Citation
An integrative approach to characterizing the estrogenicity gradient of a portion of the South Platte River

Material Information

Title:
An integrative approach to characterizing the estrogenicity gradient of a portion of the South Platte River
Creator:
Bourdon, Lisa Marie ( author )
Language:
English
Physical Description:
1 electronic file (38 pages) : ;

Subjects

Subjects / Keywords:
Endocrine disrupting chemicals in water ( lcsh )
Endocrine disrupting chemicals in water ( fast )
South Platte River (Colo. and Neb.) ( lcsh )
United States -- South Platte River ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Endocrine disrupting compounds (EDCs) from agricultural, industrial, and municipal sources can be found in many surface waters with potential adverse implications for human and ecosystem health. The South Platte River represents a significant source of water for the Denver Metro Area, yet little data exists concerning EDCs. The aim of this study is to evaluate the occurrence and effects of EDCs downstream from two major wastewater treatment plants (WWTPs). This study characterizes the estrogenicity gradient of the South Platte River in the Denver Metro area by combining data from qPCR analysis for liver vitellogenin (vtg) mRNA with liver NMR metabolomics after a 5 day in situ caged exposure of fathead minnows. Concurrent water samples collected from the start and end times of the exposures were used to determine the occurrence and concentration of wastewater contaminants. Results found 68 of 122 chemicals downstream of WWTP 1 and 73 downstream of WWTP 2, including known EDCs (e.g. nonylphenol and octylphenol). A steroidal estrogen, estrone, was only found downstream of WWTP 2. Consistent with the highest measured concentrations of wastewater estrogens, the highest levels of vtg mRNA were measured downstream of WWTP 2. Metabolomics data coincided with vtg data and showed little variation except downstream of WWTP 2, where male polar metabolomes showed increased levels of alanine and glutamate, which are utilized in VTG synthesis. PCA of male polar metabolomes showed significant separation of WWTP 2 from WWTP 1 and the reference site, further supported by PLS-DA scores plot. Female polar metabolomes showed significant separation between WWTP 1 and WWTP 2 using PLS-DA scores plot. This study demonstrates that qPCR and metabolomics data can be reliably and concurrently used to illuminate impacts from chemical exposures, although further research will better elucidate target genes and metabolites of interest.
Thesis:
Thesis (M.S.)-University of Colorado Denver.
Bibliography:
Includes bibliographic references
System Details:
System requirements: Adobe Reader.
Statement of Responsibility:
by Lisa Marie Bourdon.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
954767499 ( OCLC )
ocn954767499
Classification:
LD1193.L45 2016m B69 ( lcc )

Downloads

This item has the following downloads:


Full Text
AN INTEGRATIVE APPROACH TO CHARACTERIZING THE ESTROGENICITY
GRADIENT OF A PORTION OF THE SOUTH PLATTE RIVER
by
LISA MARIE BOURDON
B.A., Ball State University, 2010
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Master of Science
Biology Program
2016


2016
LISA MARIE BOURDON
ALL RIGHTS RESERVED


This thesis for the Master of Science degree by
Lisa Marie Bourdon
has been approved for the
Biology Program
by
Alan Vajda, Chair
Amanda Charlesworth
Kristen Keteles
April 28, 2016
n


Bourdon, Lisa Marie (M.S., Biology)
An Integrative Approach to Characterizing the Estrogenicity Gradient of a Portion of the South
Platte River
Thesis directed by Assistant Professor Alan Vajda
Abstract
Endocrine disrupting compounds (EDCs) from agricultural, industrial, and municipal sources can
be found in many surface waters with potential adverse implications for human and ecosystem
health. The South Platte River represents a significant source of water for the Denver Metro
Area, yet little data exists concerning EDCs. The aim of this study is to evaluate the occurrence
and effects of EDCs downstream from two major wastewater treatment plants (WWTPs). This
study characterizes the estrogenicity gradient of the South Platte River in the Denver Metro area
by combining data from qPCR analysis for liver vitellogenin (vtg) mRNA with liver NMR
metabolomics after a 5 day in situ caged exposure of fathead minnows. Concurrent water samples
collected from the start and end times of the exposures were used to determine the occurrence and
concentration of wastewater contaminants. Results found 68 of 122 chemicals downstream of
WWTP 1 and 73 downstream of WWTP 2, including known EDCs (e.g. nonylphenol and
octylphenol). A steroidal estrogen, estrone, was only found downstream of WWTP 2. Consistent
with the highest measured concentrations of wastewater estrogens, the highest levels of vtg
mRNA were measured downstream of WWTP 2. Metabolomics data coincided with vtg data and
showed little variation except downstream of WWTP 2, where male polar metabolomes showed
increased levels of alanine and glutamate, which are utilized in VTG synthesis. PCA of male
polar metabolomes showed significant separation of WWTP 2 from WWTP 1 and the reference
site, further supported by PLS-DA scores plot. Female polar metabolomes showed significant
separation between WWTP 1 and WWTP 2 using PLS-DA scores plot. This study demonstrates
that qPCR and metabolomics data can be reliably and concurrently used to illuminate impacts
m


from chemical exposures, although further research will better elucidate target genes and
metabolites of interest.
The form and content of this abstract are approved. I recommend its publication.
Approved: Alan Vajda
IV


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION............................................................1
II. METHODS.................................................................6
Overview................................................................6
Site Description........................................................6
Model Organism..........................................................7
Field Exposure..........................................................7
Dissection..............................................................8
vtgmRNA.................................................................8
NMR Sample Preparation and Analysis.....................................9
Analytical Chemistry...................................................10
Statistical Analysis...................................................11
III. RESULTS................................................................13
Analytical Chemistry...................................................13
vtg qPCR...............................................................16
NMR Metabolomics.......................................................16
IV. DISCUSSION AND CONCLUSIONS.............................................22
Discussion.............................................................22
Conclusions............................................................25
REFERENCES..................................................................27
v


LIST OF TABLES
TABLE
1. Chemical Detections.............................................................13
2. Detected Concentrations.........................................................15
vi


LIST OF FIGURES
FIGURE
1. Site Map.......................................................................7
2. Analytical Chemistry Methods...................................................11
3. Shared and Unique Chemical Detections..........................................14
4. Male vtg mRNA qPCR Interval Plot...............................................16
5. Male Polar NMR Metabalomics Spectra............................................17
6. Male Polar NMR Metabalomics PCA for All Classes................................18
7. Male Polar NMR Metabalomics PCA Excluding Time 0...............................18
8. PLS-DA Male Polar Metabolomes for Clear Creek and WWTP 2.......................19
9. PLS-DA Male Polar Metabolomes for WWTP 1 and WWTP 2............................19
10. Female Polar NMR Metabalomics PCA Outlier.....................................20
11. Female Polar Metabolomics PCA for All Classes.................................21
12. PLS-DA Female Polar Metabolomes for WWTP 1 and WWTP 2.........................21
vii


LIST OF ABBREVIATIONS
WWTP Wastewater Treatment Plant
VTG Vitellogenin protein
vtg Vitellogenin gene
NMR Nuclear Magnetic Resonance
qPCR Quantitative Polymerase Chain Reaction
PCA Principle Components Analysis
PLS-DA Partial Least Squares Discriminant Analysis
EE2 17 a-ethynylestradiol
vm


CHAPTERI
INTRODUCTION
With an ever-growing list of chemicals used for a plethora of anthropogenic uses, there is
a growing concern for how these chemicals may be affecting the environment as well as human
health. An important source of chemicals in the environment is municipal wastewater treatment
plant (WWTP) effluent, which is known for containing chemicals that can impact the endocrine
system 1-5. Municipal WWTP effluent contains the remnants of personal care products and
excreted pharmaceuticals, which oftentimes pass through WWTPs and enter waterways, where
their fate is unknown3,5-7. Personal care products often contain chemicals implicated in endocrine
disruption, and many pharmaceuticals also have potential adverse effects on the environment and
warrant further study 89. Pharmaceuticals in the environment were long thought to exist at
concentrations too low to cause concern; thus, data on their presence and persistence in the
environment is less extensive8. Furthermore, methods of detecting chemicals in the environment
were, and in many ways still are, limited3. Another troubling facet of WWTP effluent is its
constancy. While runoff from other sources, such as agriculture, are intermittent and short lived,
WWTP effluent discharges at regular intervals and can create pseudo-persistent exposures that
can alter normal biological and ecological processes 8.
Previous research concerning chemicals in the environment, particularly those that are
known or suspected to affect the normal function of the endocrine system, indicates that there are
instances of perturbations in normal reproductive function, and in some cases, survivorship across
taxa10,11. There are many studies evaluating altered endpoints due to endocrine disruption in
fishes, including; skewed sex ratios, altered gametogenesis, disrupted gonadal development,
intersex, and alterations in hormone mediated functions, such as vitellogenin protein (VTG)
production 12 l7. Studies in amphibians link endocrine disruption to deformities, lowered immune
function, and lower survivability12,18. Developmental exposure of reptiles can lead to altered sex
1


ratios12 as well as altered plasma steroid levels and abnormal gonadal development11. Predatory
birds exposed to estrogenic compounds may be susceptible to population declines due to eggshell
thinning and resulting reduced offspring survivorship1119,20. Mammals have had reported
incidences of altered gametogenesis, and increased incidences of reproductive cancers11,12.
Humans specifically have had increased incidences of birth defects in males such as hypospadias,
cryptorchidism and shortened anogenital distance, and early menarche in females12,21. Even some
invertebrates have demonstrated reproductive impairment associated with endocrine disruptor
exposure, although their utilization of steroid hormones is unclear and requires further study16.
Thus, it is evident that because of the well-conserved nature of the endocrine system, chemicals
are eliciting detrimental effects across taxa, including humans. Because of this, it is important
that better methods of detection and evaluation of risk-assessment be developed to identify
chemicals that warrant the most concern.
Rapid, cost-effective, and sensitive methods for risk assessment are necessitated by the
complexity of contaminant mixtures discharged into the environment. The number of chemicals,
and their potential mechanisms of action, are simply too numerous to do comprehensive toxicity
testing on every individual chemical, let alone of environmentally-relevant mixtures.
Furthermore, our analytical methods offer an incomplete picture of the chemicals present in a
mixture at any given time22. Analytical chemical methods are generally limited to known
chemicals, cannot account for mixture effects, and cannot practically account for temporal
variations3 22 24. The National Research Council has challenged the scientific community to not
only find methods that eschew traditional animal testing and rely more on cell-based assays while
also linking adverse outcomes to exposures25. As scientists better understand how cellular
responses are connected to one another, it is possible that biological pathways associated with
adverse health effects from exposures can be identified and utilized to recognize exposures in the
future. More traditional biomarker studies are generally centered around isolated measurements;
2


including: secondary sex characters, gamete production, plasma sex steroid concentrations, and
VTG levels VTG, an egg yolk precursor protein produced in many egg-laying females, is
estrogen-induced and is generally only produced in male organisms impacted by estrogenic
compounds, making it a useful biomarker of estrogen exposure26. However, traditional biomarker
studies can be limited because they are a single component in a complicated framework that can
be altered by any number of changes to other components27 and impacts on that single component
are not always informative of effects at higher levels of organization. Recent years have seen the
development of more systems-based approaches to studying the effects of chemical exposures on
organisms, including large-scale, high-throughput studies of transcriptomics, proteomics, and
metabolomics. Obtaining data about the networks of gene transcripts, proteins and metabolites,
and how exposures impact those networks, shows great potential for improved specificity in
testing for organism exposure effects27. Metabolomics in particular has great utility in toxicology
studies because changes to the metabolome occur quickly, and because alterations in the
metabolome directly impact organismic function, while transcriptomics changes do not
necessarily translate to direct organism impacts 28. Furthermore, NMR is generally less expensive
than other high-throughput methods like microarrays and still yields large data sets for analysis.
There is also the potential for non-invasive sample collections, such as the use of urine samples or
other body fluids, which could greatly limit animal usage29. Metabolomics has been utilized in
recent years to evaluate the effects of chemical exposures on organisms with reproducible success
and several common metabolite shifts30 34. While our ability to produce these data sets has
preceded our ability to fully analyze them, our knowledge of reading metabolomics NMR spectra
increases all the time and elucidates new metabolites of interest35.
The South Platte River Basin, as well as other waterways in arid or semi-arid
environments, faces a particular problem with chemical loading due to the disproportionately
high impact of WWTP effluents discharging to relatively low base-flow streams824. During low
3


flow times of year, WWTP effluent can comprise >90% of stream flow. This persistent influx of
effluents, coupled with limited dilution, can modify downstream habitats particularly in terms of
temperature, DO, salinity, nutrient and chemical loading, and overall water quality8. This
problem is further compounded by the fact that many chemicals do not have drinking water
guidelines, drinking-water health advisories, or aquatic-life criteria3. Since the human population
is expected to increase substantially in the coming decades, municipalities will increasingly come
to rely on water reclamation for drinking water resources, particularly in drier areas, and that
there will be a corresponding increase in pharmaceutical and personal care products entering
WWTPs8. Thus, it is increasingly important that scientists develop better methods of detecting
chemicals and evaluating their impacts for regulatory purposes.
The research in this manuscript targets the wastewater-dominated urban corridor of the
South Platte River through Denver, CO, USA. Denver Metro WWTP is the largest facility in the
region, and serves a population of approximately 1,700,000 with a discharge volume of 130
million gallons/day. WWTP effluent comprises over 90% of stream flow in the South Platte
River downstream of the Denver Metro WWTP. The effluent-dominated South Platte River
downstream from the Denver Metro WWTP is utilized as a source of drinking water for the
rapidly expanding urban community of Aurora, CO, and persistent concentrations of CECs
presents an emerging eco-human health challenge.
The goal of this study is to utilize multiple systems level assays integrated with chemical
sampling to assess the potential estrogenicity of a highly populated, heavily wastewater impacted
portion of the South Platte River and determine their efficacy as useful tools to screen for
exposure to estrogenic chemicals. Following a 5 day in-situ caged fish exposure, a quantitative
polymerase chain reaction (qPCR) assay was used to quantify liver vtg mRNA transcripts to
evaluate estrogenic exposure from the sites. Concurrent liver NMR metabolomics was performed
to evaluate the utility in using metabolomics as a tool for assessing chemical exposure.
4


Metabolomics has the potential to be informative of exposures other than estrogenic exposures,
unlike the presence of vtg, which is specific to estrogen exposure. Because of the high-
throughput nature of metabolomics data, inferences at higher levels of organization can be made
Transcript and metabolomic results are interpreted in the context of the measured concentration
and occurrence of WWTP contaminants.
5


CHAPTER II
METHODS
Overview
This study is focused on a 5-day in-situ caged-fish study that took place downstream of 2
WWTPs from 08/09/2012 to 08/14/2012. At the time of fish deployment and retrieval, water
samples were collected from upstream and downstream of both WWTPs, as well as at a reference
site. After retrieval, fish were sacrificed and their livers dissected out for analysis of vtg gene
expression as well as NMR metabolomics analysis. Water samples were analyzed for wastewater
contaminants.
Site Description
Two WWTPs that discharge their effluents into the South Platte River, Denver, Colorado
were chosen for this study. The wastewater treatment plants will be referred to as WWTP 1 and
WWTP 2. WWTP 1 is located at 2900 S. Platte River Drive, Englewood, CO 80110
(39.6587015, -105.00228440000001). WWTP 2 is located at 6450 York St, Denver, CO 80229
(39.8149471, -104.95789450000001). Both WWTPs have primary treatment, activated solids,
solids removal, and chlorine disinfection. WWTP 1 receives approximately 83 x 106 L/d and has
a nitrifying trickling filter treatment step as well as de-nitrification for nutrient removal. WWTP
2 receives approximately 454 x 106 L/d and has two treatment complexes, North and South. At
WWTP2, tertiary treatment, biological nutrient removal, only occurs in the North complex. A
reference site was chosen in Clear Creek upstream from Golden, CO, which is a relatively un-
impacted site that has a well-researched hydrology and at which chemical loading is minimal
(Figure 1).
6


Figure 1: Site Map
Map of North Eastern Colorado, including the South Platte River and Clear Creek.
Model Organism
The fathead minnow (Pimephales promelas) was chosen as the study organism because
they are a member of a ubiquitous family of fish that is common in chemically impacted
environments3637. There are also a commonly used species in ecotoxicology studies and
substantial data already exists to establish reliable biomarkers for chemical exposure.
Furthermore, adequate genomic data is available to utilize fathead minnows for genetic based
testing37. Adult fathead minnows were provided by the U.S. EPA Office of Research and
Development, Molecular Ecology Research Division in Cincinnati, OH.
Field Exposure
Fish were deployed in situ in cages for 5 days at 3 sites beginning on 08/09/2012 and
ending on 08/14/2012. Fish were transported to sites in oxygenated bags set in coolers with ice to
enhance survival, and then acclimated to the stream by replacing the ice with stream water. The
fish were further acclimated to stream water by mixing a 50:50 mixture of lab water and stream
water for 30 minutes. 15 male and 15 female fathead minnows were put in cages composed of a
PVC pipe with mesh on each side and deployed at sites downstream of both WWTP land WWTP
7


2 along the South Platte River as well as at Clear Creek above Golden, CO as a reference site.
The cages were suspended above the stream-bed by attaching nylon cord to a stake or to an
overhanging limb, etc. At the time of deployment, 15 male and female fish were sacrificed as
time 0 initial controls. At the time of fish deployment and retrieval, water samples were collected
from the aforementioned sites, as well as upstream of both WWTPs, for chemical analysis to
evaluate PPCPs, Pesticides, and Waste Indicators. After 5 days of exposure, the fish were
recovered from cages and transported in oxygenated buckets to a laboratory at WWTP2 for
dissection.
Dissection
After the 5 day exposure, fish were anesthetized by immersion in MS-222(0.1g/L).
Plasma samples were collected from the caudal vein for future studies, and then fish were
sacrificed by rapid decapitation. Fish were cut from cloaca to operculum and the internal organs
removed. The liver was carefully removed from the intestine and surrounding organs, and then
divided into sections for qPCR analysis of vtg and NMR metabolomics analysis, placed into
labeled cryotubes, snap frozen, and stored "80C until analysis. The gonads were then removed
from the abdominal cavity, snap frozen, and stored until "80C for future analysis. Gloves were
changed, and dissection tools were cleaned in 70% ethanol between dissections to prevent cross-
contamination between fish or between organs.
vtg mRNA
Extraction. Samples were stored at "80C until extraction. Samples were prepared as previously
described by Biales et al. 200722 on male liver samples. Each sample was placed in a labeled
microcentrifuge tube with a steel ball and lmL of tri-reagent. Samples were homogenized for 5
minutes and then allowed to sit for 10 minutes to allow for complete extraction of genetic
material. lOOpL of bromochloropropane were added to each sample tube and the tubes allowed
to stand for an additional 10 minutes to help improve yield. Samples were centrifuged at
12,000rpm and 4C for 15 minutes, then the aqueous layer pipetted off into freshly labeled
8


microcentrifuge tubes containing 50()giL of isopropanol to precipitate the genetic material.
Samples were centrigued for 10 minutes at 12,000 rpm and 4C and the isopropanol poured off of
the samples. lmL of 75% ethyl alcohol was added to each tube to wash the samples, after which
the samples were centrifuged for 8 minutes at the previously stated specifications. Following
centrifugation, the ethyl alcohol was pipetted off and DEPC water was added until the RNA pellet
was dissolved. The RNA concentrations were quantified using a Nanodrop 1000. Samples were
frozen at "80C if cDNA was not prepared immediately.
cDNA, Thirteen samples at a time were diluted in microAMP tubes to 500ng/pL and run on a
thermal cycler for 10 minutes with 3 controls, one blank with only DEPC water, one that
contained no reverse transcriptase and one that contained a high expressing pool of vitellogenin.
2pL of each sample and control were added to a well in a 96-well plate with 18pL of master mix
composed of PCR buffer, MgCl2 solution, oligo d(t) 23 VN, dNTP mix, random hexamers,
RNase inhibitor, and MuLV reverse transcriptase (for all samples excluding the applicable
control). The plate was sealed and run on the thermal cycler for 60 minutes after which 80pL of
DEPC was added to each sample well. Samples were frozen at -80C if qPCR was not performed
immediately.
qPCR. 18pL of two different mastermixes were pipetted into opposite halves of a 96-well plate.
One mastermix was for vtg RNA and was composed of DEPC, syber, and left and right vtg
primers while the other targeted 18s RNA and was composd of DEPC, syber, and 18s primer.
2pL of each of the 16 samples/controls was pipetted into three wells on either half of the 96-well
plate and the qPCR run on the PCR thermal cycler.
NMR Sample Preparation and Analysis
Samples were kept chilled during the preparation process. Samples were prepared as
described by Ekman et al. 200828. The liver samples were weighed in a pre-tared microcentrifuge
tube and then a steel ball, 400 pL of methanol and 85 pL of milli-Q water were added. The
samples were homogenized on a vortexer for five minutes or until samples were well
9


homogenized and then centrifuged for 60 seconds at l,000g in a 4C microfuge. 300 giL of
chloroform was added and the samples vortexed for 60 seconds and centrifuged again at l,000g
and 4C for one minute. 200 pL of chloroform and 200 pL of Milli-Q water were added and the
samples were vortexed again for 60 seconds, and then centrifuged at l,000g and 4C for 15
minutes. This centrifugation step separated the upper water/methanol phase, containing polar
metabolites, from the lower chloroform phase, containing lipophilic metabolites with a layer of
protein debris in the middle. The upper phase was removed via pipette and put into a labeled
microcentrifuge tube and the lower phase was transferred into a labeled glass vial, being careful
with both to not contaminate with the protein debris layer in the middle. Samples were dried via
nitrogen blow down, and then 600 pL of methanol was added to bring samples to an appropriate
volume for NMR analysis. The samples were then pipetted into labeled NMR sample tubes and
assembled into the spinner for loading into the NMR autosampler for analysis.
Acquired polar spectra were zero-fdled, line-broadened at 0.3 Hz, and Fourier
transformed (ACD/1D NMR Manager, Advanced Chemistry Development). Using an automated
routine, spectra were phase- and baseline-corrected, referenced to TSP, and binned at a width of
0.005 ppm. Regions of binned data were excluded to eliminate a residual water peak (4.38-4.55
ppm), a residual methanol peak (3.31-3.34 ppm), and a residual chloroform peak (7.20-7.80).
Remaining bins were then normalized to unit total integrated intensity.
Analytical Chemistry
Water samples were collected in 1L amber glass containers, adding 80mg of sodium
thiosulfate to remove residual chlorine as needed, when caged fish were deployed and again upon
their retrieval. One glass container each was collected for acid and base analysis. If visible
particles were present in the water samples they were filtered out and the aqueous portion of the
particles removed prior to analysis using vacuum filtration. Extraction and quantification of the
chemicals of interest followed USEPA methods 1694 and 1694 for pharmaceuticals and personal
care products in water/soil/sediment/biosolids and steroids and hormones in
10


water/soil/sediment/biosolids respectively (Figure 2). Chemical data were analyzed qualitatively
for the purposes of this study.
Group 1 Group 2 Group 3
Group 4
Figure 2: Analytical Chemistry Methods
Flow chart for determination of pharmaceutical and personal-care products by LC/MS/MS38.
Statistical Analysis
qPCR. All data transformation was performed using LinRegPCR software, and statistical
analysis performed using Minitab Statistical Software v. 17. AACT data were used to perform an
analysis of variance with Bonferroni correction to reduce chance of Type I error. Significance
levels were set at 0.05.
11


NMR. Principal component analysis (PCA) was used to evaluate overall patterns in the data
(SIMCA-13.0, Umetrics Inc.). Outliers were identified with a Hotellings T2 test at the 95%
confidence interval. Then, partial least-squares discriminant analysis (PLS-DA) models were
created to help visually determine the extent of impact for a given treatment, in relation to
impacts of different treatments. CV-ANOVA (Analysis of Variance testing of Cross-Validated
predictive residuals) was used within SIMCA-P+ for PLS-DA models to assess their validity.
Class score values in PLS-DA scores plots were averaged and are shown with the associated
standard error.
12


CHAPTER III
RESULTS
Analytical Chemistry
Of 122 tested chemicals, none were detected at the Clear Creek reference site. 13 were
detected upstream of WWTP 1, spiking to 68 chemicals in the effluent. Upstream of WWTP 2,
24 chemicals were detected, spiking to 73 chemicals downstream of WWTP 2 (Table 1).
Furthermore, there were 58 common chemicals that were found downstream of both WWTPs
(Figure 3). There were several chemicals that were found in both effluents, including: DEET,
galaxolide, atenolol, tramadol, caffeine, and several organophosphates, which are commonly used
as pesticides. Both WWTP effluents also had measurable levels of common wastewater
contaminants including nonylphenol, fluoxetine, 4-tert-octylphenol, triclosan, and metoprolol.
The steroidal estrogen estrone, and the endocrine-disrupting herbicide Atrazine were only found
in WWTP 2 effluent samples (Table 2). As seen in Table 2, several of the chemicals measured
were substantially higher, in some cases 2-3 times higher, in the WWTP 2 effluent compared to
WWTP 1 effluent.
Table 1: Chemical Detections
Comparisons of the numbers of chemicals found at each site sampled for analytical chemistry, as
well as the number of those chemicals that was unique to that sample site in this study.
Site # of Chemicals Found # of Unique Chemicals Found
Clear Creek (Reference) 0 0
Upstream WWTP 1 13 0
Effluent WWTP 1 68 8
Upstream WWTP 2 24 1
Effluent WWTP 2 73 7
13


Upstream
WWTP 2
1
Downstream
WWTP 1
8
17
Figure 3: Shared and Unique Chemical Detections
Numbers of shared or unique chemical detections by sites.
14


Table 2: Detected Concentrations
Measured concentrations of some detected chemicals by sample site (ng/L). Chemicals are
denoted by endocrine impacts with blue indicating chemicals with suspected endocrine impacts,
green indicating chemicals with both estrogenic and androgenic impacts, red indicating chemicals
with estrogenic impacts, yellow indicating chemicals with observed endocrine impacts with
unknown mechanisms of action, and purple indicating steroidal estrogens.
Chemical Upstream WWTP 1 Effluent WWTP 1 Upsteam WWTP 2 Effluent WWTP 2
DEET 41.6 377, 356, 467 39.6 168,257
Metropolol ND 1030, 839, 1300 10.7 510,609
Atenolol 16.7 371,453,295 11.4 2140, 2790
Tramadol 37.6 1560, 1710, 1320 58.1 1140, 1180
Galaxolide 149 3360, 3320, 3860 103 3480, 4140
Tris(2-butoxyethyl) Phosphate 173 1720, 800, 1850 97.4 1240, 2680
84.0 752,783,787 143 716, 944
Tris(dichloroisopropyl) Phosphate

Nonylphenol ND 264, 285 ND 700,772
4-tert-octvlphenoI ND 91.9, 75.1, 97.8 ND 335
fncTo ND 42.1, 321, 29.4, 38.6, 303 ND 27.6, 11.0, 540.
Qisphenol A ND 214 ND ND
Fluoxetine ND 20.1, 11.0, 21.3 ND 15.0, 16.7
Caffeine 29.5 209, 192, 234, 274 54.4, 51.1 189, 251, 297, 421
Atrazine ND ND 24.1, 30.4 32.1, 55.0
Estrone ND ND ND 169
15


vtg qPCR
Relative expression of vtg in males was normalized to Time 0 and was statistically absent
at the Clear Creek reference site and WWTP 1. However, vtg expression was significantly higher
at WWTP 2 (p<0.05) than the other sites, showing an increase of at least 2 orders of magnitude
(Figure 4).
Interval Plot of RQvs Site
95% Cl for the Mean
Figure 4: Male vtg mRNA qPCR Interval Plot
Relative quantities of vtg gene expression at Time 0, Clear Creek Reference, and both WWTPs 1
and 2.
NMR Metabolomics
Male Metabolomes. Polar metabolite profiles of male fathead minnows showed minimal impact
at WWTP 1, but alanine and glutamate peaks were observed to significantly increase at WWTP 2
(Figure 5). Due to small liver size, relatively few liver fragments were collected for NMR
metabolomics analysis from fish at the Clear Creek Site, limiting comparisons. PCA analysis of
the metabolite profiles reveals that when all 4 classes are evaluated together, Time 0 fish separate
16


the most from the other three groups (Figure 6). This is potentially due to the stress of transport
and not having been fed during transport. Removal of Time 0 fish from the PCA emphasizes the
separation of WWTP 2 from the other two sites (Figure 7). Follow up PLS-DA analysis for
classes excluding Time 0 fish demonstrate significant separation of WWTP 2 metabolomes from
the other two classes (p=0.0059), as does analysis of both Clear Creek reference and WWTP 1
analyzed with WWTP 2 individually,(p=0.0007 and p=6.7e"005 respectively) (Figure 8, Figure 9
respectively) as the model was not valid when all three classes were combined. PLS-DA
maximizes differences between groups, when model is validated with CV-ANOVA; thus, it is
evident that there is a statistically significant difference between the metabolomes of males at
WWTP2 compared to Clear Creek (reference) and WWTP1.
Figure 5: Male Polar NMR Metabolomics Spectra
Characteristic NMR spectra of male fish livers downstream of WWTP 1 (bottom) and male fish
downstream of WWTP 2 (top) minus the spectra of metabolites from the male fish at the Clear
Creek reference site.
17


Figure 6: Male Polar PCA for All Classes
PCA for male polar metabolomes for all treatment classes to determine broad differences in
metabolomes by site.
Con (CQ
WWTPI
. WWTP2
till
R2X[l] = 0.229 R2X[2] = 0.197 Ellipse: Hotelling's T2 (95#)
Figure 7: Male Polar NMR Metabolomics PCA Excluding Time 0
PCA for male polar metabolomes for classes excluding Time 0 to evaluate broad differences in
metabolomes of Clear Creek, WWTP 1, and WWTP 2.
18


Con (CC)
WWTP2
3 SD = 0.927799
2 SD = 0.618533
M55
M62
#Mbl
OM23
M25 O
OM29
3"*!!?? -0.618533
M52 #M56
M53 m59
MS, MSS** M
-----M57 M60------ M&M
-3 SD = -0.927799
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Num
R2X[1] = 0.224
SIMCA 13,0 2/23/2016 9:31:20 AM (UTC-5)
Figure 8: PLS-DA Male Polar Metabolomes for Clear Creek and WWTP 2
PLS-DA for Clear Creek male and WWTP2 male metabolomes to maximize separation between
Clear Creek and WWTP 2 metabolomes.
WWTPl
WWTP2
3 SD = 0.903703
2 SD = 0.602469 "
M55
M52# M54
_ #M56 M62
M53#* _
M58#^0M6O M65
M59 M63
M57# M61 ^166 4
----------------------------
M40 O M41
S439 (
- 2SD ;~-o760246? hMS"
'^bM39 QM48
3 SD = -0.903703
8 10 12 14 16 18 20 22
Num
R2X[1] = 0.201
7:MCA 13.0 2/23/2016 9:33:33 AM (UTC-5)
Figure 9: PLS-DA Male Polar Metabolomes for WWTP 1 and WWTP 2
PLS-DA for WWTPl male and WWTP2 male polar metabolomes to maximize differences
between WWTP 1 and WWTP 2 metabolomes.
19


Female Metabolomes. Female sample 25 was identified as a strong outlier due to extremely high
levels of bile acids observed and was removed from the data set as aberrant data (Figure 10).
PCA demonstrated some separation (Figure 11), and the PLS-DA model for WWTP1 and
WWTP2 female polar metabolomes showed significant differences between the two sites (p=
0.035) (Figure 12). As with male samples, there were few female liver samples obtained for the
Clear Creek reference site.
Q Con (CC)
| Time 0
wi
W2
1
F35P F^SBP F4PfiF10P
I F1P
fU F58P
---- F26P
p^Sgp F57P
#F15P
-1
-1.5-
O F25P
-2 r 1 1 1 1 1 1 1 >
-1.5 -1 -0.5 0 0.5 1
tin
>26 R2X[2] = 0.19 ...... Ellipse: Hotelling's T2. (9
Figure 10: Female Polar NMR Metabolomics PCA Outlier
PCA of all female metabolome data, demonstrating sample F25 as an extreme outlier.
20


H Con (CC)
TimeO
WWTP1
?68 R2x[2] = 0.144 Ellipse: Hotelling's. T2--C?
Figure 11: Female Polar Metabolomics PCA for All Classes
PCA for polar female metabolites for all treatment classes and excluding the extreme outlier to
demonstrate broad separation of metabolomes by site.
WWTPl
WWTP2
t[l]
R2X[2] =0.241
Figure 12: PLS-DA Female Polar Metabolomes for WWTP 1 and WWTP 2
PLS-DA for female polar metabolomes at WWTPl and WWTP2 to maximize differences
between metabolomes.
21


CHAPTER IV
DISCUSSION AND CONCLUSIONS
Discussion
Utilizing biomarkers at multiple levels of biological organization combined with
analytical chemistry methods allowed for characterization of the estrogenicity gradient of a
portion of the South Platte River, Denver, CO. Because of the transient, dynamic nature of
chemical mixtures in aquatic environments, one would expect differences in observed chemical
loading from WWTP effluents, even those in close proximity, with varying degrees of impact on
endocrine function in organisms present2324. Indeed, these preliminary investigations did identify
WWTPs as a major source of CECs with the potential to adversely impact endocrine function in
fishes in the South Platte River basin. We have characterized the occurrence and concentration of
diverse WWTP contaminants including natural and synthetic estrogens and neuro-active
pharmaceuticals. Both WWTP effluents contained a complex and dynamic mixture of wastewater
contaminants that varied in composition between sampling dates; however, there was a greater
occurrence and concentration of known estrogenic CECs (e.g. estrone) downstream from WWTP
2 than downstream from WWTP 1. It is also notable that chemicals were detected upstream of
WWTP 2 that were not detected at other upstream sites, some of which have potential agricultural
or groundwater non-point sources (e.g. atrazine) (Table 2). Still others were found downstream
of WWTP 1, but were not detected upstream of WWTP 2, indicating possible partitioning,
dilution, or degradation mechanisms at work (e.g. nonylphenol, fluoxetine, 4-tert-
octylphenol)(Table 2). Similar complexity and diversity in chemical mixtures, particularly in
terms of the occurrence of pharmaceuticals and personal care products in waterways, has been
observed in waterways across the country3,39.
This higher incidence of chemicals downstream from WWTP 2 is consistent with the
increased vtg mRNA levels found in male fish deployed downstream of that site. The expression
of vtg was found to be at least 4 orders of magnitude higher downstream of WWTP 2 compared
22


to all other test sites. VTG production in fishes can be utilized to determine its 17a-
ethynylestradiol (EE2) estrogen equivalency40. Employing EE2 induced VTG expression data in
fathead minnows obtained by Schwindt41 we are able to estimate an EE2 equivalency of ~5.0
ng/L based on relative vtg expression observed at WWTP 2. The measured concentrations of
steroidal (estrone) and non-steroidal (nonylphenol, and ocylphenol) estrogens at this site supports
this estimate.
The observed increase in vtg mRNA in males also agrees with the altered WWTP 2
metabolomes, particularly the observed increase in alanine and glutamate in males, which have
both been linked to increased production of vtg2. Furthermore, it is evident that some level of
disturbance occurred in both males and females due to the significant shifts seen in WWTP 2
metabolomes compared to WWTP 1. While this impact was greater in males than in females, it is
evident that some metabolic impact is observable in both genders. These gender-specific shifts
have been observed in other studies as well. A study done on fathead minnows exposed to EE2
showed greater impact on male metabolomes compared to females28, while the inverse was
observed in fathead minnows exposed to 17(3-trenbolone, an adrogenic compound1.
These data, combined with a growing body of work by other labs .
demonstrate that integrative approaches to assessing impacts of chemical exposures are becoming
progressively useful. The agreement we see between chemical, vtg gene expression, and
metabolomics data indicates that these methods can be successfully utilized to obtain much-
needed data concerning the presence of estrogenic compounds in uncharacterized waterways. It
is evident from this study and those mentioned above that analytical chemistry alone cannot
provide an accurate picture of the chemicals present in a system; particularly because our
detection limits were not low enough to detect many environmentally relevant concentrations of
chemicals. There is great potential that chemicals were present in the water samples that we
simply could not detect.
23


Likewise, it is clear that utilizing a biomarker like vitellogenin is only useful if certain
chemicals are predominating in a chemical mixture. While vtg expression in a sample reliably
predicts exposure to an estrogenic compound, the lack of such expression cannot necessarily be
used to assume a lack of estrogenic compounds in a chemical mixture. A study done by Finne et
al. found that exposing rainbow trout hepatocytes to chemicals with different known or suspected
mechanisms of action resulted in a mitigated gene expression response compared to what would
be expected from a single chemical exposure42. Because of this mitigating phenomenon, it is
impossible to determine if a lack of vtg expression in a sample is due to a lack of estrogenic
compounds, or if there are other chemicals present that may be mitigating the estrogenic
response. Thus; it is important to carefully choose biomarker tests, understand their limitations,
and ideally pair methods together than can reduce such limitations.
While metabolomics data has great potential, for the time being its applications are still
somewhat limited. In our study, our metabolomics data was complicated by the fact that our
reference site yielded few male or female fish with livers substantial enough for metabolomics
testing. During necropsy the liver was divided for qPCR and metabolomics analysis; however, if
livers were too small, qPCR was prioritized over metabolomics. While a lack of livers for
metabolomics could be indicative of smaller livers, which in turn could indicate healthier, less
chemical impacted fish; it could also have been a consequence of novice dissectors. In future,
devising a way of obtaining an accurate liver weight measurement during necropsy could help
shed light on a potential cause for such issues. Thus, it is possible that the few livers obtained
from the female fish may have confounded our ability to elucidate the relationship of the Clear
Creek reference sites impact on female polar metabolome data. Yet, seeing less impact on
female polar metabolomes was not an unexpected result, particularly in a suspected estrogenic
environment since females appear to have better compensatory mechanisms for estrogen
exposure than males.
24


Metabolomics is growing in utility as its use in ecotoxicology grows. As seen in this
study, metabolomics data agreed with both chemical and qPCR data, but also has great potential
for making connections between a chemical exposure and impacts on an organism at higher levels
of organization. While vtg expression is indicative of estrogenic exposure, it is difficult to
connect such expression to an effect at the organism level43. We dont always know how vtg
induction effects organism survival, reproductive success, etc. However, metabolomics data,
particularly in conjunction with transcriptomics and proteomics data, could allow scientists to
piece together functional networks that could connect an exposure response to higher level
impacts27. An organism is more than just its component parts; it is an amalgamation of parts that
work together in delicate balance. Better understanding these individual components, and how
they impact each other, will allow for better understanding of how chemicals are altering that
balance. In turn, scientists can evaluate what the implications may be for the organism, and
possibly the population as a whole.
It is also clear that qPCR can be utilized successfully to identify expression of genes of
interest. Furthermore, because mRNA is a short-lived molecule, it can be indicative of more
acute exposure than if the protein itself is measured. While the vtg protein can remain present in
an animal for weeks-months, mRNA degrades within days22. Therefore, in studies of short
duration exposures, qPCR has potentially greater utility than analyzing protein biomarkers.
Identifying and isolating gene markers for chemicals with other mechanisms of action is an
important goal going forward. It seems as though finding ways of linking more traditional, apical
well-studied biomarker responses with more molecular, systems level responses is an important
and ongoing shift.
Conclusions
The development of reliable and cost effective methods for assessing and managing the
potential impacts of chemicals found in aquatic environments is likely to become increasingly
important in coming decades. This study represents a good pilot study on a previously
25


uncharacterized watershed using methods that utilize biomarkers at different levels of
organization to clearly demonstrate impacts from endocrine disrupting chemicals, particularly
estrogenic compounds. Utilizing molecular and cellular responses in ecotoxicology studies is a
relatively new method; thus, scientists are constantly learning new information about metabolites
and genes of interest and how to interpret such data sets. Going forward, it is important to
determine more useful metabolites and genes of interest to elucidate impacts from chemicals.
Perhaps then we can find ways of understanding these complicated biological networks and how
chemical perturbations alter their function for regulatory purposes.
26


REFERENCES
1. Ekman, D. R. el al. Use of gene expression, biochemical and metabolite profiles to
enhance exposure and effects assessment of the model androgen 17??-trenbolone in fish.
Environ. Toxicol. Chem. (2011). doi:10.1002/etc.406
2. Teng, Q., Ekman, D. R., Huang, W. & Collette, T. W. Impacts of 17 alpha -
ethynylestradiol exposure on metabolite profiles of zebrafish ( Danio rerio ) liver cells.
Aquat. Toxicol. 130-131, 184-191 (2013).
3. Kolpin, D. W. el al. Pharmaceuticals Hormones and Other Organic Wastewater
Contaminants in U S Streams 1999 2000 : A National Reconnaissance. Environ. Sci.
Technol. 36, 1202-1211 (2002).
4. Kolpin, D. W., Skopec, M., Meyer, M. T., Furlong, E. T. & Zaugg, S. D. Urban
contribution of pharmaceuticals and other organic wastewater contaminants to streams
during differing flow conditions. Sci. Total Environ. 328, 119-30 (2004).
5. Limpiyakom, T., Homklin, S. & Ong, S. K. Fate of Estrogens and Estrogenic Potentials in
Sewerage Systems. Crit. Rev. Environ. Sci. Technol. 41, 1231-1270 (2011).
6. H. Jones, O. a., Voulvoulis, N. & Lester, J. N. Human Pharmaceuticals in Wastewater
Treatment Processes. Crit. Rev. Environ. Sci. Technol. 35, 401-427 (2005).
7. Boxall, A. B. A. el al. Pharmaceuticals and personal care products in the environment:
what are the big questions? Environ. Efeal. Perspect. 120, 1221-1229 (2012).
8. Brooks, B. W., Riley, T. M. & Taylor, R. D. Water quality of effluent-dominated
ecosystems: Ecotoxicological, hydrological, and management considerations.
Hydrobiologia 556, 365-379 (2006).
9. Oulton, R. L., Kohn, T. & Cwiertny, D. M. Pharmaceuticals and personal care products in
effluent matrices: A survey of transformation and removal during wastewater treatment
and implications for wastewater management. J. Environ. Monit. 12, 1956-78 (2010).
10. Mills, L. J. & Chichester, C. Review of evidence: are endocrine-disrupting chemicals in
the aquatic environment impacting fish populations? Sci. Total Environ. 343, 1-34 (2005).
11. Tyler, C. R., Jobling, S. & Sumpter, J. P. Endocrine Disruption in Wildlife: A Critical
Review of the Evidence. Crit. Rev. Toxicol. 28, 319-361 (1998).
12. Iguchi, T. & Katsu, Y. Commonality in Signaling of Endocrine Disruption from Snail to
Human. Bioscience 58, 1061 (2008).
13. Vajda, A. M. ei al. Reproductive Disruption in Fish Downstream from an Estrogenic
Wastewater Effluent. Environ. Sci. Technol. 42, 3407-3414 (2008).
14. Gimeno, S., Komen, H., Jobling, S., Sumpter, J. & Bowmer, T. Demasculinisation of
sexually mature male common carp, Cyprinus carpio, exposed to 4-tert-pentylphenol
during spermatogenesis. Aquat. Toxicol. 43, 93-109 (1998).
15. Gimeno, S., Komen, H., Gerritsen, A. G. M. & Bowmer, T. Feminisation of young males
of the common carp, Cyprinus carpio, exposed to 4-tert-pentylphenol during sexual
differentiation. Aquat. Toxicol. 43, 77-92 (1998).
27


16. Jobling, S. et al. Comparative responses of molluscs and fish to environmental estrogens
and an estrogenic effluent. Aquat. Toxicol. 65, 205-220 (2003).
17. Vajda, A. M. et al. Demasculinization of male fish by wastewater treatment plant effluent.
Aquat. Toxicol. 103,213-21 (2011).
18. Briihl, C. a, Pieper, S. & Weber, B. Amphibians at risk? Susceptibility of terrestrial
amphibian life stages to pesticides. Environ. Toxicol. Chem. 30, 2465-72 (2011).
19. Wiemeyer, S. N. et al. Organochlorine pesticide, polychlorobiphenyl, and mercury
residues in bald eagle eggs-1969-79 and their relationships to shell thinning and
reproduction. Arch. Environ. Contam. Toxicol. 13, 529-549 (1984).
20. Curtis, O. E. & Jenkins, A. R. Shell thickness and size of Peregrine Falcon, Falco
peregrinus minor eggs from two areas in South Africa. Ostrich 73, 64-66 (2002).
21. Soto, A. M., Maffini, M. V, Schaeberle, C. M. & Sonnenschein, C. Strengths and
weaknesses of in vitro assays for estrogenic and androgenic activity. Best Tract. Res. Clin.
Endocrinol. Metab. 20, 15-33 (2006).
22. Biales, A. D., Bencic, D. C., Flick, R. W., Fazorchak, J. & Fattier, D. F. Quantification
and associated variability of induced vitellogenin gene transcripts in fathead minnow
(Pimphales promelas) by quantitative real-time polymerase chain reaction assay. Environ.
Toxicol. Chem. 26, 287-296 (2007).
23. Martinovic, D., Denny, J. S., Schmieder, P. K., Ankley, G. T. & Sorensen, P. W.
Temporal Variation in the Estrogenicity of a Sewage Treatment Plant Effluent and Its
Biological Significance. Environ. Sci. Technol. 42, 3421-3427 (2008).
24. Nelson, E. D., Do, H., Fewis, R. S. & Carr, S. a. Diurnal Variability of Pharmaceutical,
Personal Care Product, Estrogen and Alkylphenol Concentrations in Effluent from a
Tertiary Wastewater Treatment Facility. Environ. Sci. Technol. 45, 1228-1234 (2011).
25. National Research Council. Toxicity Testing in the 21st Century. (2007).
doi: 10.17226/11970
26. Gerbron, M., Geraudie, P., Rotchell, J. & Minier, C. A new in vitro screening bioassay for
the ecotoxicological evaluation of the estrogenic responses of environmental chemicals
using roach (Rutilus rutilus) liver explant culture. Environ. Toxicol. 25, 510-6 (2010).
27. Garcia-Reyero, N. & Perkins, E. J. Systems biology: leading the revolution in
ecotoxicology. Environ. Toxicol. Chem. 30, 265-273 (2011).
28. Ekman, D. R. et al. Investigating compensation and recovery of fathead minnow
(Pimephales promelas) exposed to 17a-ethynylestradiol with metabolite profiling.
Environ. Sci. Technol. 42, 4188-4194 (2008).
29. Connor, S. C. et al. Development of a multivariate statistical model to predict peroxisome
proliferation in the rat, based on urinary 1H-NMR spectral patterns. Biomarkers 9, 364-85
(2004).
30. Samuelsson, F. M., Forlin, F., Karlsson, G., Adolfsson-Erici, M. & Farsson, D. G. J.
Using NMR metabolomics to identify responses of an environmental estrogen in blood
plasma of fish. Aquat. Toxicol. 78, 341-349 (2006).
28


31. Ekman, D. R. et al. Metabolite profiling and a transcriptional activation assay provide
direct evidence of androgen receptor antagonism by bisphenol A in fish. Environ. Sci.
Technol. 46, 9673-9680 (2012).
32. Lankadurai, B. P., Wolfe, D. M., Simpson, A. J. & Simpson, M. J. lHNMR-based
metabolomic observation of a two-phased toxic mode of action in Eisenia fetida after sub-
lethal phenanthrene exposure. Environ. Chem. 8, 105-114 (2011).
33. Garcia-Reyero, N. et al. Integrated approach to explore the mechanisms of aromatase
inhibition and recovery in fathead minnows (Pimephales promelas). Gen. Comp.
Endocrinol. 203, 193-202 (2014).
34. Katsiadaki, I. et al. Hepatic transcriptomic and metabolomic responses in the Stickleback
(Gasterosteus aculeatus) exposed to ethinyl-estradiol. Aquat. Toxicol. 97, 174-187 (2010).
35. Keun, H. & Ebbels, T. Geometric trajectory analysis of metabolic responses to toxicity
can define treatment specific profiles. Chem. Res. ... 17, 579-587 (2004).
36. Jensen, K. M., Korte, J. J., Kahl, M. D., Pasha, M. S. & Ankley, G. T. Aspects of basic
reproductive biology and endocrinology in the fathead minnow (Pimephales promelas).
Comp. Biochem. Physiol. Part C Toxicol. Pharmacol. 128, 127-141 (2001).
37. Ankley, G. T. & Villeneuve, D. L. The fathead minnow in aquatic toxicology: Past,
present and future. Aquat. Toxicol. 78, 91-102 (2006).
38. USEPA. Method 1694 : Pharmaceuticals and Personal Care Products in Water, Soil,
Sediment, and Biosolids by HPLC / MS / MS. EPA Method 77 (2007).
39. Schultz, M. M. et al. Antidepressant pharmaceuticals in two U.S. effluent-impacted
streams: Occurrence and fate in water and sediment and selective uptake in fish neural
tissue. Environ. Sci. Technol. 44, 1918-1925 (2010).
40. Van Den Belt, K., Berckmans, P., Vangenechten, C., Verheyen, R. & Witters, H.
Comparative study on the in vitro/in vivo estrogenic potencies of 17(3-estradiol, estrone,
17a-ethynylestradiol and nonylphenol. Aquat. Toxicol. 66, 183-195 (2004).
41. Schwindt, A. The population ecology of fathead minnows (Pimephales promelas) in
estrogen contaminated environments. (2013).
42. Finne, E. F., Cooper, G. A., Koop, B. F., Hylland, K. & Tollefsen, K. E. Toxicogenomic
responses in rainbow trout (Oncorhynchus mykiss) hepatocytes exposed to model
chemicals and a synthetic mixture. Aquat. Toxicol. 81, 293-303 (2007).
43. Cheek, A. O. et al. Experimental evaluation of vitellogenin as a predictive biomarker for
reproductive disruption. Environ. Health Perspect. 109, 681-690 (2001).
29


Full Text

PAGE 1

AN INTEGRATIVE APPROACH TO CHARACTERIZING THE ESTROGENICITY GRADIENT OF A PORTION OF THE SOUTH PLATTE RIVER by LISA MARIE BOURDON B.A., Ball State University, 2010 A thesis submitted to the Faculty of the Graduate School of the University of Colora do in partial fulfillment of the requirements for the degree of Master of Science Biology Program 2016

PAGE 2

2016 LISA MARIE BOURDON ALL RIGHTS RESERVED

PAGE 3

ii This thesis for the Master of Science degree by Lisa Marie Bourdon has be en approved for the Biology Program by Alan Vajda, Chair Amanda Charlesworth Kristen Keteles April 28, 2016

PAGE 4

iii Bourdon, Lisa Marie (M.S., Biology) An Integrative Approach to Characterizing the Estrogenicity Gradient of a Portion of the South Platte River Thesis directed by Assistant Professor Alan Vajda Abstract Endocrine disrupting compounds (EDCs) from agricultural, industrial, and municipal sources can be found in many surface waters with potential adverse implications for human and ecosystem hea lth. The South Platte River represents a significant source of water for the Denver Metro Area, yet little data exists concerning EDCs. The aim of this study is to evaluate the occurrence and effects of EDCs downstream from two major wastewater treatme nt plants (WWTPs) This study characterizes the estrogenicity gradient of the South Platte River in the Denver Metro area by combining dat a from qPCR analysis for liver vitellogenin ( vtg ) mRNA with liver NMR metabolomics after a 5 day in situ caged exposure of fathead minnows Co ncurrent water samples collected from the start and end times of the exposures were used to determine the occurrence and concentration of wastewater contaminants R esults found 68 of 122 chemicals downstream of WWTP 1 and 73 downstr eam of WWTP 2, including known EDCs (e.g. nonylphenol and octylphenol) A steroidal estrogen, estrone, was only found downstream of WWTP 2. Consistent with the highest measured concentrations of wastewater estrogens, the highest levels of vt g mRNA were m easured downstream of WWTP 2 Metabolomics data coincided with vtg data and showed little variation except downstream of WWTP 2, where male polar metabolomes showed increased levels of alanin e and glutamate, which are utilized in VTG synthesis PCA of ma le polar metabolomes showed significant separation of WWTP 2 from WWTP 1 and the reference site, further supported by PLS DA scores plot. Female polar metabolomes showed significant separation between WWTP 1 and WWTP 2 using PLS DA scores plot. This stu dy demonstrates that qPCR and metabolomics data can be reliably and concurrently used to illuminate impacts

PAGE 5

iv from chemical exposures, although further research will better elucidate target genes and metabolites of interest. The form and content of this abst ract are approved. I recommend its publication. Approved: Alan Vajda

PAGE 6

v TABLE OF CONTENTS CHAPTER I. II. METHODS ...6 Site Desc vtg NMR Sample Preparation and A III. RESULTS 13 vtg IV. 22

PAGE 7

vi LIST OF TABLES T ABLE 1. Chemical Dete 13 2. 15

PAGE 8

vii LIST OF FIGURES FIGURE 1. 7 2. Analytical Chemistry Me 11 3. ..14 4. Male vtg mRNA qPCR Interval Plot 16 5. Male Polar NMR Metabalomic s Spectr a .. .. 17 6. Male Polar NMR Metabalomics P CA for All ... 18 7. Male Polar NMR Metabalomics PCA Excluding Time 0 .. 18 8. PLS DA Male Polar Me t a bolomes for Clear Creek and WWTP 2 19 9. PLS DA 19 10. F emale Polar NMR Metabalomics PCA Outlier ... .. 20 11. Female Polar Metabolomics PCA for All Classes 21 12. PLS DA Female Polar Metabolomes ... 21

PAGE 9

viii LIST OF ABBREVIATIONS WWTP Wastewater Treatment Plant VTG Vitellogenin protein vtg Vitellogenin gene NMR Nuclear Magnetic Resonance qPCR Quan t itative Polymerase Chain Reaction PCA Principle Components Analysis PLS DA Partial Least Squares Discriminant Analysis EE2 ethynylestradiol

PAGE 10

1 CHAPTER I INTRODUCTION With an ever growing list of chemicals used for a plethora of anthropogenic uses, there is a growing concern for how these chemicals may be a ffecting the environme nt as well as human health. A n important s ource of chemicals in the environment is municipal wastewater treatment plant (WWTP) effluent, which is known for containing chemicals that can impact the endocrine system 1 5 M unicipal WWTP effluent contains the remnants of personal care products and excreted pharmaceuticals, which oftentimes pass through WWTPs and enter waterways where their fate is unknown 3,5 7 Personal care products often contain chemicals implicated in endocrine disruption, and many pharmaceuticals also have potential adverse effects on the environment and warra nt further study 8,9 Pharmaceuticals in the environment were long thought to ex ist at concentrations to o low to cause concern; thus, data on th eir presence and persistence in the environment is less extensive 8 Furthermore, methods of detecting chemicals in the environment were, and in many ways still are, limited 3 Ano ther troubling facet of WWTP effluent is its constancy While runoff from other sources, such as agriculture, are intermittent and short lived, WWTP effluent discharges at r egular intervals and can create pseudo persistent exposures that can alter normal biological and ecological processes 8 Previous research concerning chemicals in the environment, particularly those that are known or suspected to affect the normal function of the endocrine system indicates that there are instances of perturbations in normal reproductive function, and in some cases, survivorship across taxa 10,11 There are many studies evaluating altered en dpoints due to endocrine disruption in fishes, including; skewed sex ratios, altered gametogenesis, disrupted gonadal development, intersex, and alterations in hormone mediated functions, such as vitellogenin protein (VTG) production 12 17 Studies in amphibians link endocrine disruption to deformities, lowered immune function, and lower surv ivability 12,1 8 Developmental exposure of r eptiles can lead to altered sex

PAGE 11

2 ratios 12 as well as altered plasma steroid levels and abnormal gonadal development 11 Predatory birds exposed to estrogenic compounds may be susceptible to population declines due to eggshell thinning and resulting reduced offspri ng survivorship 11,19,20 Mammals have had reported incidences of altered gametogenesis, and increased incidences of reproductive cancers 11,12 Humans specifically have had increased incidences of birth defects in males such as hypospadias, cryptorchidism and shortened anogenital distance, and early menarche in female s 12,21 Even some invertebrates have demonstrated reproductive impairment associated with endocrine disruptor exposure, although their utilization o f steroid hormones is unclear and requires further study 16 Thus, it is evident that because of the well conserved nature of the endocrine system, chemicals are eliciting detrimental effects across taxa including humans. Because of this, it is important that better methods of detection and evaluation of risk assessment be developed to identify chemicals that warrant the most concern. R apid, cost effective and sensitive methods for risk assessment are necessitated by the complex ity of con taminant mixtures discharged into the environment. The number of chemicals, and their potential mechanisms of action, are simply too numerous to do comprehensive toxicity testing on every individual chemical, let alone of environmentally relevant mixtures Furthermore, our analytical methods offer an incomplete picture of the chemicals present in a mixture at any given time 22 Analytical chemi cal methods are generally limited to known chemicals, ca nnot account for mixture effects, and c annot practically account for temporal variations 3,22 24 The National Research Council has challenged t he scientific community to not only find methods that eschew traditional animal testing and rely more on cell based assays while also linking adverse outcomes to exposures 25 As scientists better understand how cellular respon se s are connected to one another it is possible that biological pathways associated with adverse health effects from exposures can be identified and utilized to recognize exposures in the future. More traditional biomarker studies are generally centered around isolated measurements;

PAGE 12

3 including: secondary sex characters, gamete production, plasma sex steroid concentrations, and VTG levels 1 VTG, an egg yolk precursor protein produced in many egg laying females, is estrogen induced and is generally only produced in male organisms impacted by estrog enic compounds, making it a useful biomarker of estrogen exposure 26 However, traditional biomarker studies can be limited because they are a single component in a complicated framework that can be altered by any number of changes to other components 27 and impacts on that single component are not always informative of effects at higher levels of organization Recent years have seen the development of more systems based approaches to studying the effects of chemical exposures on organisms, including large scale, high throughput studies of transcriptomics, proteomics, and metabolomics. Obtaining data about the networks of gene transcripts, proteins and metabolites, and how exposures impact those networks, shows great potential for improved specificity in testing for organism exposure effects 27 Metabolomics in particular has great utility in toxicology studies because changes t o the metabolome occur quickly, and because alterations in the metabolome directly impact organism ic function, while transcriptomics changes do not necessa rily translate to direct organism impacts 28 Furthermore, NMR is generally less expensive than other high throughput methods like microarrays and still yields large data sets for analysis. There is also the potential for non invasive sample collections, such as the use of urine samples or other body fluids, which could greatly limit animal usage 29 Metabolomics has been utilized in recent years to evaluate the effects o f chemical exposures on organisms with reproducible success and several common metabolite shifts 30 34 While our ability to produce these data sets has preceded our ability to fully analyze them, our knowledge of reading metabolomics NMR spectra increases all the time and elucidates new metabolites of interest 35 The South Platte River Basin, as well as other waterways in arid or semi arid environments faces a particular p roblem with chemical loading due to the disproportionately high impact of WWTP effluents discharging to relatively low base flow streams 8,24 During low

PAGE 13

4 flow times of year, WWTP effluent can comprise >90% of stream flow. This persistent influx of effluents coupled with limited dilution, can modify downstream habitat s particularly in terms of temperature, DO, salinity, nutrient and chemical loading, and overall water quality 8 This problem is further compounded by the fact that many chemicals do not have drinking water guidelines, drinking water health advisories, or aquatic life criteria 3 Since the human population is expected to increase substantially in the coming decades, municipalities will increasingly come to rely on water reclamation for drinking water resources, particularly in drier areas, and that there will be a corresponding increase in pharmaceutical and personal care products entering WWTPs 8 Thus, it is increasingly important that scientists develop better methods of detecting chemicals and evaluating their impacts for regulatory purposes. The research in this manuscript targets the wastewater dominated urban corridor of the South Platte River through De nver, CO, USA Denver Metro WWTP is the largest facility in the region, and serves a population of approximately 1,700,000 with a discharge volume of 130 million gallons/day. WWTP effluent comprises over 90% of stream flow in the South Platte River down stream of the Denver Metro WWTP. The effluent dominated South Platte River downstream from the Denver Metro WWTP is utilized as a source of drinking water for the rapidly expanding urban community of Aurora, CO, and persistent concentrations of CECs presen ts an emerging eco human health challenge. The goal of this study is to utilize multiple systems level assays integrated with chemical sampling to assess the potential estrogenicity of a highly populated, heavily wastewater impacted portion of the South P latte River and determine their efficacy as useful tools to screen for e xposure to estrogenic chemicals. Following a 5 day in situ caged fish exposure a quantitative polymerase chain reaction (qPCR) assay was used to quantify liver vtg mRNA transcripts to evaluate estrogenic exposure from the sites. Concurrent liver NMR metabolomics was performed to evaluate the utility in using metabolomics as a tool for assessing chemical exposure

PAGE 14

5 Metabolomics has the potential to be informative of exposures other tha n estrogenic exposures, unlike the presence of vtg which is specific to estrogen exposure. Because of the high throughput nature of metabolomics data, inferences at higher levels of organization can be made. Transcript and metabo lomic results are interp reted in the context of the measured concentration and occurrence of WWTP contaminants

PAGE 15

6 CHAPTER II METHODS Overview This study is focused on a 5 day in situ caged fish study that took place downstream of 2 WWTPs from 08/09/2012 to 08/14/201 2. At the time of fish deployment and retrieval, water samples were collected from upstream and downstream of both WWTPs, as well as at a reference site. After retrieval, fish were sacrificed and their livers dissected out for analysis of vtg gene expres sion as well as NMR metabolomics analysis. Water samples were analyzed for wastewater contaminants. Site Description Two WWTPs that discharge their effluents into the South Platte River, Denver, Colorado were chosen for this study. The wastewater trea tment plants will be referred to as WWTP 1 and WWTP 2. WWTP 1 is located at 2900 S. Platte River Drive, Englewood, CO 80110 (39.6587015 105.00228440000001) WWTP 2 is located at 6450 York St, Denver, CO 80229 (39.8149471, 104.95789450000001) Both WW TPs have primary treatment, activated solids, solids removal, and chlorine disinfection. WWTP 1 receives approximately 83 x 10 6 L/d and has a nitrifying trickl ing filter treatment step as well as de nitrification for nutrient removal. WWTP 2 receives app roximately 454 x 10 6 L/d and has two treatment complexes, North and South. At WWTP2, t ertiary treatment, biological nutrient removal, only occurs in the North complex. A reference site was chosen in Clear Creek upstream from Golden, CO which is a relati vely un impacted site that has a well researched hydrology and at which chemical loading is minimal (Figure 1)

PAGE 16

7 Figure 1: Site Map Map of North Eastern Colorado, including the Sout h Platte River and Clear Creek. Model Organism The fathead minnow ( Pime phales promelas ) was chosen as the study organism because they are a member of a ubiquitous family of fish that is common in chemically impacted environments 36,37 There are also a commonly used species in ecotoxicology studies and substantial data alrea dy exists to establish reliable biomarkers for chemical exposure. Furthermore, adequa te genomic data is available to utilize fathead minnows for genetic based testing 37 Adult fathead minnows were provided by the U.S. EPA Office of Research and Development, Molecular Ecology Research Division in Cincinnati OH. Field Exposure Fish were deployed in situ in cages for 5 days at 3 sites beginning on 08/09/2012 and ending on 08 /14/2012 Fish were transported to sites in oxygenated bags set in coolers with ice to enhance survival, and then acclimated to the stream by replacing the ice with stream water. The fish were further acclimated to stream water by mixing a 50:50 mixture of lab water and stream water for 30 minutes. 15 male and 15 female fathead minnows were put in cages composed of a PVC pipe with mesh on each side and deployed at sites downstream of both WWTP 1 and WWTP

PAGE 17

8 2 along the South Platte River as well as at Clear Creek above Golden, CO as a reference site. The cages were suspended above the stream bed by attaching nylon cord to a stake or to an overhanging limb, etc. At the time of deployment, 15 male and female fish were sacrificed as time 0 initial controls A t the time of fish deployment and retrieval, water samples were collected from the aforementioned sites as well as upstream of both WWTPs for chemical analysis to evaluate PPCPs, Pesticides, and Waste Indicators. After 5 days of exposure, the fish were recovered from cages and transported in oxygenated buckets to a laboratory at WWTP 2 for dissection. Dissection After the 5 day exposure, fish were anesthetized by immersion in MS 222(0.1g/L). Plasma samples were collected from the caudal vein for future studies and then fish were sacrificed by rapid decapitation. Fish were cut from cloaca to operculum and the internal organs removed. The liver was carefully removed from the intestine and surrounding organs, and then divided into sections for qPCR anal ysis of vtg and NMR metabolomics analysis, placed into labeled cryotubes snap frozen and stored 80 C until analys is The gonads were then removed from the abdominal cavity, snap frozen, and stored until 80 C for future analysis. Gloves were changed, and dissection tools were cleaned in 70% ethanol between dissections to prevent cross contamination between fish or between organs. vtg mRNA Extraction. Samples were stored at 80C until extraction. Samples were prepared as previously described by Bial es et al. 2007 22 on male liver samples. Each sample was placed in a labeled microcentrifuge tube with a steel ball and 1mL of tri reagent. Samples were homogenized for 5 minutes and then allowed to sit for 10 minutes to allow for complete extraction of genetic material. 100L of bromochloropropane were added to each sample tube and the tub es allowed to stand for an additional 10 minutes to help improve yield. Samples were centrifuged at 12,000rpm and 4C for 15 minutes, then the aqueous layer pipetted off into freshly labeled

PAGE 18

9 microcentrifuge tubes containing 500L of isopropanol to precipitate the genetic material. Samples were centrigued for 10 minutes at 12,000 rpm and 4C and the isopropanol po ured off of the samples. 1mL of 75% ethyl alcohol was added to each tube to wash the samples, after which the samples were centrifuged for 8 minutes at the previously stated specifications. Following centrifugation, the ethyl alcohol was pipetted off and DEPC water was added until the RNA pellet was dissolved. The RNA concentrations were quantified using a Nanodrop 1000. Samples were frozen at 80C if cDNA was not prepared immediately. cDNA. Thirteen samples at a time were diluted in microAMP tubes t o 500ng/L and run on a thermal cycler for 10 minutes with 3 controls, one blank with only DEPC water, one that contained no reverse transcriptase and one that contained a high expressing pool of vitellogenin. 2L of each sample and control were added to a well in a 96 well plate with 18L of master mix composed of PCR buffer, MgCl 2 solution, oligo d(t) 23 VN, dNTP mix, random hexamers, RNase inhibitor, and MuLV reverse transcriptase (for all samples excluding the applicable control). The plate was seale d and run on the thermal cycler for 60 minutes after which 80L of DEPC was added to each sample well. Samples were frozen at 80C if qPCR was not performed immediately. qPCR. 18L of two different mastermixes were pipetted into opposite halves of a 96 well plate. One mastermix was for vtg RNA and was composed of DEPC, syber, and left and right vtg primers while the other targeted 18s RNA and was composd of DEPC, syber, and 18s primer. 2L of each of the 16 samples/controls was pipetted into three well s on either half of the 96 well plate and the qPCR run on the PCR thermal cycler. NMR Sample Preparation and Analysis Samples were kept chilled during the preparation process. Samples were prepared as described by Ekman et al. 2008 28 The liver samples were weighed in a pre tared microcentrifuge tube and then a steel ball, 400 L of methanol and 85 L of milli Q water were added. The samples were homogenized on a vortexer for five minutes or until samples were well

PAGE 19

10 homogenized and t hen centrifuged fo r 60 seconds at 1,000 g in a 4C microfuge. 300 L of chloroform w as added and the samples vortexed for 60 seco nds and centrifuged again at 1,000 g and 4C for one minute 200 L of chloroform and 200 L of Milli Q water were added and th e samples were vortexed again for 60 seco nds, and then centrifuged at 1,000 g and 4C for 15 minutes This centrifugation step separate d the upper water/methanol phase, containing polar metabolites from the lower chloroform phase, containing lipophilic met abolites with a layer of protein debris in the middle. The upper phase was removed via pipette and put into a labeled microcentrifuge tube and the lower phase was transferred into a labeled glass vial, being careful with both to not contaminate with the p rotein debris layer in th e middle. Samples were dried via nitrogen blow down and then 600 L of methanol was added to bring samples to an appropriate volume for NMR analysis. The samples were then pipetted into labeled NMR sample tubes and assembled int o the spinner for loading into the NMR autosampler for analysis. Acquired polar spectra were zero filled, line broadened at 0.3 Hz, and Fourier transformed (ACD/1D NMR Manager, Advanced Chemistry Development). Using an automated routine, spectra were pha se and baseline corrected, referenced to TSP, and binned at a width of 0.005 ppm. Regions of binned data were excluded to e 7.80). Remaining bins were then normalized to unit total integrated intensity. Analytical Chemistry Water samples were collect ed in 1L amber glass containers, adding 80mg of sodium thiosulfate to remove residual chlorine as needed when caged fish were deployed and again upon their re trieval One glass container each was collected for acid and base analysis. If visible particle s were present in the water samples they were filtered out and the aqueous portion of the particles removed prior to analysis using vacuum filtration. E xtraction and quantification of the chemicals of interes t followed USEPA methods 1694 and 1694 for pharm aceuticals and personal care products in water/soil/sediment/biosolids and steroids and hormones in

PAGE 20

11 water/soil/sediment/bi osolids respectively (Figure 2 ). Chemical data were analyzed qualitatively for the purposes of this study. Figure 2: Analytical C hemistry Methods Flow chart for determination of pharmaceutical and personal care products by LC/MS/MS 38 Statistical Analysis qPCR All data transformation was performed using LinRegPCR software, and statistical analysis performed using Minitab St atis tical Software v. 17. data were used to perform an analysis of variance with Bonferroni correction to reduce chance of Type I error Significance levels were set at 0.05.

PAGE 21

12 NMR Principal component analysis (PCA) was use d to evaluate overall patterns in the data (SIMCA 2 test at the 95% confidence interval. Then partial least squares discriminant analysis (PLS DA) models were created to help visually determine the extent of impact fo r a given treatment, in relation to impacts of different treatments. CV ANOVA (A n alysis o f V a riance testing of Cross Validated predictive residuals) was used within SIMCA P+ for PLS DA mo dels to assess their validity. Class s core values in PLS DA scores plots were averaged and are shown wit h the associated standard error.

PAGE 22

13 CHAPTER III RESULTS Analytical Chemistry Of 122 tested chemicals, none were detected at the Clear Creek reference site. 13 were detected u pstream of WWTP 1, spiking to 68 chemicals in the effluent. Upstream of WWTP 2, 24 chemicals were detected, spiking to 73 chemicals downstream of WWTP 2 (Table 1). Furthermore, there were 58 common chemicals that were found downstream of both WWTPs (Figure 3 ). There were several ch emicals that were found in both effluents including: DEET, galaxolide, atenolol, tramadol, caffeine, and several organophosphates which are commonly used as pesticides. Both WWTP effluents also had measurable levels of common wastewater contaminants inc luding nonylphenol, fluoxetine, 4 tert octylphenol, triclosan, and metoprolol The steroidal estrogen estrone and the endocrine disrupting herbicide A trazine were only found in WWTP 2 effluent samples (Table 2). As seen in Table 2, several of the chemi cals measured were substantially higher, in some cases 2 3 times higher, in the WWTP 2 effluent compared to WWTP 1 effluent. Table 1: Chemical Detections Comparisons of the numbers of chemicals found at each site sampled for analytical chemistry, as well as the number of those chemicals that was unique to that sample site in this study. Site # of Chemicals Found # of Unique Chemicals Found Clear Creek (Reference) 0 0 Upstream WWTP 1 13 0 Effluent WWTP 1 68 8 Upstream WWTP 2 24 1 Effluent WWTP 2 73 7

PAGE 23

14 Figure 3: Shared and Unique Chemical Detections Numbers of shared or unique chemical detections by sites.

PAGE 24

15 Table 2: Detected Concentrations Measured concentrations of some detected chemicals by sample site (ng/L). Chemicals are denoted by endocrine impacts with blue indicating chemicals with suspected endocrine impacts, green indicating chemicals with both estrogenic and androgenic impacts, red indicating chemicals with estrogenic impacts, yellow indicating chemicals with observe d endocrine impacts with unknown mechanisms of action, and purple indicating steroidal estrogens. Chemical Upstream WWTP 1 Effluent WWTP 1 Upsteam WWTP 2 Effluent WWTP 2 DEET 41.6 377, 356, 467 39.6 168, 257 Metropolol ND 1030, 839, 1300 10.7 510, 609 Atenolol 16.7 371, 453, 295 11.4 2140, 2790 Tramadol 37.6 1560, 1710, 1320 58.1 1140, 1180 Galaxolide 149 3360, 3320, 3860 103 3480, 4140 Tris(2 butoxyethyl) Phosphate 173 1720, 800, 1850 97.4 1240, 2680 Tris(dichloroisopropyl) Phosphate 84.0 752, 783 787 143 716, 944 Nonylphenol ND 264, 285 ND 700, 772 4 tert octylphenol ND 91.9, 75.1, 97.8 ND 335 Triclosan ND 42.1, 321, 29.4, 38.6, 303 ND 27.6, 11.0, 540, 1010 Bisphenol A ND 214 ND ND Fluoxetine ND 20.1, 11.0, 21.3 ND 15.0, 16.7 Caffeine 29.5 209, 192, 234, 274 54.4, 51.1 189, 251, 297, 421 Atrazine ND ND 24.1, 30.4 32.1, 55.0 Estrone ND ND ND 169

PAGE 25

16 vtg qPCR Relative expression of vtg in males was normalized to Time 0 and was statistically absent a t the Clear Cr eek reference site and WWTP 1. H owever, vtg expression was significantly higher at WWTP 2 (p<0.05) than the other sites s howing an increase of at least 2 orders of magnitude (Figure 4 ). Figure 4 : Male vtg mRNA qPCR Interval Plot Relative quantities of vtg gene expression at Ti me 0, Clear Creek Reference, and both WWTPs 1 and 2. NMR Metabolomics Male Metabolomes. Polar m etabolite profiles of male fathead minnows showed minimal impact at WWTP 1, but alanine and glutamate peaks were observed to s ignificantly increase at WWTP 2 ( Figure 5 ) Due to small liver size, relatively few liver fragments were collected for NMR metabolomics analysis from fish at the Clear Creek Site limiting comparisons PCA analysis of the metabolite profiles reveals that when all 4 classes are evaluated together, Time 0 fish separate

PAGE 26

17 the most from the other three groups (Figure 6 ) T his is potentially d ue to the s tress of transport and not having been fed during transport. Removal of Time 0 fish from the PCA emphasizes the separation of WWTP 2 from the other two sites (Figure 7 ) Follow up PLS DA analysis for classes excluding Time 0 fish demonstrate significant separation of WWTP 2 metabolomes from the other two classes (p=0.0059), as do es analysi s of both Clear Creek reference and WWTP 1 analyzed wit h WWTP 2 individually, (p=0.0007 and p=6.7e 005 respectively) (Figure 8, Figure 9 respectively) as the model was not valid when all three classes were combined PLS DA maximizes differences between groups, when model is validated with CV ANOVA; thus, it is evident that there is a statistically significant difference between the metabolomes of males at WWTP2 compared to Clear Creek (reference) and WWTP1. Figure 5 : Male Polar NMR Metabolomics Spectra Characteristic NMR spectra of male fish liver s downstr eam of WWTP 1 (bottom) and male fish downstream of WWTP 2 (top) minus the spectra of metabolites from the male fish at the Clear Creek reference site.

PAGE 27

18 Figure 6 : Male Polar PCA for All Classes PCA for male polar metabo lomes for all treatment classes to determine broad differences in metabolomes by site. Figure 7 : Male Polar NMR Metabolomics PCA Excluding Time 0 PCA for male polar metabolom es for classes excluding Time 0 to evaluate broad differences in metabolomes of Clear Creek, WWTP 1, and WWTP 2.

PAGE 28

19 Figure 8 : PLS DA Male Polar Meta bolomes for Clear Creek and WWTP 2 PLS DA for Clear Creek male and WWTP2 male metabolomes to maximize separation between Clear Creek and WWTP 2 metabolomes. Fi gure 9 : PLS DA Male Polar Metabolomes for WWTP 1 and WWTP 2 PLS DA for WWTP1 male and WWTP2 male polar metabolomes to maximize differences between WWTP 1 and WWTP 2 metabolomes.

PAGE 29

20 Female Metabolomes Female sample 25 was identified as a strong outlier due to ext remely high levels of bile acids observed an d was removed from the data set as aberrant data (Figure 10 ) PCA demonstrated some separation (Figure 11 ), and the PLS DA model for WWTP1 and WWTP2 female polar metabolomes showed significant differences between the two sites (p= 0.035) (Figure 12 ). As with male samples, there were few female liver samples obtaine d for the Clear Creek reference site. Figure 10 : Female Polar NMR Metabolomics PCA Outlier PCA of all female metabolome data, demonstrating sample F25 as an extreme outlier.

PAGE 30

21 Figure 11 : Fe male Polar Met abolomics PCA for All Classes PCA for polar female metabol ites for all treatment classes an d excluding the extreme outlier to demonstrate broad separation of metabolomes by site. Figur e 12 : PLS DA Female Polar Metabolomes for WWTP 1 a nd WWTP 2 PLS DA for female polar metabolomes at WWTP1 an d WWTP2 to maximize differences between metabolomes

PAGE 31

22 CHAPTER IV DISCUSSION AND CONCLUSIONS Discussion Utilizing biomarkers at multiple levels of biological organization combined with analytical chem istry methods allowed for characterization of the estrogenicity gradient of a portion of the South Platte River, Denver, CO. Because of the transient, dynamic nature of chemical mixtures in aquatic environment s one would expect differences in observed ch emi cal loading from WWTP effluents, even those in close proximity, with varying degrees of impact on endocrine function in organisms present 23,24 Indeed, t hese preliminary investigations did identify WWTPs as a major source of CECs with the potential to adversely impact endocrine function in fishes in the South Platte River basin. We have characterized the occurrence and concentration of diverse WWTP contaminants including natural and synthetic estrogens and neuro active pharmaceuticals. Both WWTP effluents contained a complex and dynamic mixture of wastewater contaminants that varied in composition between sampling dates; however, there was a greater occurrence and concentration of known estrogenic CECs (e.g. estrone) downstream from WWTP 2 than downstream from WWTP 1. It is also notable that chemicals were detected upstream of WWTP 2 that were not detected at other upstream sites, some of which have potential agricultural or groundwater non point sources (e.g. atrazine) (Table 2) Still others were found downstream of WWTP 1, but were not detected upstream of WWTP 2, indicating possible partitioning, dilut ion, or degradation mechanisms at work (e.g. nonylphenol, fluoxetine, 4 tert octylphenol) (Table 2) Similar complexity and diversity in chemical mixtures, particularly in terms of the occurrence of pharmaceuticals and personal care products in waterways has been observed in waterways across the country 3,39 This higher incidence of chemicals downstream from WWTP 2 is consistent with the increased vtg mRNA levels found in male fish deployed downstream of that site. The expression of vtg was found to be at least 4 orders of magnitude higher downstream of WWTP 2 compared

PAGE 32

23 to all other test sites. VTG production in f ish es can be uti lized to determine its ethynylestradiol ( EE2 ) estrogen equivalency 40 Employing EE2 induced VTG expression data in fathead minnows obtained by Schwindt 41 we are able to est imate an EE2 equivalency of ~5.0 ng/L based on relative vtg expression observed at WWTP 2. The measured concentrat ion s of steroidal ( estrone ) and non steroidal (nonylphenol, and ocylphenol) estrogens at this site supports this estimate. The observed increase in vtg mRNA in males also agrees with the altered WWTP 2 metabolomes, particularly the observed increase in alanine and glutamate in males, which have both been linked to increased production of vtg 2 Furthermore, it is evident that some level of disturbance occurred in both males and females due to the significant shifts seen in WWTP 2 metabolomes compared to WWTP 1. While this impact was greater in males than in females, it is evident that some metabolic impact is observable in bot h genders. These gender specific shifts have been observed in other studies as well. A study done on f athead minnows exposed to EE2 showed greater impact on male metabolomes compared to females 28 while the inverse was observed in fat trenbolone, an adrogenic compound 1 These data, combined with a growing body of work by other labs 1,2,28,30,32,34,35 demonstrate that integrative approaches to assessing impacts o f chemical exposures are becoming progressively useful. The agreement we see between chemical, vtg gene expression, and metabolomics data indicates that these methods can be successfully utilized to obtain much needed data concerning the presence of estro genic compounds in uncharacterized waterways. It is evident from this study and those mentioned above that analytical chemistry alone cannot provide an accurate picture of the chemicals present in a system; particularly because our detection limits were n ot low enough to detect many environmentally relevant concentrations of chemicals. There is great potential that chemicals were present in the water samples that we simply could not detect.

PAGE 33

24 Likewise, it is clear that utilizing a biomarker like vitelloge nin is only useful if certain chemicals are predominating in a chemical mixture. While vtg expression in a sample reliably predicts exposure to an estrogenic compound, the lack of such expression cannot necessarily be used to assume a lack of estrogenic c ompounds in a chemical mixture. A study done by Finne et al. found that exposing rainbow trout hepatocytes to chemicals with different known or suspected mechanisms of action resulted in a mitigated gene expression response compared to what would be expec ted from a single chemical exposure 42 Because of this mitigating phenomenon, it is imposs ible to determine if a lack of vtg expression in a sample is due to a lack of estrogenic compounds, or if there are other chemicals present that may be mitigating the estrogenic response. Thus; it is important to carefully choose biomarker tests understa nd their limitations, and ideally pair methods together than can reduce such limitations While metabolomics data has great potential, for the time being its applications are still somewhat limited. In our study, our metabolomics data was complicated by t he fact that our reference site yielded few male or female fish with livers substantial enough for metabolomics testing. During necropsy the liver was divided for qPCR and metabolomics analysis; however, if livers were too small, qPCR was prioritized over metabolomics. While a lack of livers for metabolomics could be indicative of smaller livers, which in turn could indicate healthier, less chemical impacted fish; it could also have been a consequence of novice dissectors. In future, devising a way of ob taining an accurate liver weight measurement during necropsy could help shed light on a potential cause for such issues. Thus, it is possible that the few livers obtained from the female fish may have confounded our ability to elucidate the relationship o f the Clear female polar metabolomes was not an unexpected result, particularly in a suspected estrogenic environment since females appear to have better compensator y mechanisms for estrogen exposure than males

PAGE 34

25 Metabolomics is growing in utility as its use in ecotoxicology grows. As seen in this study, metabolomics data agreed with both chemical and qPCR data, but also has great potential for making connections betw een a chemical exposure and impacts on an organism at higher levels of organization. While vtg expression is indicative of estrogenic exposure, it is difficult to connect such expression to an effect at the organism level 43 vtg induction effects organism survival, reproductive success, etc. However, metabolomics data, particularly in conjunction with transcriptomics and proteomics data, could al low scientists to piece together functional networks that could connect an exposure response to higher level impacts 27 An organism is more than just its component parts; it is an amalgamation of parts that work toge ther in delicate balance. Better understanding these individual components, and how they impact each other, will allow for better understanding of how chemicals are altering that balance. In turn, scientists can evaluate what the implications may be for the organism, and possibly the population as a whole. It is also clear that qPCR can be utilized successfully to identify expression of genes of interest. Furthermore, because mRNA is a short lived molecule, it can be indicative of more acute exposure tha n if the protein itself is measured. While the vtg protein can remain p resent in an animal for weeks months, mRNA degrades within days 22 Therefore, in studies of short duration exposures, qPCR has potentially greater utility than analyzing protein biomarkers. Identifying and isolating gene markers for chemicals with other mechanisms of action is an important goal going forward. It seems as though finding ways of linking more traditional, apica l w ell studied biomarker responses with more molecular, systems level responses is an important and ongoing shift. Conclusions The develop ment of reliable and cost effective methods for assessing and managing the potential impacts of chemicals found in aquati c environments is likely to become increasingly important in coming decades. This study represents a good pilot study on a previously

PAGE 35

26 uncharacterized watershed using methods that utilize biomarkers at different levels of organization to clearly demonstrat e impacts from endocrine disrupting chemicals, particularly estrogenic compounds. Utilizing molecular and cellular responses in ecotoxicology studies is a relatively new method; thus, scientists are constantly learning new information about metabolites an d genes of interest and how to interpret suc h data sets. Going forward, it is important to determine more useful metabolites and genes of interest to e lucidate impacts from chemicals. Perhaps then we can find ways of understanding these complicated biolo gical networks and how chemical perturbations alter their function for regulatory purposes

PAGE 36

27 R EFERENCES 1. Ekman, D. R. et al. Use of gene expression, biochemical and metabolite profiles to enhance exposure and effects assessment of the model androgen 17?? trenbolone in fish. Environ. Toxicol. Chem. (2011). doi:10.1002/etc.406 2. Teng, Q., Ekman, D. R., Huang, W. & Collette, T. W. Impacts of 17 alpha ethynylestradiol exposure on metabolit e profiles of zebrafish ( Danio rerio ) liver cells. Aquat. Toxicol. 130 131, 184 191 (2013). 3. Kolpin, D. W. et al. Pharmaceuticals Hormones and Other Organic Wastewater Contaminants in U S Streams 1999 Environ. Sci. Technol. 36, 1202 1211 (2002). 4. Kolpin, D. W., Skopec, M., Meyer, M. T., Furlong, E. T. & Zaugg, S. D. Urban contribution of pharmaceuticals and other organic wastewater contaminants to streams during di ffering flow conditions. Sci. Total Environ. 328, 119 30 (2004). 5. Limpiyakorn, T., Homklin, S. & Ong, S. K. Fate of Estrogens and Estrogenic Potentials in Sewerage Systems. Crit. Rev. Environ. Sci. Technol. 41, 1231 1270 (2011). 6. H. Jones, O. a., Voulv oulis, N. & Lester, J. N. Human Pharmaceuticals in Wastewater Treatment Processes. Crit. Rev. Environ. Sci. Technol. 35, 401 427 (2005). 7. Boxall, A. B. A. et al. Pharmaceuticals and personal care products in the environment: what are the big questions? E nviron. Heal. Perspect. 120, 1221 1229 (2012). 8. Brooks, B. W., Riley, T. M. & Taylor, R. D. Water quality of effluent dominated ecosystems: Ecotoxicological, hydrological, and management considerations. Hydrobiologia 556, 365 379 (2006). 9. Oulton, R. L. Kohn, T. & Cwiertny, D. M. Pharmaceuticals and personal care products in effluent matrices: A survey of transformation and removal during wastewater treatment and implications for wastewater management. J. Environ. Monit. 12, 1956 78 (2010). 10. Mills, L J. & Chichester, C. Review of evidence: are endocrine disrupting chemicals in the aquatic environment impacting fish populations? Sci. Total Environ. 343, 1 34 (2005). 11. Tyler, C. R., Jobling, S. & Sumpter, J. P. Endocrine Disruption in Wildlife: A Cri tical Review of the Evidence. Crit. Rev. Toxicol. 28, 319 361 (1998). 12. Iguchi, T. & Katsu, Y. Commonality in Signaling of Endocrine Disruption from Snail to Human. Bioscience 58, 1061 (2008). 13. Vajda, A. M. et al. Reproductive Disruption in Fish Downs tream from an Estrogenic Wastewater Effluent. Environ. Sci. Technol. 42, 3407 3414 (2008). 14. Gimeno, S., Komen, H., Jobling, S., Sumpter, J. & Bowmer, T. Demasculinisation of sexually mature male common carp, Cyprinus carpio, exposed to 4 tert pentylphen ol during spermatogenesis. Aquat. Toxicol. 43, 93 109 (1998). 15. Gimeno, S., Komen, H., Gerritsen, A. G. M. & Bowmer, T. Feminisation of young males of the common carp, Cyprinus carpio, exposed to 4 tert pentylphenol during sexual differentiation. Aquat. Toxicol. 43, 77 92 (1998).

PAGE 37

28 16. Jobling, S. et al. Comparative responses of molluscs and fish to environmental estrogens and an estrogenic effluent. Aquat. Toxicol. 65, 205 220 (2003). 17. Vajda, A. M. et al. Demasculinization of male fish by wastewater tre atment plant effluent. Aquat. Toxicol. 103, 213 21 (2011). 18. Brhl, C. a, Pieper, S. & Weber, B. Amphibians at risk? Susceptibility of terrestrial amphibian life stages to pesticides. Environ. Toxicol. Chem. 30, 2465 72 (2011). 19. Wiemeyer, S. N. et al. Organochlorine pesticide, polychlorobiphenyl, and mercury residues in bald eagle eggs 1969 79 and their relationships to shell thinning and reproduction. Arch. Environ. Contam. Toxicol. 13, 529 549 (1984). 20. Curtis, O. E. & Jenkins, A. R. Shell thicknes s and size of Peregrine Falcon, Falco peregrinus minor eggs from two areas in South Africa. Ostrich 73, 64 66 (2002). 21. Soto, A. M., Maffini, M. V, Schaeberle, C. M. & Sonnenschein, C. Strengths and weaknesses of in vitro assays for estrogenic and andr ogenic activity. Best Pract. Res. Clin. Endocrinol. Metab. 20, 15 33 (2006). 22. Biales, A. D., Bencic, D. C., Flick, R. W., Lazorchak, J. & Lattier, D. L. Quantification and associated variability of induced vitellogenin gene transcripts in fathead minnow (Pimphales promelas) by quantitative real time polymerase chain reaction assay. Environ. Toxicol. Chem. 26, 287 296 (2007). 23. Martino Temporal Variation in the Estrogenicity of a Sewage Treatment Plant Effluent and Its Biological Significance. Environ. Sci. Technol. 42, 3421 3427 (2008). 24. Nelson, E. D., Do, H., Lewis, R. S. & Carr, S. a. Diurnal Variability of Pharmaceutical, Personal Care Product, Estrogen and Alkylphenol Concentrations in Effluent from a Tertiary Wastewater Treatment Facility. Environ. Sci. Technol. 45, 1228 1234 (2011). 25. National Research C ouncil. Toxicity Testing in the 21st Century (2007). doi:10.17226/11970 26. Gerbron, M., Geraudie, P., Rotchell, J. & Minier, C. A new in vitro screening bioassay for the ecotoxicological evaluation of the estrogenic responses of environmental chemicals u sing roach (Rutilus rutilus) liver explant culture. Environ. Toxicol. 25, 510 6 (2010). 27. Garcia Reyero, N. & Perkins, E. J. Systems biology: leading the revolution in ecotoxicology. Environ. Toxicol. Chem. 30, 265 273 (2011). 28. Ekman, D. R. et al. Inv estigating compensation and recovery of fathead minnow (Pimephales promelas) exposed to 17 ethynylestradiol with metabolite profiling. Environ. Sci. Technol. 42, 4188 4194 (2008). 29. Connor, S. C. et al. Development of a multivariate statistical model to predict peroxisome proliferation in the rat, based on urinary 1H NMR spectral patterns. B iomarkers 9, 364 85 (2004). 30. Samuelsson, L. M., Frlin, L., Karlsson, G., Adolfsson Erici, M. & Larsson, D. G. J. Using NMR metabolomics to identify responses of an environmental estrogen in blood plasma of fish. Aquat. Toxicol. 78, 341 349 (2006).

PAGE 38

29 31. Ekman, D. R. et al. Metabolite profiling and a transcriptional activation assay provide direct evidence of androgen receptor antagonism by bisphenol A in fish. Environ. Sci. Technol. 46, 9673 9680 (2012). 32. Lankadurai, B. P., Wolfe, D. M., S impson, A. J. & Simpson, M. J. 1H NMR based metabolomic observation of a two phased toxic mode of action in Eisenia fetida after sub lethal phenanthrene exposure. Environ. Chem. 8, 105 114 (2011). 33. Garcia Reyero, N. et al. Integrated approach to explore the mechanisms of aromatase inhibition and recovery in fathead minnows (Pimephales promelas). Gen. Comp. Endocrinol. 203, 193 202 (2014). 34. Katsiadaki, I. et al. Hepatic transcriptomic and metabolomic responses in the Stickleback (Gasterosteus aculeatus ) exposed to ethinyl estradiol. Aquat. Toxicol. 97, 174 187 (2010). 35. Keun, H. & Ebbels, T. Geometric trajectory analysis of metabolic responses to toxicity can define treatment specific profiles. 17, 579 587 (2004). 36. Jensen, K. M., Korte J. J., Kahl, M. D., Pasha, M. S. & Ankley, G. T. Aspects of basic reproductive biology and endocrinology in the fathead minnow (Pimephales promelas). Comp. Biochem. Physiol. Part C Toxicol. Pharmacol. 128, 127 141 (2001). 37. Ankley, G. T. & Villeneuve, D. L. The fathead minnow in aquatic toxicology: Past, present and future. Aquat. Toxicol. 78, 91 102 (2006). 38. Sediment and Biosolids by HPLC / MS / MS. EPA Method 77 (20 07). 39. Schultz, M. M. et al. Antidepressant pharmaceuticals in two U.S. effluent impacted streams: Occurrence and fate in water and sediment and selective uptake in fish neural tissue. Environ. Sci. Technol. 44, 1918 1925 (2010). 40. Van Den Belt, K., Be rckmans, P., Vangenechten, C., Verheyen, R. & Witters, H. Comparative study on the in vitro/in vivo estrogenic potencies of 17 estradiol, estrone, ethynylestradiol and nonylphenol. Aquat. Toxicol. 66, 183 195 (2004). 41. Schwindt, A. The population ecology of fathead minnows (Pimephales promelas) in estrogen contaminated environments. (2013). 42. Finne, E. F., Cooper, G. A., Koop, B. F., Hylland, K. & Tollefsen, K. E. Toxicogenomic responses in rainbow trout (Oncorhynchus mykiss) hepatocytes exposed to model chemicals and a synthetic mixture. Aquat. Toxicol. 81, 293 303 (2007). 43. Cheek, A. O. et al. Experimental evaluation of vitellogenin as a predictive biomarker for reproductive disruption. Environ. Health Perspect. 109, 681 690 (2001).