors should be considered because fish are exposed to much greater numbers of
contaminants as well as other biological stressors including temperature changes,
dissolved oxygen fluctuations, starvation and predation.
The multiple stressor effect on mortality has been documented in other species
exposed to different stressors. Schisler & Bergersen (2000) examined the combined
effects of disease and temperature on rainbow trout, Oncorhynchus mykiss. Elevated
water temperatures on their own had minimal effect on mortality of trout; however,
mortality was significantly increased in trout infected with whirling disease, Myxobolus
cerabralis.
The mortality that was experienced during this mixture study could possibly be
attributed to a high reproductive output. It is important to point out that the control
reproductive success was much higher during this study than any previous flow-through
diluter experiment conducted by the U.S. EPA in Cincinnati, OH. This increased
reproduction may have been a cause for the unexpected mortality in the mixture
concentration because the fish could have spent more energy on reproduction than other

50

similar aged and sized FHM used in previous flow-through diluter experiments; the
added stress of EE2 and NH3 in combination may have been enough to cause mortality.
Females could have used more energy producing and laying eggs while the males may
have spent more energy courting females and guarding the eggs. This experiment should
be repeated to make sure the mortality effect is consistent. Gunderson (1997) determined
that there was a tradeoff between reproductive effort and adult survival in fish species.
The NH3 baseline study also had a high reproductive effort experienced by the control
fish when compared to previous flow-through diluter studies; however mortality was not
an issue during this study (Chapter 2). If further testing proves the mortality experienced
in the mixture study to be a justifiable effect, continued output of WWTP effluent
containing concentrations similar to those tested in this study could be causing
detrimental effects on wild FHM populations, as well as other fish species.
The mixture study tested a 0.25ng/L 17α-ethinylestradiol (EE2) and a 0.03mg/L
NH3 concentration. The EE2 sample analysis revealed a consistent dosage of superstock
throughout the study except for one sample collection, sample #4 (Figure 12). This was
deemed as an outlier and was probably due to an error in sample collection or analysis.
The EE2 analysis for the 0.25ng/L testing chambers was below the U.S. EPA chemistry
group’s method detection limits (± 0.2ng/L EE2) and therefore was not considered
reliable data. Because the EE2 superstock was within the 60 to 80% nominal test
acceptability (Grist et al. 2003) and the 0.03mg/L NH3 concentration in the tanks was
also near expected concentrations, a reasonable assumption can be made that EE2 within
the tanks was also near expected concentrations.

51

Another disadvantage of using a flow-through diluter system is that EE2
chemistry data of testing solution is extremely hard to distinguish from background levels
of natural estrogens that are produced by the fish. The U.S. EPA has established methods
to determine the nominal testing concentrations of EE2 (U.S. EPA 2007). Sixty to eighty
percent nominal is acceptable in most studies (Grist et al. 2003). The mixture study
produced better than most studies in EE2 analysis as a 96.8% nominal concentration in
the EE2 superstock was determined. The 0.25ng/L EE2 concentration in the tanks could
not be determined as it is below the method detection limit for the U.S. EPA chemistry
group in Cincinnati, OH.
This pilot study suggested an underlying mechanism regulating the impact of
ammonia exposure on the induction of vitellogenin in male FHM. Ammonia may be
stimulating the production of cortisol, which has been known to decrease the production
of vitellogenin in fish (Giesy et al. 2003) and potentially act antagonistically from EE2
effects. Another hypothesis may be that the males are shutting down unnecessary
mechanisms, such as the production of vitellogenin, in response to the environmental
stressor. Further investigation into the MOA of ammonia on vitellogenin production
should also be conducted in order to fully understand why the antagonism exists. By
understanding these mechanisms at which vitellogenin production is slowed due to
contaminant exposure, computational models could be created in order to predict the
effects of other contaminants on vitellogenin production, and eliminate the need to test
every multiple stressor combination in a costly experimental design.
The protocol used in this experiment was successful in determining multiple
stressor responses to anthropogenic and natural stressors. For example, a mortality effect

52

was not significant in either individual FHM exposure to NH3 or EE2, but a significant
effect was found in the exposure to their mixture. EE2 appeared to increase vitellogenin
concentration in the livers of male FHM but when EE2 was mixed with NH3, the
production of vitellogenin was blocked. These findings indicate the significance of
testing with multiple stressors rather than single contaminant exposures and how
protocols such as the 21 day flow-through diluter study can be used to test the impacts of
multiple stressors. Importantly, the procedures outlined here were designed to test both
the individual effect of both toxicants and a mixture of both contaminants, which is
necessary to determine synergistic and antagonistic effects.
The findings from this study suggest that multiple contaminant exposures should
be tested and used for ecological risk assessments rather than data derived from single
contaminant dose response studies. The possibility of chemical interactions within a fish
between EDCs, other stressors and mating systems could have a population-level effect
through reduced fertility, fecundity, or mortality. Multiple stressors that manipulate
reproductive activity and male secondary sexual characteristics may strongly affect
sexual selection and mate choice in fish (Nash et al. 2004). This study tested only two out
of potentially hundreds of interacting chemicals found in WWTP effluent. WWTP
effluent contains a cocktail of chemicals that could have numerous reproductive effects
on fish populations. Therefore, understanding how the chemicals within WWTP effluent
interact with one another in terms of reproductive effects on fish and other aquatic
species should be a high priority for future toxicological studies.
Because the field of testing with multiple stressors is relatively new, very little
data exists regarding how different estrogenic contaminants interact with one another and

53

therefore it is hard to predict the outcomes of multiple stressor studies. Therefore, more
studies testing a variety of other EDC’s should be conducted in multiple stressor
situations in order to fully understand how estrogenic contaminants interact and affect
fish reproduction, especially at environmentally relevant concentrations. Estrogen mimics
can occur without chemical exposures due to sub-optimal temperatures, restricted food
supply, low pH, environmental pollutants, and/or parasites (Jobling & Tyler 2003). Most
multiple stressor studies have analyzed the effects of two natural stressors such as
predation and nutrition while very few studies have researched combinations of natural
and anthropogenic stressor relationships (Folt et al. 1999). These experiments have
shown that the combination of these two stressor types can cause detrimental effects at a
level much lower than individual chemical exposures. Testing estrogenic contaminants
with other stressor situations such as low dissolved oxygen concentrations and high or
low temperatures should also be considered as fish may be exposed to these
environmental situations as well.
Hanazato and Dodson (1995) tested the effects of low oxygen concentration when
coupled with the effects of a predatory response and a pesticide exposure. Synergistic
relationships were determined when a cladoceran, Daphnia pulex, was exposed to a
combination of 2 or more of these stressors. Growth rates, neonatal development and
maturation size were all reduced when a combination of these stressors were present.
These findings

determined that

the synergistic relationships

found

between

environmental factors and chemical exposures are often underestimated in laboratory
tests looking at the effects of single contaminant exposures. Most laboratory tests are held
at high dissolved oxygen concentrations and their study determined that effects of

54

chemical exposures can be amplified when environmental stressors such as low dissolved
oxygen concentrations are present.
Short-lived species such as FHM have a greater risk of population collapse when
exposed to EDCs, but chronic exposures of longer-lived species may result in extinction
as well (Kidd et al. 2007). FHM and other short-lived species are vital to many aquatic
food webs as forage fish (Jenkins & Burkhead 1993). If FHM populations, as well as
other small, short-lived species, are threatened and/or reduced, larger and longer-lived
species could be at risk due to a decreased food supply. Long-term and lifetime exposures
to multiple estrogenic contaminants could also put larger species at risk due to potential
endocrine disruption and reduced reproduction success. Long-lived species, such as
sturgeon, become sexually mature at an old age compared to short-lived species. The
longer lived fish also have shorter spawning seasons and in some cases can have
multiyear intervals between spawning events (Gross et al. 2002). These species may be
more susceptible to population collapse if their reproductive cycle is adversely altered
due to a lower reproductive output. Reproductive effects, in particular the
intergenerational effects on fertility, are difficult to assess in long-lived species due to
their low reproductive output and long generation times compared to short-lived species.
Understanding how multiple stressor exposures affect the reproductive success of fish
populations is important because continued estrogenic contaminant inputs into the
environment from municipal wastewaters could threaten their sustainability (Kidd et al.
2007).

55

CHAPTER FOUR: CONCLUSIONS AND FUTURE DIRECTION
Expectations for this study were to define the multiple stressor effects of EE2 and
NH3 on FHM reproduction during a 21 day experiment using a flow-through diluter
system. Because the testing of multiple stressors is a relatively new area of study and
very few data exist on how chemicals interact during exposure studies, this study served
as an exploratory one to provide preliminary data for future multiple stressor experiments
as well as to update un-ionized ammonia toxicity relative to FHM reproduction.
The NH3 baseline experiment suggested that U.S. EPA water quality criteria for
NH3 may currently be set above the NOEC on fish reproduction. U.S. EPA water quality
criteria for waters containing adult, non-salmonid fish is currently set at 0.094mg/L (at a
pH of 7.3 and 25°C temperature similar to the parameters of the dilution water tested in
this study (Chapter 2)). This concentration is over 56% higher than the LOEC determined
by this study (0.06mg/L) in which I found a 29% reduction in cumulative fecundity. A
reduction this severe, in addition to other threats to minnow populations such as other
contaminant exposures and predation, has the potential to threaten population survival. If
fish in the wild are being exposed to concentrations of ammonia high enough to cause
severe reproductive effects, theoretically it is possible that entire populations within a
water body are at risk.
Density dependent factors play an important role in population regulation and should be
considered when determining population risk due to a contaminant. If fish within a healthy water
body are producing more eggs than the environment can support, density dependent factors such
as competition will limit population growth. It is because of density dependence that a reduction
in fecundity as a result of contaminant exposure may not necessarily lead to a direct population
collapse. For example, if a population of fish produces 1500 eggs in a given year and the

56

population can only support an increase of 1000 individuals, density dependent factors
will result in 500 of those eggs from reaching maturity in order to maintain the
population at its carrying capacity. If that same population is exposed to a contaminant
that reduces the fecundity from 1500 to 1000 eggs, there still will be 1000 new
individuals reaching maturity within that population; hence the contaminant will not have
an effect on the number of individuals within that population. It is even possible that the
loss in fecundity will produce healthier fish because competition for food availability,
sexual mate selection and territorial defense may be reduced. However, it is my view that
exposure to a known contaminant which causes reproductive effects in fish should not be
considered as potentially beneficial because other detrimental factors resulting from the
exposure are likely to occur.
Extrapolating the results of this study which used laboratory fish to individuals
found in the wild can be a difficult task at best. The FHM at the U.S. EPA AWBERC
facility in Cincinnati, OH have been raised for hundreds of generations in conditions
specific to increasing survival and reproduction. The genetic diversity in these fish is
minimized because no new genetic sources of FHM have been introduced to this
laboratory in years. These fish are not exposed to a variety of stressors seen in wild
populations including predation, foraging, low water quality and high and low
temperature situations. Wild fish face different evolutionary pressures and also have a
much higher genetic diversity than laboratory raised fish because new genes are
constantly being added to the gene pool because fish either migrate or are introduced
from different water bodies.

57

Extrapolating this data to other species of fish with longer lifespans and lower
reproduction can be even more difficult. Long-lived fish tend to have shorter spawning
seasons and in some cases can have multiyear intervals between spawning events (Gross
et al. 2002). These species may be more susceptible to contaminant exposures as their
reproductive output is relatively low compared to short-lived species such as FHM.
Reproductive effects of contaminant exposures using long-lived fish are difficult to
implement due to the long maturation rate and short and intermittent spawning seasons.
Therefore, mechanistic models using laboratory data collected from short-lived fish can
be used to predict the exposure response of long-lived species.
The mixture study verified that the estimated NOEC values as determined from
the respective baseline experiments were justified. Neither the 0.03mg/L NH3 nor
0.25ng/L EE2 concentrations caused a FHM reproductive effect. The mixture of these
two contaminants also did not cause a reproductive effect. This non-response nullifies the
hypothesis that the mixture of these two contaminants at their respective environmentally
relevant concentration would result in a reproductive effect. The other hypothesis from
this study was that vitellogenin production in male FHM would not be induced during
exposure to EE2. This hypothesis was verified as the EE2 exposed males showed a slight
increase in vitellogenin while the mixture exposed males showed virtually no vitellogenin
in either the plasma or liver. This inhibition of vitellogenin production in the males is
most likely influenced by the increased stress and production of cortisol which has been
shown to reduce vitellogenin production. Another interesting finding that should be
explored further is why there wasn’t a reduction of vitellogenin production in the females
exposed to NH3.

58

My experiment found significant and alarming evidence that the mixture of EE2
and NH3 significantly increased both male and female mortality. The control treatment
only experienced one male and/or female mortality over the course of the 21 day study.
In contrast, the mortality increased by 600% for both males and females in the mixture
concentration where 6 males and 6 females died (mortality was experienced in 5 of the 7
replicates) during the 21 day study. Twenty-eight percent (28%) of the total fish exposed
to the mixture concentration experienced mortality. This is a significant increase in
mortality over a relatively short period. The average WWTP effluent contains a 0.1mg/L
NH3 and greater than 1.0ng/L EE2 concentration. Previous lethality concentrations for
NH3 and EE2 have been reported as 0.91mg/L and greater than 1ng/L, respectively.
These previous estimates are far above the concentrations tested in the mixture
(0.03mg/L NH3 and 0.25ng/L EE2). It is recommended that this study be repeated to see
if a similar mortality effect occurs. If further studies determine that a mortality effect at
these environmentally relevant concentrations of NH3 and EE2 exists, it is recommended
that the current U.S. EPA water quality criteria be re-visited and possibly amended to
incorporate this significant finding.
The mixture study tested environmentally relevant concentrations of EE2 and
NH3. Surveys of WWTP effluent have found concentrations of NH3 around 0.1mg/L and
EE2 greater than 1ng/L. This study tested 0.03mg/L NH3 and 0.25ng/L EE2. Therefore it
is assumed that fish in the wild are being exposed to a mixture of these contaminants at or
above the concentrations tested in this study in areas that are heavily impacted by WWTP
effluent. The only way to avoid exposure of these chemicals is through more advanced

59

treatment of WWTP effluent and educating the public on proper disposal of unused or
unwanted medication.
The removal of ammonia from wastewater treatment effluent has been thoroughly
researched and many types of treatments are available. The traditional method for
ammonia removal is through biological treatments such as nitrifying bacteria (Jorgensen
& Weatherley 2003). Many biological methods require warm temperatures to sufficiently
remove ammonia. Gersberg et al. (1986) found that planting vegetation such as bulrushes
and reeds in the wastewater treatment ponds can play a significant role in nitrogen
removal. While vegetation may be successful in removing ammonia, during the winter
months this method may not be as efficient. Ion exchange is a method that has proven to
be successful in removing ammonia at a wide range of temperatures (Jorgensen &
Weatherley 2003). Ion exchangers such as zeolite were found to be extremely successful
in removal of ammonia in wastewater effluent even during high spikes in ammonia
concentration, another variable that biological removal methods cannot handle (Jorgensen
& Weatherley 2003).
Because EE2 is the main ingredient in many of today’s oral contraceptives and
will most-likely always be a part of human birth control and hormone replacement
therapy, the control of EE2’s release into the environment must take place at the WWTP.
Vader et al. (2000) determined that nitrifying activated sludge in the presence of
ammonia was able to oxidize EE2 into hydrophilic compounds. The authors reported that
this sludge degraded a 50µg/L concentration of EE2 within 6 days. The enzyme,
ammonium monooxygenase, initiated the pathway of ammonium oxidation. While
oxidizing ammonia, the bacteria were also reported to cometabolize EE2 (Vader et al.

60

2000, Shi et al. 2004). It is important to note that in the presence of an ammonia
oxidation inhibitor, such as the compound Allylthiourea, oxidation was unable to occur
thus limiting the nitrifying bacteria sludge’s potential to degrade EE2 (Shi et al. 2004).
EE2 hydroxylated metabolites are known to have a substantially lesser estrogenic effect
on fish than the EE2 parent compound (Vader et al. 2000).
Another control of the environmental release of EE2 (and possibly many other
pharmaceuticals) can be through educating the public on the proper disposal of expired or
unwanted pharmaceutical drugs. Thirty-five percent (35%) of American’s dispose of
these pharmaceuticals in the sink or toilet (Bound and Voulvoulis 2005). This disposal
process increases the likelihood of these chemicals being released into the environment at
concentrations above effect levels. By educating the public of the environmental fate of
drugs, the improper disposal of these pharmaceuticals could significantly be reduced.
Other forms of disposal that could have a lesser effect on the environment include
discarding them in household waste where they can be sent out to a landfill (which would
allow these drugs to have a lesser chance of reaching the water system) or returning them
to an institution that offers a return and disposal service (Bound and Voulvoulis 2005).
As part of my graduate program, I created a brochure which outlined the effects of
improper prescription disposal and provided alternatives to properly dispose of unwanted
or expired medications. This brochure can be obtained electronically through Michigan
State University’s Extension (MSUE) website at:
http://web2.msue.msu.edu/bulletins/Bulletin/PDF/E3110.pdf

61

Future Directions:
Toxicity testing needs to incorporate multiple stressor effects to better assess how
a mixture of chemicals in WWTP effluent interact with one another in situ and potentially
affect the reproduction of aquatic organisms. Future studies should focus on testing a
variety of EDC’s in multiple stressor situations in order to fully understand how
estrogenic

contaminant

interactions,

especially

at

environmentally

relevant

concentrations, impact fish reproduction. Estrogen mimics occur without chemical
exposures due to biological stressors in the environment. Testing estrogenic contaminants
with biological stressors should also be considered if a full understanding of multiple
stressor response is to be acquired. Most multiple stressor studies have focused on testing
the effects of two biological stressors while very few studies have researched
combinations of biological and chemical relationships.
Further investigation into the mode of action of ammonia on vitellogenin
production should also be conducted in order to fully understand why the antagonism
between ammonia exposure and vitellogenin production exists. Models have been
established relating vitellogenin to egg production. If NH3’s mode of action on
vitellogenin production was understood, data from studies such as this could be used to
predict population trajectories. These models could also be used in place of expensive
and time consuming flow-through diluter system experiments if the effects of similar
contaminants with similar modes of action were being analyzed. The mixture study
determined that there is a mortality effect to exposed fish in a mixture of NH3 and EE2 at
their respective environmentally relevant concentration. Further investigation in to why
this mortality is occurring and why the synergistic response between these two

62

contaminants exists should also be thoroughly researched. These future studies would
also be useful in determining if the high mortality experienced in this study could be
attributed to the high reproductive output across all concentrations when compared to
previous flow-through diluter experiments.

63

APPENDICES

64

APPENDIX A: MODEL ORGANISM AND DESIGN OF THE FLOW-THROUGH
DILUTER SYSTEM
Model Organism:
The FHM used in this study were cultured in-house at the U.S. EPA CincinnatiAWBERC location using established methods (U.S. EPA 2006). FHM have been used in
laboratory studies as a model species in toxicological research for decades. They are a
representative of the ecologically important Cyprinidae family (Jensen et al. 2001,
Ankley and Villeneuve 2006), have a broad distribution in both lentic and lotic
environments across North America (Ankley and Villeneuve 2006), and are an
opportunistic omnivore tolerant of a wide range of water types (Ankley and Villeneuve
2006). They have been used extensively in chronic life stage and early life stage survival
and development tests (Ankley et al. 2001) partly because culturing FHM in a laboratory
setting is relatively easy (U.S. EPA 2006). FHM are a good model organism because of
their life-history: they are fractional spawners; produce clutches of 50-100 eggs every 3-5
days and have a rapid life cycle reaching maturity within 4-5 months (Ankley et al.
2001). Controlling their reproductive life cycle is readily achieved through alterations in
temperature and photoperiod (Jensen et al. 2001). Previous studies also suggest that FHM
response to chemicals is comparable to responses observed in a variety of threatened and
endangered fish (Ankley et al. 2001). Extrapolating data from the laboratory to the field
has been a challenging task for ecotoxicologists. Laboratory FHM have played a key role
in making this linkage as they have been an important species in predicting the
bioavailability of inorganics and other contaminants through the use of modeling and
caged deployments (Ankley & Villeneuve 2006).
65

Flow-Through Diluter System:
An established U.S. EPA protocol for conducting flow-through diluter
experiments (Ankley et al. 2001) was implemented to explore how both EE2 and NH3
act together to disrupt reproduction of the FHM. A summary of the experimental design
including test conditions and endpoints of interest is outlined in Appendix C. Laboratory
line water was made by passing tempered tap water through a set of activated carbon
filters, for the removal of chlorine and organics. Liquid calcium chloride was then added
to the water to supplement the hardness. Following treatment, the water moved through
four 500 gallon fiberglass tanks for conditioning. After conditioning, the water was
pumped into the water delivery system and then fed to the diluter system. The diluter
system was designed to provide a continuous flow of toxicant at three toxicant
concentrations including a control to the glass aquarium testing chambers.
The setup of the flow-through diluter system allowed for a continuous renewal of
toxicant to thirty chambers (Figure 19). Each day, three amounts of highly concentrated
stock solution were added to three separate 18L dilution water pots to produce the diluted
stock concentrations, from here on referred to as a superstock. This superstock was mixed
up fresh daily in order to provide a continuous diluted stock for the flow-through diluter
system. For each treatment including the control, a 22qt stainless steel stockpot
containing a 30x toxicant (or control) superstock was placed on an electric stir plate. The
superstock from each stockpot was drawn at a rate of 10ml/min using a 4 cartridge
Watson-Marlow 205u series peristaltic pump. The superstocks were then diluted with
laboratory line water, which was pulled from a single 24qt stainless steel stockpot at a
rate of 290ml/min using a Watson-Marlow 500 series peristaltic pump, to create the

66

testing solutions. Two 32-cartridge Watson-Marlow peristaltic pumps then distributed the
testing solutions into randomly selected 12L testing chambers each at a rate of 20ml/min,
which resulted in fresh testing solution being supplied to the tanks at rate of 40ml/min.
This flow-rate was sufficient enough for 6 water changes per day. The setup allowed for a
fail-safe system in the event that one pump fails or if a tube breaks during the exposure
period. In this event, the other 32 cartridge pump would still distribute the testing solution
into each testing chamber at a rate of 20ml/min, thus allowing the dosing to continue.
The diluter system was designed to test twenty-eight testing chambers and two
blank tanks. The blank tanks were needed to measure background water chemistry in the
event of high control mortality. The testing chambers were randomly selected for the
treatments using the online randomizer found at http://www.randomizer.org. Each
chamber was equipped with a drainage system to allow for a continuous renewal of 9.5L
of testing solution. The temperature within the chambers was monitored weekly and
heated using a water bath to 25 ± 1°C. The water bath was heated to 27°C using a
submersible heating cord. Seven replicates, each consisting of 2 male and 4 female adult
FHM were used in each of three testing concentrations and control. Three spawning tiles,
made of PVC pipe sections, were also placed into each replicate chamber. Having three
tiles in each chamber allowed both males to each court a female for spawning and
allowed an additional chamber to serve as a hiding place for the other females.

67

A: Laboratory line water inlet
B: Either EE2 or NH3 stock solution (18L)
C: Control/dilution water holding tank (20L) with float valve
D: Control/dilution water 500 series Watson-Marlow pump
E: Toxicant and dilution water mixing tank
F: 32 cartridge 205u series Watson-Marlow pump
G: Excess sink with carbon filtration
H: Toxicant delivery 4 cartridge 205u series Watson-Marlow pump

I: Lighting system
J: Fish holding chamber (14L) with drain system
K: Drain with carbon filtration
L: Laboratory line water flow for heating bath
M: Heating cord

Figure 19: Flow-through diluter system used to expose fathead minnow breeding pairs to mixtures of 17α-ethinylestradiol and
un-ionized ammonia for 21 days

68

APPENDIX B: SUMMARY OF TEST CONDITIONS FOR PIMEPHALES
PROMELAS
TEST PARAMETER
Test Type

CONDITION
Flow-through

Test Duration

14 day acclimation period
21 day exposure period

Temperature

25°C (±1°C)

Photoperiod

16 h light: 8 h dark

Test Chamber Size

12 L

Test Solution Volume

9.5 L

Renewal of Test Solution

Continuous

Age of Test Organisms

4-6 month old Pimephales promelas

Weight of Test Organisms

Male: 3.0 – 4.0g
Female: 1.0 – 2.0g

No. Organisms/Test Chamber

2 male: 4 female

No. Replicate Test Chambers

7

No. Organisms/concentration

14 male: 28 female

Feeding Regime

1ml frozen Artemia salina (1x daily)
20ml newly hatched Artemia salina
(2x daily)

Control and/or Dilution Water

Laboratory Line Water

Sample/Test Material

EE2, NH3, and EE2 + NH3

Endpoints

Survival, Fecundity, Fertility,
Gonadosomatic Index, Fatpad Index,
Vitellogenin Analysis

Test Acceptability

90% or greater control survival
85% or greater control fertility
<20% CV for analytical chemistries

69

REFERENCES

70

REFERENCES
Adams, S.M. (2005). Assessing cause and effect of multiple stressors on Marine systems.
Marine Pollution Bulletin, 51, 649-657.
Alexander, H.C. Latvaitis, B.P., & Hopkins, D.L. (1986). Site-specific toxicity of unionized ammonia in the Tittabawassee River at Midland, Michigan: Overview.
Environmental Toxicology and Chemistry, 5, 427 – 435
Ankley, G.T., Jensen, K.M., Kahl, M.D., Korte, J.J., & Makynen, E.A. (2001).
Description and evaluation of a short-term reproductive test with the fathead
minnow (Pimephales promelas). Environmental Toxicology and Chemistry,
20(6): 1276-1290.
Ankley, G.T. & Villeneuve D.L. (2006). The fathead minnow in aquatic toxicology: past,
present, and future. Aquatic Toxicology, 78: 91-102.
Armstrong, B.M. (2010). How to properly dispose of unwanted or expired medication.
Michigan State University Extension Bulletin, E-3110.
Bonga, S.E. (1997). The stress response in fish. Physiological Reviews, 77, 591 – 625.
Bound, J.P. & Voulvoulis, N. (2005). Household disposal of pharmaceuticals as a
pathway for aquatic contamination in the United Kingdom. Environmental Health
Perspectives, 113(12), 1705-1711.
Di Giulio, R.T. & Hinton, D.E. (2008). The toxicology of fishes. CRC Press, Taylor and
Francis Group.
Filby, A.L., Thorpe, K.L., Maack, G. & Tyler, C.R. (2007). Gene expression profiles
revealing the mechanisms of anti-androgen and estrogen-induced feminization in
fish. Aquatic Toxicology, 81, 219-231.
Folt, C.L., Chen, C.Y., Moore, M.V. & Burnaford, J. (1999). Synergism and antagonism
among multiple stressors. Limnology and Oceanography, 44 (3:2), 864-877.
Gersberg, R.M, Elkins, B.V., Lyon, S.R. & Goldman, C.R. (1986). Role of aquatic plants
in wastewater treatment by artificial wetlands. Water Research, 20(3), 363 – 368.
Giesy, J.P., Snyder, E.M., Nichols, K.M., Snyder, S.A., Villalobos, S.A., Jones, P.D., &
Fitzgerald, S.D. (2003). Examination of reproductive endpoints in goldfish
(Carassius auratus) exposed in situ to municipal sewage treatment plant effluent
discharges in Michigan, USA. Environmental Toxicology and Chemistry, 22(10),
2416-2431.

71

Grist, E.P., Wells, N.C., Whitehouse, P., Brighty, G., & Crane, M. (2003). Estimating the
effects of 17a-ethinylestradiol on populations of the fathead minnow Pimephales
promelas: Are conventional toxicological endpoints adequate? Environmental
Science and Technology, 37, 1609-1616.
Gross, M.R., Repka, J., Robertson, C.T., Secor, D.H & Van Winkle, W. (2002). Sturgeon
conservation: Insights from elasticity analysis. American Fisheries Society
Symposium, 28, 13 - 30
Gunderson, D. R. (1997). Trade-off between reproductive effort and adult survival in
oviparous and viviparous fishes. Canadian Journal of Fisheries and Aquatic
Sciences, 54, 990 – 998.
Hanazato, T. & Dodson, S.I. (1995). Synergistic effects of low oxygen concentration ,
predator kairomone, and a pesticide on the cladoceran Daphnia pulex. Limnology
and Oceanography, 40(4), 700 – 709.
Hawkins, M.B., Thornton, J.W., Crews, D., Skipper, J.K., Dotte, A., & Thomas, P.
(2000). Identification of a third distinct estrogen receptor and reclassification of
estrogen receptors in teleosts. Proceedings National Academy of Sciences, USA,
97, 10751 – 10756.
Hodges, S.W. (2007). Movement, survival and refuge use of three minnow species
(Campostoma anomalum, Notropis boops, and Semotilus atromaculatus) during
seasonal drying in an intermittent Ozark mountain stream. (Doctoral dissertation,
University of Arkansas, 2007). Abstract retrieved November 27, 2010 from
ProQuest database.
Islinger, M. Willimski, D., Völkl, A. & Braunbeck, T. (2003). Effects of 17aethinylestradiol on the expression of three estrogen-responsive genes and cellular
ultrastructure of liver and testes in male zebrafish. Aquatic Toxicology, 62, 85103.
Jenkins, R.E & Burkhead, N.M. (1993). Freshwater Fishes of Virginia. American
Fisheries Society, Bethesda, Maryland.
Jensen, K.M., Korte, J.J., Kahl, M.D., Pasha, M.S., & Ankley, G.T. (2001). Aspects of
basic reproductive biology and endocrinology in the fathead minnow (Pimephales
promelas). Comparative Biochemistry and Physiology Part C, 128: 127-141.
Jenssen, B.M. (2006). Endocrine-disrupting chemicals and climate change: A worst-case
combination for Arctic marine mammals and seabirds? Environmental Health
Perspectives, 114(1): 76-80.
Jobling, S., & Tyler, C.R. (2003). Endocrine disruption, parasites and pollutants in wild
freshwater fish. Parasitology, 126, S103 – S108

72

Jorgensen, T.C. & Weatherley, L.R. (2003). Ammonia removal from wastewater by ion
exchange in the presence of organic contaminants. Water Research, 37(8), 1723 –
1728.
Kidd, K.A., Blanchfield, P.J., Mills, K.H., Palace, V.P., Evans, R.E., Lazorchak, J.M., &
Flick, R.W. (2007). Collapse of a fish population after exposure to a synthetic
estrogen. Proceedings of the National Academy of Sciences, 104(21): 8897-8901
Kime, D.E. (1998). Endocrine disruption in fish. Kluwer Academic, London.
King County. (2007). Survey of endocrine disruptors in King County surface waters.
Prepared by Richard Jack and Deb Lester, Water and Land Resources Division.
Seattle, Washington.
Korte, J.J., Mylchreest, E. & Ankley, G.T. (2004). Comparative evaluation of ELISAs for
detecting vitellogenin in the fathead minnow (Pimephales promelas) – a response
to Tyler et al. Comparative Biochemistry and Physiology Part C, 138: 533-536.
Lange, R., Hutchinson, T.H., Croudace, C.P., Siegmund, F., Schweinfurth, H., Hampe,
P., Panter, G.H., & Sumpter, J.P. (2001). Effects of the synthetic estrogen 17αethinylestradiol on the life-cycle of the fathead minnow (Pimephales promelas).
Environmental Toxicology and Chemistry, 20(6), 1216 – 1227.
Lazorchak, J. M. & Smith, M.E. (2004). National Screening Survey of EDCs in
Municipal Wastewater Treatment Effluents. EPA/600/R-04/171. U.S.
Environmental Protection Agency, Cincinnati, OH.
Ma, T., Wan, X., Huang, Q., Wang, Z. & Liu, J. (2005). Biomarker responses and
reproductive toxicity of the effluent from a Chinese large sewage treatment plant
in Japanese medaka (Oryzias latipes). Chemosphere, 59, 281-288.
Mayes, M.A., Alexander, H.C. & Hopkins, D.L. (1986). Acute and chronic toxicity of
ammonia to freshwater fish: a site-specific study. Environmental Toxicology &
Chemistry, 5, 437-442
McNaught, A.D. & Wilkinson, A. (1997). Compendium of Chemical Terminology.
Accessed on May 21, 2009 from the International Union of Pure and Applied
Chemistry website at http://goldbook.iupac.org/index.html
Mercer, B.W., Ames, L.L., Touhill, C.J., Van Slyke, W.J. & Dean, R.B. (1970).
Ammonia removal from secondary effluents by selective ion exchange. Water
Pollution Control Federation, 42(2), R95-R107
Miller, D.H., Jensen, K.M., Villeneuve, D.L., Kahl, M.D., Makynen, E.A., Durhan, E.J.
& Ankley, G.T. (2007). Linkage of biochemical responses to population-level
effects: A case study with vitellogenin in the fathead minnow (Pimephales
promelas). Environmental Toxicology and Chemistry, 26(3), 521 – 527

73

Murphy, C.A., Rose, K.A. & Thomas, P. (2005). Modeling vitellogenesis in female fish
exposed to environmental stressors: predicting the effects of endocrine
disturbance due to exposure to a PCB mixture and cadmium. Reproductive
Toxicology, 19, 395 – 409
Nash, J.P., Kime, D.E., Van der Ven, L.T.M., Wester, P.W., Brion, F., Maack, G.,
Stahlschmidt-Allner, P. & Tyler, C.R. (2004). Long-term exposure to
environmental concentrations of the pharmaceutical ethynylestradiol causes
reproductive failure in fish. Environmental Health Perspectives, 112(17): 17251733
Osborne, N., Sherry, J., Rendell, J.L, & Currie, S. (2007). The role of hsp90 in 17αethynylestradiol-induced endocrine disruption in rainbow trout hepatocytes.
Ecotoxicology & Environmental Safety, 68, 13 – 19.
Parks, L.G., Cheek, A.O., Denslow, N.D., Heppell, S.A., McLachlan, J.A., Leblanc,
G.A., & Sullivan, C.V. (1999). Fathead minnow (Pimephales promelas)
vitellogenin: purification, characterization and quantitative immunoassay for the
detection of estrogenic compounds. Comparative Biochemistry and Physiology
Part C, 123, 113-125.
Parrott, J.L. & Blunt, B.R. (2005). Life-cycle exposure of fathead minnows (Pimephales
promelas) to an ethinylestradiol concentration below 1 ng/L reduces egg
fertilization success and demasculinizes males. Wiley InterScience. 131-141.
Passell, H.D., Dahm, C.N. & Bedrick, E.J. (2007). Ammonia modeling for assessing
potential toxicity to fish species in the Rio Grande, 1989 – 2002. Ecological
Applications, 17(7), 2087 – 2099.
Pawlowski, S., van Aerle, R., Tyler, C.R. & Braunbeck, T. (2004). Effects of 17aethinylestradiol in a fathead minnow (Pimephales promelas) gonadal
recrudescence assay. Ecotoxicology and Environmental Safety, 57: 330-345.
Randall, D.J. & Tsui, T.K.N. (2002). Ammonia toxicity in fish. Marine Pollution
Bulletin, 45, 17 – 23.
Robbins, J.W.D., Howells, D.H. & Kriz, G.J. (1972). Stream pollution from animal
production units. Journal of Water Pollution Control Federation, 44(8), 1536 –
1544.
Schisler, G.J. & Bergersen, E.P. (2000). Effects of multiple stressors on morbidity and
mortality of fingerling rainbow trout infected with Myxobolus cerebralis.
Transactions of the American Fisheries Society, 129, 859 – 865.

74

Schlosser, I.J. (1995). Dispersal, boundary processes, and trophic level interactions in
streams adjacent to beaver ponds. Ecology, 76, 908-925.
Schultz, I.R., Orner, G., Merdink, J.L. & Skillman, A. (2001). Dose-response
relationships and pharmacokinetics of vitellogenin in rainbow trout after
intravascular administration of 17a-ethynylestradiol. Aquatic Toxicology, 51, 305318.
Shi, J., Fujisawa, S., Nakai, S. & Hosomi, M. (2004). Biodegredation of natural and
synthetic estrogens by nitrifying activated sludge and ammonia-oxidizing
bacterium Nitrosomonas europaea. Water Research, 38, 2323-2330.
Skillman, A.D., Nagler, J.J., Hook, S.E., Small, J.A. & Schultz, I.R. (2006). Dynamics of
17a-ethynylestradiol exposure in rainbow trout (Oncorhynchus mykiss):
absorption, tissue distribution, and hepatic gene expression pattern.
Environmental Toxicology and Chemistry, 25(11): 2997-3005.
Starrett, W.C. (1951). Some factors affecting the abundance of minnows in the Des
Moines River, Iowa. Ecology, 32(1), 13 – 27.
Thurston, R.V., Russo, R.C. & Meyn, E.L. (1986). Chronic toxicity of ammonia to
fathead minnows. Transactions of the American Fisheries Society, 115, 196 – 207
U.S. EPA. (1999). 1999 update of ambient water quality criteria for ammonia.
EPA/822/R-99/014. Office of Water. Washington DC.
U.S. EPA. (2002). A short-term method for assessing the reproductive toxicity of
endocrine disrupting chemicals using the fathead minnow (Pimephales promelas).
EPA/600/R-01/067. Washington D.C.
U.S. EPA. (2005). Standard operating procedure for Tri Reagent RNA isolation
protocol. National Exposure Research Laboratory number MIRB 031.2
U.S. EPA. (2006). Standard operating procedure for fathead minnow culture method
(Pimephales promelas) Number T03-100. Dynamac/NERL/EP-D-06-096.
U.S. EPA. (2007). Standard operating procedure for the analysis of estrogen compounds
from water samples by GC/MS. National Exposure Research Laboratory Number
MIRB 051.0E.
U.S. EPA (2009). Standard operating procedure for the qualitative “real time” PCR
(QPCR) vitellogenin protocol using HS DyNamo SYBR Green Kit. National
Exposure Research Laboratory Number MIRB 032.3.
U.S. EPA. (2009). 2009 update of ambient water quality criteria for ammonia freshwater. EPA/822/D-09/001. Office of Water. Washington DC.

75

Vader, J.S., van Ginkel, C.G., Sperling, F.M.G.M., de Jong, J., de Boer, W., de Graaf,
J.S., van der Most, M. & Stokman, P.G.W. (2000). Degradation of ethinyl
estradiol by nitrifying activated sludge. Chemosphere, 41, 1239-1243.
Weber, L.P., Balch, G.C., Metcalfe, C.D. & Janz, D.M. (2004). Increased kidney, liver,
and testicular cell death after chronic exposure to 17a-ethinylestradiol in medaka
(Oryzias latipes). Environmental Toxicology and Chemistry, 23(3), 792-797.
Wolfand, J. (2007). Active ingredient in oral contraceptives (17a-ethinylestradiol) alters
male competitive courtship behaviors and secondary sexual characteristics in
fathead minnows (Pimephales promelas). Journal of U.S. Stockholm Junior Water
Prize. Volume 2. Water Environment Federation.
Wood, C.M. (2001). Chapter 1: Target responses of the gill. IN Target organ toxicity in
marine and freshwater teleosts, Volume 1 Organs. Edited by: Schlenk, D.
&Benson, W.H. London, Taylor & Francis: 1-85.

76