Contaminants of Emerging Concern in

the Great Lakes

Science to Inform Management Practices for
Protecting the Health and Integrity of Wildlife
Populations from Adverse Effects

GLRI Action Plan I, Focus Area 1, Goal 5
by

Daniel L. Villeneuve1, Steven R. Corsi2, Christine M. Custer3, W. Edward Johnson4,
Stephanie L. Hummel5, Heiko L. Schoenfuss6, Edward J. Perkins7, Sarah A. Zack8

U.S. EPA, Mid-Continent Ecology Division, Duluth, MN, USA

U.S. Geological Survey, Upper Midwest Water Science Center, Middleton, Wl, USA

U.S. Geological Survey, Upper Midwest Environmental Sciences Center, La Crosse,
Wl, USA

NOAA National Centers for Coastal Ocean Science, Silver Spring, MD, USA

U.S. Fish and Wildlife Service, Ecological Services, Region 3, Bloomington, MN, USA

St. Cloud State University, St. Cloud, MN, USA

U.S. Army Engineer Research and Development Center, Vicksburg, MS, USA

Illinois-Indiana Sea Grant College Program, University of Illinois Extension,
Woodstock, IL, USA.


-------
Dedication

This report is dedicated to Tom Custer, in
recognition of his 44 years of federal service, his

impact and productivity as a scientist, and the
inspiring energy and enthusiasm that he brought to
all his work to protect Great Lakes ecosystems.

ii


-------
Table of Contents

Executive Summary	1

Introduction	4

Integrated Results and Discussion	8

CEC Occurrence	8

Priority CECs	10

Screening Values and Benchmarks	10

Exposure Activity Ratios (EARs)	14

Weight of Evidence	14

CECs and Other Stressors	15

Weight of Evidence Evaluations	15

PAHs as a Threat to Great Lakes Fish and Wildlife	15

Estrogens as a Threat to Great Lakes Fish and Wildlife	17

Overarching Management Implications	21

CECs are Widespread	21

Effects are Often Subtle and Implications for Ecological Fitness Unclear	21

Identifying and Reducing Sources of PAHs Could Reduce Contaminant-Related
Stress in a Number of Great Lakes Tributaries	22

Complementary Methods Show Promise for CECs Surveillance and Monitoring .... 22

Implementation of Management Practices is Influenced by Contaminant Source ... 22

Data Availability	24

References	25

Appendix A. Organic Contaminants, Microplastics, Waterborne Pathogens, and Host-
Associated Bacteria Surveillance and Potential Biological Effects in Great Lakes
Tributaries	29

Introduction	29

Organic Contaminants	30

Objectives	30

Methods	30

Key Findings	32

Microplastics	35

Objectives	36

Methods	36

Key Findings	38

Microorganisms	41

Objectives	41

iii


-------
Methods	42

Key Findings	43

Knowledge Gaps	48

Management Implications	49

Acknowledgements	51

Products	51

References	52

Appendix B. Monitoring of Contaminants of Emerging Concern by Great Lakes Mussel
Watch	57

Introduction	57

Methods	58

Results and Discussion	60

Polycyclic Aromatic Hydrocarbons (PAHs)	60

Polybrominated Diphenyl Ethers (PBDEs)	62

Pharmaceuticals and Personal Care Products (PPCPs)	65

Alkylphenols and Alkylphenol Ethoxylates	67

Bivalve Health	69

DNA Damage	69

Metabolomics	71

Cellular Biomarkers	72

Key Findings	74

Management Implications	74

Knowledge Gaps	75

Acknowledgements	75

References	75

Appendix C. Exposure and Effects of Bioaccumulative Contaminants of Emerging
Concern in Tree Swallows Nesting across the Laurentian Great Lakes	77

Problem Statement and Study Overview	77

Introduction	78

Methodology	78

Key Findings	81

Polycyclic Aromatic Hydrocarbons (PAHs)	81

Polybrominated Diphenyl Ethers (PBDEs)	85

Per- and polyfluorinated Alkyl Substances (PFAS)	87

Management Implications	90

Knowledge Gaps	90

Disclaimer	90

iv


-------
Acknowledgements	91

References	91

Appendix D. Survey of Contaminants of Emerging Concern and Their Effects to Fish
and Wildlife in Great Lakes Tributaries	93

Problem Statement and Scope	93

Methodology	95

Key Findings	96

CECs Are Ubiquitous in Great Lakes Waterbodies	96

Biological Effects Are Subtle and Widespread	98

Effects of CECs Differ between Species	100

Some CECs Occur at Concentrations Predicted to Cause Biological Effects... 101

Management Implications	103

Knowledge Gaps	104

Disclaimer	105

Acknowledgements	105

References	105

Appendix E. Detecting and Evaluating Biological Effects of Contaminants of Emerging
Concern in Great Lakes tributaries	109

Problem Statement and Scope	109

Detecting Biological Activities Associated with Mixtures of CECs in Great Lakes
Tributaries	110

Strategy - In Vitro Assays	112

Strategy - In Situ Exposure of Model Organisms	113

Endpoints and Analyses	116

Key Findings/Progress - Detecting Biological Activities	117

Associating Chemicals with Specific Biological Activities	123

Sample Collection Approaches that Minimize Variables that Can Confound the
Association of Chemical Stressors and Biological Effects	123

Multi-Variate Statistical Approaches for Associating Chemicals with Biological
Effects	124

Evidence-Based Approaches for Associating Chemicals with Biological Effects
	125

Chemical-Pathway Interaction Networks	125

Exposure:Activity Ratios (EARs)	126

Prospective Application of Evidence-Based Approaches	127

Key Findings/Progress - Associating Detected Chemicals with Biological Effects
	128

Linking Biological Activities to Adverse Apical Effects in Wildlife	129

V


-------
The Adverse Outcome Pathway (AOP) Framework and Effects-Based Monitoring
	129

Coupling Population Models with Pathway-Based Data	129

Extrapolating Pathway-Based Data across Species	130

Key Findings/Progress - Development of Translational Tools and Frameworks
for Linking Pathway-Based Measurements of Adverse Effects	131

Application Case Studies	132

Integrated Application of Pathway-Based Approaches for Source Attribution ..132

A Site Assessment Workflow Moving from Pathway-Based Monitoring to AOP-
Informed Hypothesis Testing and Hazard Verification	132

Effects-Based Monitoring Leading to Reassessment of the Significance of a
Common Environmental Contaminant	133

Pathway-Based Surveillance to Prioritize Future Monitoring	133

Management Implications and Future Directions	134

Future Development	133

Disclaimer	136

Acknowledgements	136

Action Plan I Research Products Cited	137

Other References	139

Appendix F. Transcriptional Effects-Based Monitoring of Contaminants of Emerging
Concern in Great Lakes Tributaries: Biological Effects and Chemical Specific
Impacts	141

Problem Statement and Scope	141

Detecting Biological Activities Associated with Mixtures of CECs in Great Lakes
Tributaries using Transcriptomics	141

Strategy - In Situ Exposure of Model Organisms	141

Strategy - Laboratory Exposure of Model Organisms	142

Endpoints and Analyses	142

Case Study: Effects-Based Monitoring of Wastewater Treatment Plant Effluent at
Increasing Distances from Point of Discharge	145

Key Findings/Progress - Detecting Biological Activities with Transcriptomics . 145

Associating Detected Chemicals with Biological Effects Using Transcriptomics and
Evidence-Based Approaches	145

Statistical Approaches for Associating Chemicals with Transcriptomics and
Biological Effects	145

Evidence-Based Approaches for Associating Chemicals with Biological Effects
	146

Single Chemical Gene Expression Effects	146

Causal Networks Linking Gene Expression and Chemicals	146

vi


-------
Hazard Quotients	147

Case Study: Transcriptomics of Bisphenol A Laboratory Exposures to Assess
Bisphenol A Contribution to Estrogenicity in Wastewater Treatment Plant Effluent
	147

Case Study: Integrated Application of Transcriptomics and Evidence-Based
Approaches Assessing the Estrogenic Impact of CECs	147

Case Study: Determining CECs Associated with Transcriptomics Effects in Great
Lakes Tributaries	150

Key Findings/Progress - Associating Detected Chemicals with Biological Effects
Using Transcriptomics and Evidence-Based Approaches	153

Linking Biological Activities to Adverse Apical Effects in Wildlife	153

The Adverse Outcome Pathway (AOP) Framework with Transcriptomics Effects-
Based Monitoring	153

Coupling Organ Damage Models with Transcriptomics Data	154

Case Study: Transcriptional Effect-Based Monitoring of Potential for CEC
Mixtures to Cause Tumor Development	156

Key Findings/Progress - Development of Translational Tools and
Frameworks for Linking transcriptional Measurements to Adverse Effects in
Wildlife	156

Management Implications and Future Directions	157

Action Plan I Products	158

Other References	159

ADDENDUM 1 - Contaminants of Emerging Concern in the Great Lakes: GLRI
Integrated Phase II Group Progress Report. GLRI CEC 2016. 22 pages.

vii


-------
Executive Summary

Under Action Plan I (2010-2014) of the Great Lakes Restoration Initiative (GLRI),
Federal and Academic partners began an investigation of the presence and distribution
of contaminants of emerging concern (CECs) in the Great Lakes and potential impacts
on fish and wildlife. The term CECs is applied to a broad range of chemicals that are
currently in use but for which we currently lack good understanding of whether fish,
wildlife, or humans are being exposed and/or whether negative health or environmental
effects are expected if exposure occurs. Pharmaceuticals, personal care products,
flame retardants, many current use pesticides, and poly- and perfluorinated chemicals
are some well-known groups of CECs, but there is no definitive or comprehensive list
that can be used to support the management of CECs to reduce impacts on the Great
Lakes ecosystem.

Four overarching goals were identified for this collaborative investigation:

1.	Evaluate the sources, occurrence, and distribution of CECs across the Great
Lakes Basin.

2.	Examine associations between the distribution of CECs and land-use patterns.

3.	Review both scientific literature and field-generated data to determine the
potential for CECs to cause adverse effects on Great Lakes fish and wildlife
populations.

4.	Develop efficient strategies to survey and/or monitor for threats that CECs may
pose in order to take effective management actions before those threats evolve
into large scale impacts on Great Lakes ecosystems or the services they provide.

Achievement of these goals ensures progress towards Focus Area 1: Toxic Substances
and Areas of Concern from GLRI Action Plan I, Goal 5: "The health and integrity of
wildlife populations and habitat are protected from adverse chemical and biological
effects associated with the presence of toxic substances in the Great Lakes Basin".

This large-scale research effort was comprised of individual and collaborative projects
from multiple federal agencies and academic institutions, involving over 85
investigators, and overseen by the U.S. Environmental Protection Agency (EPA) Region
5, Great Lakes National Program Office. Partners include the United States Geological
Survey, the National Oceanic and Atmospheric Administration, U.S. Fish and Wildlife
Service, Saint Cloud State University, the U.S. EPA Office of Research and
Development, and the U.S. Army Corps of Engineers.

Key findings:

1. Contaminants of emerging concern were found throughout the monitored
Great Lakes tributaries, but types and concentrations vary in association with
regional land use. CECs were detected in nearly all samples collected. The type
and concentration of the specific contaminants detected varied considerably among
field sites and in association with land use type, such as urban, agricultural, wetland,

l


-------
or forest. Contaminants were detected in the water column, sediment, and tissues of
all species surveyed in the current work (mussels, aquatic insects, fish, and insect-
eating birds).

2.	There were over 20 contaminants for which CEC concentrations approached
or exceeded those reported to cause toxicity in laboratory experiments. This
was based on detection in water, sediments and or biota at one or more field sites.
These contaminants represent compounds that warrant further investigation and
monitoring with respect to potential impacts in certain areas of the Great Lakes
basin. Based on the present investigation, compounds of greatest concern include:
polycyclic aromatic hydrocarbons, associated with oil-based products and
combustion of organic matter; atrazine, an herbicide; dichlorvos, an insecticide; and
ibuprofen and venlafaxine, both pharmaceuticals.

3.	Results suggest that mixtures of CECs presently found in most Great Lakes
tributary locations surveyed may elicit subtle biological effects, but likely are
not, alone, causing obvious detriment to current communities of fish and
wildlife. CECs detected in the Great Lakes were associated with subtle biological
effects like changes in gene expression, altered circulating glucose, etc. in both wild-
caught and laboratory-reared organisms. These effects were generally not indicative
of reproductive failure or mortality. However, the effects may have more serious
implications when combined with other sources of stress like habitat degradation,
changing climate conditions, and competition with invasive species. Due to limited
historical data, it is unknown whether severe CEC-related impacts have already
affected aquatic communities in waterbodies that have received long-term inputs of
these contaminants. Likewise, under Action Plan I, biological effects were not
necessarily evaluated at the sites where CEC concentrations exceeding laboratory
toxicity thresholds were detected. As a result, strategic, ongoing surveillance and
monitoring of CECs is warranted.

This collaborative investigation resulted in new tools, approaches, and data that can be
used to inform and support the management of CECs to reduce their impacts on Great
Lakes natural resources. The following products of this research effort are available
through https://commynities.geoplatform.gov/glri/ or by contacting the investigators (see
technical chapters found in Appendices A-F):

1.	Database of CEC occurrence and concentrations in US tributary streams. The

database includes CEC detections in water, sediment, and fish and wildlife tissues,
and represents the most comprehensive survey of CECs in the Great Lakes Region.

2.	Synopses of results and key findings. Integrated summaries of results,
conclusions, and management implications of the CEC research are available
through reports, topical fact sheets, and presentations.

3.	Technical publications: This collaborative research effort has resulted in over 50
peer-reviewed publications, agency reports, and data releases that can be of use to
resource managers, the scientific community, and members of the public.

4.	Innovative tools. Innovative monitoring devices, sampling equipment, conceptual
frameworks, and software applications were developed over the course of this

2


-------
research. These tools are transferable to stakeholders via internet accessibility or via
specifications, instructions, and demonstration detailed in technical publications.

Hypotheses to guide CECs research under Action Plan II. Findings from 2010-2014
were used to guide further research in 2015-2018 for basin-wide surveillance of CECs
and for sites warranting further study of potential biological impacts of CECs. Additional
surveillance included both evaluation of additional classes of contaminants and
expanded lists for chemical classes shown to be of greatest concern. Mixtures of some
of the most frequently detected contaminants were also tested in laboratory studies to
understand whether long term exposures to multiple contaminants may result in effects
not evident from uncontrolled, short-term field experiments.

3


-------
Introduction

The Great Lakes Restoration Initiative (GLRI) accelerates efforts to protect and restore
the Great Lakes, the largest system of fresh surface water in the world. Built upon the
foundation of the Great Lakes Regional Collaboration Strategy, GLRI answered a
challenge from the governors of the Great Lakes states to commit to creating a new
standard of care that will leave the Great Lakes better for the next generation. Since
2010, the multi-agency GLRI has provided funding to strategically target the greatest
threats to the Great Lakes ecosystem and to accelerate progress toward achieving long
term goals, including protecting the health and integrity of wildlife populations and
habitat from adverse chemical and biological effects associated with the presence of
toxic substances.

Under GLRI Action Plan I, two broad classes of toxic substances were identified:

1.	Legacy contaminants -identified as persistent, bioaccumulative, and toxic
pollutants largely left over from past practices, but that continue to recirculate
through the ecosystem (GLRI 2010). Examples of legacy contaminants include
polychlorinated biphenyls, mercury, polychlorinated dioxins and furans, and
several highly persistent pesticides (e.g., DDT, aldrin, dieldrin, chlordane, etc.).
These chemicals are known to be persistent, bioaccumulative, and toxic.

2.	Contaminants of emerging concern (CECs) - identified as chemicals detected in
the Great Lakes over the past several years that may pose threats to the
ecosystem (GLRI 2010). As it is used here, the term CECs applies to a broad
diversity of chemical structures used in a range of commercial products including
flame retardants (e.g., polybrominated diphenyl ethers; PBDEs), household
chemicals, pharmaceuticals and personal care products (PPCPs), current use
pest control agents, as well as a variety of industrial chemicals. Broadly
speaking, some unifying attributes of CECs include a lack of regulatory
standards, recent discovery/detection of occurrence in the environment, and
limited information regarding effects on aquatic life at environmentally relevant
concentrations. A few CECs, such as bisphenol A, 4-nonylphenol, and some per-
and polyfluorinated alkyl substances (PFAS), have received considerable public
attention and a greater level of study. Additionally, compounds including PAHs
and widely-used agricultural chemicals, including atrazine, have also been well-
studied and have existing regulatory standards, but are included as CECs for the
purposes of this report due to their on-going input into aquatic systems.

To better understand the potential threats CECs may pose to the health and integrity of
Great Lakes wildlife populations, several specific research objectives were defined:

1. Evaluate the sources, occurrence, and distribution of CECs across the
Great Lakes Basin. This effort involved using analytical chemistry methods to
quantify concentrations of selected CECs in water, sediment, and wildlife tissues
that were collected from sites across the Great Lakes basin. Emphasis was

4


-------
placed on tributary streams rather than open lake or nearshore habitats, as
tributaries provide pathways for CECs to enter the Great Lakes.

2.	Examine associations between the distribution of CECs and land-use
patterns. Land use in the Great Lakes Basin varies, and a range of
environments (e.g., urban, agricultural, wetland) can be found. Evaluating CECs
in relationship to land use helps to identify predictive approaches for identifying
locations around the Great Lakes region that might be more, or less, likely to
have certain types and concentrations of CECs. The ability to associate CECs
with specific land use types may also inform development of CEC source
reduction strategies.

3.	Review both existing data and research and novel field-generated data to
determine the potential for adverse effects from CECs on Great Lakes fish
and wildlife populations. Existing toxicity information was used in conjunction
with new sources of data or models in order to better understand potential
differences in sensitivity among species.

4.	Develop efficient strategies to survey and/or monitor environmental threats
from CECs, thereby allowing management actions to be implemented
before threats evolve into large-scale impacts on Great Lakes ecosystems
or the services they provide. Effects-based monitoring approaches can be
used in order to evaluate potential threats from CECs. Effects-based monitoring
refers to use of biological responses observed in cell cultures in water samples,
laboratory-reared organisms, and/or wild-caught organisms as an indication of
exposure to contaminants and/or potential harm to wildlife resources and the
ecosystem functions they provide.

To address these objectives, a multi-partner research effort was developed and
implemented. The project team included several federal agency and academic partners
and was facilitated by the U.S. Environmental Protection Agency (EPA) Great Lakes
National Program Office (GLNPO). Collectively, these efforts employed a range of
complementary approaches that addressed the objectives and the overall goal of
understanding the potential threats CECs may pose to the health and integrity of Great
Lakes wildlife populations.

•	The U.S. Geological Survey (USGS) Great Lakes Area Water Science Centers
conducted a basin-wide survey of organic contaminants, microplastics, and
pathogens and evaluated their association with land-use attributes. Both water and
sediment were sampled.

•	The USGS Upper Midwest Environmental Sciences Center expanded their historical
monitoring of contaminant exposure and biological effects in tree swallows nesting at
Great Lakes Areas of Concern (AOCs) to include an expanded range of CECs and a
broader diversity of biological measurements. Tree swallow monitoring is well-suited
to provide insight into potential transfer of CECs from the aquatic food base to
terrestrial nesting birds.

5


-------
•	The National Oceanic and Atmospheric Administration Mussel Watch Program
expanded their historical monitoring of persistent toxic substances in the Great
Lakes basin to include monitoring of CECs and added several biological
measurements. The sedentary lifestyle of dreissenid mussels provides for precise
place-based monitoring and an examination of the effects of CEC exposure on biota.
Although dreissenid mussels are non-native invasive species, they were used
because of the extensive monitoring data available, abundant availability throughout
the Great Lakes' Basin, lack of concern for removing individuals from wild
populations for research use, and the possibility that they could be used as
indicators for determining site-specific exposure.

•	The U.S. Fish and Wildlife Service CEC Team and St. Cloud State University
provided more in-depth local characterization of the concentrations and distribution
of CECs within selected Great Lakes tributary streams. They coupled chemical
monitoring in water, sediment, and resident fish tissues with biological analyses of
both wild-caught and laboratory-reared caged fish.

•	The U.S. EPA's Mid-Continent Ecology Division used cell culture-based assays to
screen water samples for biological activities potentially associated with adverse
health effects in fish and other wildlife. Additionally, laboratory-reared fish were
caged at selected field locations and a variety of biological analyses were conducted
to evaluate early warning signs indicative of exposure to CECs.

•	The U.S. EPA's Exposure Methods and Measurements Division used advanced
analytical instrumentation including nuclear magnetic resonance spectroscopy and
high-resolution mass spectrometry to measure changes in the internal biochemistry
of organisms following exposure to water from Great Lakes tributaries. These
measurements provide insights into how the organism may be responding to stress
or using energy following exposure to CECs and can provide an early indicator of
damage to internal organs and tissues that would not be evident upon visual
examination.

•	The U.S. Army Corps of Engineers Research and Development Center employed
cutting-edge biotechnologies to examine the effects of CECs occurrence and
concentrations in Great Lakes water on molecular responses in exposed fish.

At the onset of Action Plan I, the agencies planned and executed CEC studies
independently. Consequently, the technical appendices (Appendices A-F) were
organized by partner organization. As work progressed, US EPA GLNPO Focus Area
leads/sub-leads coordinated biannual meetings that provided a platform for partners to
share their approaches and results and identify opportunities for collaboration and
coordination. By the end of Action Plan I, the partners were organized into an
interagency CECs workgroup. Overarching results of the entire research effort and
resulting management implications are presented in an integrated fashion in Chapters 2
and 3, respectively. These integrated results served as the foundation for a more
coordinated and targeted research effort carried out under Action Plan II. Additional
details of the research are presented in over 50 peer-reviewed journal articles and

6


-------
technical reports that resulted from the interagency CEC workgroups efforts under
Action Plan I.

7


-------
Integrated Results and Discussion

Partners employed a diverse range of approaches to evaluate presence and variability
of CECs across a representative range of Great Lakes tributary streams, and the
potential biological effects of the contaminant mixtures present at selected sites. Water
and sediment samples from Great Lakes tributaries were collected and analyzed for a
broad suite of CECs to provide information on occurrence and distribution (Appendices
A, D). Wildlife tissues were examined for presence of CECs, including those from
dreissenid mussels (Appendix B), egg, plasma, and other tissues from tree swallows
(Appendix C), and fish tissues (Appendix D). Biological effects along the continuum
from molecular and biochemical to physiological to individual level outcomes (e.g.,
impacts on survival, growth, and/or reproduction) were also evaluated on dreissenid
mussels, fish, and birds (Appendices B-E). Collectively, these studies have produced
complementary information and a multi-faceted data matrix on CECs that provides
evidence for initial screening and prioritization of CECs that are of biological concern in
Great Lakes tributaries.

CEC Occurrence

Contaminants of emerging concern are ubiquitous in monitored Great Lakes tributaries
and vary according to association with land use. Analysis of water and sediment
samples indicated that CECs were present in nearly all tributaries sampled, but the
actual contaminants detected varied substantially by site (Baldwin et al. 2016; Elliot et
al. 2017; Dila et al. 2018). The contaminants detected most commonly in water samples
(Table 1) included PAHs, herbicides and insecticides, contaminants commonly
associated with wastewater (e.g., alkylphenols, dyes and pigments, fecal indicators,
flavors and fragrances, PPCPs) and flame retardants (PBDEs). Contaminants were
generally most prevalent in urban areas, followed by agricultural areas. Pesticides were
evenly associated with agricultural and urban land use, although with herbicides more
dominant in agricultural area and insecticides more prevalent in urban watersheds.
Watersheds dominated by forest and wetlands had the lowest concentrations of CECs.

The most prevalent contaminants (> 30% occurrence) detected in sediment samples
included PAHs, industrial chemicals (e.g., 9,10-anthraquinone), wastewater
contaminants (e.g., alkylphenols, fecal indicators, hormones, PPCPs, solvents)
components of plastic, and phenolic chemicals (which have natural and anthropogenic
sources; Elliot et al. 2017, Table 1). PAHs were the most common class of
contaminants and were observed at the greatest concentrations among all the
contaminants detected in sediment.

Fish liver tissues were sampled for CECs at a subset of locations studied (Appendix D).
Plasticizers and flame retardants were detected in all fish liver samples, while all other
chemical classes had at least one sample with non-detects observed. Atrazine, 17-
alpha-ethynylestradiol,17-alpha-estradiol, androstenedione, carbamazepine, diazepam,
diclofenac, estrone, meprobamate, naproxen, sulfamethoxazole, and estriol were not
detected in any benthic or pelagic fish samples (Choy et al. 2017), even though they

8


-------
were commonly detected in water and sediment samples (Table 1). The most
commonly detected chemical class in fish liver tissue was PFAS while PBDEs were
detected in just a few samples.

Table 1. Contaminant of emerging concern use categories sampled (X) for different
environmental sample types, 2010-2014. NS = Not sampled. The analytes evaluated in tissues
were more limited.







Sample Type





Water

Sediment

Mussel
Tissue

Fish
Tissue

Tree
Swallow
Tissues

Alkyphenols ws

X

X

X

NS

NS

Dyes and pigments w

X

NS

NS

NS

NS

Fecal indicators ws

X

X

NS

NS

NS

Flame retardants1 w

X

NS

NS

NS

NS

Flavors and fragrances w

X

X

NS

NS

NS

Herbicides w

X

X

NS

NS

NS

Hormones s

X

X

NS

X

NS

Industrial chemicalss

X

X

NS

NS

NS

Insecticides w

X

X

NS

NS

NS

Microplastics

X

NS

NS

NS

NS

PAHs ws

X

X

X

NS

X

PBDEs

NS

NS

X

X

X

PFAS

NS

NS

NS

NS

X

Phenolic chemicals s

X

X

NS

NS

NS

Plastics components s

X

X

NS

X

NS

PPCPs ws

X

X

X

X

NS

Solvents ws

X

X

NS

NS

NS

1 Refers to flame retardants other than PBDEs, which are identified separately.

w Contaminants commonly detected in water samples.
s Contaminants commonly detected in sediment samples.

Common contaminants reported in mussel tissue (Appendix B) included PAHs, PBDEs,
PPCPs, alkylphenols and alkylphenol ethoxylates (Table 1). Results generally indicated
that compounds were more concentrated in dreissenid mussels from harbor and
tributary samples than nearshore or offshore samples; however harbor and tributary
sites were limited in scope to the Milwaukee and Niagara River areas. Of the CECs
detected in mussel tissues, PAHs had the greatest concentrations.

The inclusion of tree swallows in the study provided insights into the potential transfer of
CECs from the aquatic to terrestrial environment via food web interactions. Among the
suite of contaminants monitored in tree swallow samples and nestling diets (prey items
as well as gut contents), compounds detected included PAHs, PBDEs, and PFAS

9


-------
(Table 1, Appendix C). The sites with the greatest PAH concentrations were near
industrial urban areas (Appendix C; Custer et al. 2017a). PBDEs and PFAS were
detected in all samples, but at relatively low concentrations at most sites. The greatest
concentrations of PBDEs in tree swallows were detected at industrial sites (Appendix
C), and the greatest concentrations of PFAS were near airfields where aqueous film
forming foam was used in fire-suppression training exercises (Custer et al. 2019; Custer
et al. 2017; Nakayama et al. 2010).

Overall, monitored CECs varied across water, sediment, and wildlife tissues, with a
common conclusion that PAHs were a primary concern due to their prevalence in many
watersheds across the Basin. Water sample results indicated that wastewater-related
chemicals and pesticides were prevalent at many sites and are important to consider
when evaluating potential biological effects. Flame retardants such as PBDEs and
phosphate-based chemicals were commonly detected, but at relatively low
concentrations relative to reported biological effect concentrations.

Priority CECs

Data describing the presence of hundreds of contaminants of emerging concern in
tributaries of the Great Lakes are becoming more available. However, understanding
which of these contaminants may elicit negative consequences and at which locations is
challenging. Nonetheless, identifying the CECs likely to be causing negative impacts is
important in order to focus management and monitoring resources in a way that most
effectively mitigates the threats that CECs may pose to Great Lakes fish and wildlife.
The CECs for which there was compelling evidence associating environmental
concentrations with the potential for negative biological effects include:

•	Select PAHs, products of combustion

•	Atrazine, an herbicide

•	Dichlorvos, an insecticide

•	Ibuprofen and Venlafaxine, pharmaceuticals

Identification of priority CECs was based on multiple approaches. Each approach
utilized different types of input data and yielded different levels of uncertainty regarding
the potential to detect negative effects. Consequently, the overall weight of evidence,
considering lines of evidence from multiple approaches and sources, were used
wherever possible to inform prioritization

Screening Values (SV) and Benchmarks

In cases where toxicity information was available either from regulatory guideline
studies or from the peer-reviewed scientific literature, concentrations of CECs detected
in this study were compared to those previously demonstrated to adversely affect
organisms (Appendices A, D). This was done by dividing concentration of the chemical
required to produce a negative biological response by the concentration measured in
the environment (i.e., calculating a toxicity quotient). Values greater than one indicate a
strong likelihood of effect. Two complementary methods were used; either comparing

10


-------
water chemistry results to established water quality benchmarks (Appendix A; Baldwin
et al. 2016) or comparing water chemistry results to custom developed water quality
screening values (Appendix D; Gefell et al. 2019a). Established water quality
benchmarks were compiled from agencies such as U.S. Environmental Protection
Agency, Canadian Council of Ministers of the Environment, and Environment and
Climate Change Canada (Baldwin et al. 2016) and were available for 27 CECs (Table
2). Screening values were developed for an additional 14 CECs that were detected
frequently in water samples but for which benchmark standards did not already exist
(Gefell et al. 2019a). Screening values were based on concentrations reported in the
scientific literature to cause negative impacts on survival, growth, reproduction, or
selected developmental processes or behaviors in fish that could be readily linked to
population demographics (Gefell et al. 2019b). Screening values were developed for an
additional seven effect categories related to fish health (endocrine, neurological,
physiological/metabolic, etc.). Two types of screening values were derived; "SVlow" as a
concentration below which no hazard to fish is expected and "SVhigh" as a
concentration above which there is a strong expectation of adverse effects. Collectively,
toxicity screening-based methods were applied for 41 unique CECs (Table 2).

One or more established water quality benchmarks were exceeded in samples from 20
of the 57 sites evaluated (Appendix A). In total, 11 unique chemicals including five
PAHs, five pesticides, and one detergent metabolite (4-nonylphenol; Table 2; Baldwin et
al. 2016) exceeded established water quality benchmarks at one or more sites.
Chemicals with the most frequent exceedances were PAHs, 4-nonylphenol, and
atrazine. Up to nine different compounds exceeded established water quality
benchmarks at a single site.

Based on custom screening values (Appendix D; Gefell et al. 2019b), measured
concentrations exceeded SVlow values for all but one of the 14 CECs evaluated
(lidocaine). Exceedance of SVhigh values in population-relevant effect categories (e.g.,
survival, growth, reproduction) was detected for two of the 14 chemicals, both of which
were pharmaceutical compounds (ibuprofen and venlafaxine). Each of the 24 project
locations evaluated using SVs had at least one exceedance of a SVhigh and 17 of 24
locations had at least one SVhigh exceedance. Variability in the exceedance of SVhigh
was observed between different tributaries and exceedance of screening values was
not detected at all sites on the same tributary, suggesting spatial heterogeneity. As one
might expect, within a given tributary system, exceedances were typically downstream
of point sources.

Both methods indicated the potential for adverse biological effects at multiple sites, and
both methods highlight the need to evaluate biological effects of co-occurring chemicals.
Current methods using screening values or benchmarks only assess the risks of each
individual chemical and do not address the cumulative risk associated with exposure to
co-occurring mixtures of CECs. Consequently, complementary methods that can help to
better understand the risk and effects of chemical mixtures on fish and wildlife are
needed in addition to toxicity quotient-based approaches.

li


-------
Use of established water quality benchmarks is of value for evaluating concentrations in
a biological context, but these benchmarks are available for only fraction of the CECs
detected in these studies. For example, of the organic waste compounds monitored by
USGS water quality benchmarks were available for less than half (27/69). US Fish and
Wildlife service derived SVs by drawing on a broader range of laboratory studies
published in the peer-reviewed scientific literature, not just standardized regulatory
toxicity tests. However, for the 25 most commonly detected chemicals, out of over 150
monitored, SVs could be developed for just 14 of the 25. An on-going and recognized
challenge in assessing the biological significance of CECs is the lack of available
toxicity information for many compounds. Because traditional laboratory-based toxicity
testing data are lacking for many CECs, alternative approaches for associating CECs
with biological effects that are less reliant on traditional toxicity data were also
evaluated.

12


-------
Table 2. Chemicals for which water chemistry data was evaluated using established water quality
benchmarks (BM) or custom screening values (SV). "No" indicates a water quality benchmark or
screening value was available, but not exceeded in any samples. "Yes" indicates the water
quality benchmark or custom screening value SVlow was exceeded in at least one sample.
"**YES**" indicates the water quality benchmark was exceeded by a factor of 10 or more or
custom screening value SVhigh was exceeded in at least one sample.

Chemical Class

Chemical Name

SVa or BMb
Exceeded?

Antimicrobial

Phenol

No

Disinfectants

Triclosan

Yes a

Detergent Metabolites

4-Nonylphenol

Yes b

Fire Retardants

Tris(2-butoxyethyl) phosphate (TBEP)

Yes a

Flavors and Fragrances

Hexahydrohexamethyl cyclopentabenzopyran (HHCB)

Yes a

Fuels

1 -Methylnaphthalene
2-Methylnaphthalene
Isopropylbenzene (cumene)

No
No
No



Atrazine

**YES** k



Bromacil

No

Herbicides

Metalaxyl
Metolachlor

No
Yes b



Pentachlorophenol
Prometon

Yes b
No



Androstene-dione

Yes a

Hormone

B-sitosterol

Yes a



Estrone

Yes a

Insecticides

Carbaryl

Chlorpyrifos

Diazinon

Yes b
No
No



Dichlorvos

**YES** k



N, N-diethyl-meta-toluamide (DEET)

Yes a

Other

p-Dichlorobenzene

No

Tribromomethane (bromoform)

No



Anthracene

**YES** k

PAHs

Benzo[a]pyrene
Fluoranthene

**YES** k
**YES** ^

Naphthalene
Phenanthrene

No
Yes b



Pyrene

**YES** ^

Plastics component

Bis(2-ethylhexyl) phthalate (DEHP)
Bisphenol A
Diethyl phthalate

No
Yes a
No

PPCP

Carbamazepine

Citalopram

Diphenhydramine

Ibuprofen

Lidocaine

Yes a
Yes a
Yes a
**YES** 9
No



Venlafaxine

**YES** 9

Solvents

Isophorone
Tetrachloroethene

No
No

a Peer-reviewed scientific literature-derived screening value (Gefell et al. 2019b).
b Established water quality benchmark (Baldwin et al. 2016).

13


-------
Exposure Activity Ratios (EARs)

One of the most promising approaches for associating CECs with their potential
biological effects involved comparing the CEC concentrations measured in
environmental samples with concentrations at which the chemicals elicit responses in
ToxCast high throughput screening assays (Kavlock et al. 2012; Blackwell et al. 2017).
ToxCast employs a battery of mostly cell culture-based assays that measure the ability
of chemicals to bind to and activate certain proteins, inhibit specific biochemical
reactions, alter gene expression, change cell shapes or functions, etc. (Kavlock et al.
2012). These biological activities do not directly equate to clear impacts on survival,
growth, or reproduction. However, they can provide a potentially conservative estimate
of the relative potency of a chemical to elicit a biological response, which may be
mechanistically linked to toxicity that can be expected to occur if the magnitude and
duration of exposure are adequate (Appendix E; Krewski et al. 2010).

The ratio of detected chemical concentration in the environment to the concentration at
which it elicits activity in a high throughput assay is termed an exposure-to-activity ratio
(EAR). While not as definitive as water quality benchmarks or SVs for identifying
ecological risks, the data needed to calculate an EAR are available for many more
chemicals than those for which benchmarks or SVs based on traditional toxicity testing
are. For example, of the 69 chemicals monitored by Baldwin et al. 2016 (Appendix A),
water quality benchmarks were available for just 27 (39%) chemicals, while EARs could
be calculated for 48 chemicals (70%). Additionally, even when some traditional toxicity
data are available, the EAR approach may identify subtle, sublethal impacts that
exposure to CECs may cause (e.g., endocrine disruption, potential for behavioral or
developmental effects, etc.) that may not be detected in traditional toxicity tests.
Furthermore, EARs can be summed for all chemicals detected in a sample that cause
activity in the same assay. This provides a means to evaluate cumulative impacts of
mixtures. Due to the potential utility of this method, further development and application
of the EAR approach was identified as an important goal/focus under Action Plan II.

Weight of Evidence

In some instances, statistical inference approaches were used to prioritize CECs that
may have a biological effect when traditional toxicity data or alternative data from
programs like ToxCast were not available. Statistical approaches (e.g., partial least
squares regression, context likelihood of relatedness) can examine potential co-
variation between the concentrations of a chemical detected at a given site and the
magnitude of effect on a broad range of biological effect endpoints. These endpoints
may be measured in organisms exposed on-site (Appendices B-F) or in laboratory-
based assays following exposure to extracts of environmental samples (Appendix E).
As part of Action Plan I, multi-variate statistical approaches were applied to identify co-
variation between chemical concentrations and the abundance of small molecule
metabolites (Appendix E) or messenger RNA (Appendix F) extracted from tissues or
fluids from fish exposed on site. Because these approaches identify correlations, rather
than causation, they are not definitive for hazard identification or risk assessment.
Nonetheless, they were shown to be useful for reducing a long list of detected CECs to

14


-------
a shorter list that could be prioritized for subsequent monitoring and testing (Appendix
E; Davis et al. 2016). Additionally, along with complementary approaches like water
quality benchmarks, screening values, or EARs, they can provide another line of
evidence for use in weight of evidence-based approaches for identifying which CECs
may threaten the integrity and health offish and wildlife populations.

CECs and other Stressors

Some CEC concentrations exceeded available water quality benchmarks and/or
screening values. Nonetheless, monitoring of biological responses in either laboratory-
reared fish caged in the field or resident fish and wildlife generally did not indicate overt
toxicity was occurring. Results to date indicate that individually, CECs are not likely to
cause serious declines to current fish and wildlife populations or impair ecosystem
function. However, CECs detected in the Great Lakes were associated with subtle
biological effects. These effects included changes in gene expression, increased
circulating glucose concentrations, altered fish health parameters such as organ
weights as a percent of body weight in both resident and laboratory-reared organisms.
These subtle effects may have more serious implications when combined with other
sources of stress like habitat degradation, changing climate conditions, and competition
with invasive species. Ideally, results from this CEC research should be integrated with
other monitoring activities conducted as part of the GLRI or on-going management
programs to consider a more holistic ecosystem context in which CECs represent one
of several stressors to Great Lakes fish and wildlife. It is also noted that biological
effects monitoring conducted as part of this effort was not always co-located at the sites
where CEC concentrations exceeding water quality benchmarks were detected. Thus,
overt impacts are still a possibility at select locations, and these are potential priorities
for follow-up investigations.

Weight of Evidence Evaluations

As noted in the introduction (Chapter 1), at the inception of this work on CECs
experimental designs, analytical methods, and site selection for different aspects of the
research were not coordinated among partners. Consequently, the full spectrum of
approaches employed in the research could not necessarily be applied at each site, for
each chemical of interest, and/or for all species evaluated. The following examples
highlight situations where the weight of evidence collected by the various partners
contributed to a broader understanding of key findings. The increased interagency
coordination that emerged over the course of GLRI Action Plan I and the associated
integration case studies set the stage for a more coordinated cross-agency effort under
GLRI Action Plan II

PAHs as a Threat to Great Lakes Fish and Wildlife

Polycyclic aromatic hydrocarbons were one of the classes of chemicals that were
evaluated by all partners involved in the CEC research. Often, PAHs are considered
legacy contaminants due to the availability of toxicity benchmarks and known inputs
from recirculation. However, PAHs were included in this CEC-focused effort due to their

15


-------
pervasiveness and in recognition of the diversity of on-going sources that continue to
introduce PAHs into the Great Lakes basin. Results detailed in the Appendices (A-F)
provide evidence that PAHs are likely a significant contributor to biological effects
detected in several areas and species studied.

Measurable concentrations of PAHs were found in the surface water of 57 different
Great Lakes tributaries and at 43% of all sites surveyed (Appendix A). At 20 of these 57
sites, five PAHs (anthracene, benzo[a]pyrene, fluoranthene, phenanthrene and pyrene)
frequently exceeded water quality benchmarks, with concentrations predicted to cause
biological impacts. Fluoranthene concentrations in Maumee and Detroit rivers were
great enough to potentially cause tumors in fish (Appendix F). Likewise, PAHs were
among the contaminants frequently detected in sediment samples.

In general, PAHs are rapidly metabolized by vertebrates, consequently, they do not
accumulate in fish or birds as do more persistent legacy contaminants. Nonetheless,
there was ample evidence for PAH exposure and accumulation among the invertebrates
sampled, noting that invertebrates do not metabolize PAHs as readily as vertebrates.
For example, PAHs were widely detected in tissues of mussels deployed for monitoring
in harbors, tributaries, nearshore, and offshore sites around the Great Lakes Basin
(Appendix B). Additionally, across the Great Lakes, emergent aquatic insects ingested
by tree swallows were contaminated with PAHs, indicating that animals that eat these
invertebrates are subsequently exposed these chemicals (Appendix C). There was
enough exposure to PAHs at some sites, as measured in the diet of swallows, to elicit
both a physiological response and a reproductive effect in the swallows. Since aquatic
insects are part of the diet of many other aquatic and wildlife species, many more
animals are likely to ingest PAHs. These observations, along with the exceedance of
water quality benchmarks for PAHs at several sites, provide multiple lines of evidence
that exposure to and uptake of PAHs is likely widespread in Great Lakes fish and
wildlife.

Given the presence of PAHs in water and sediment, accumulation in invertebrates, and
probable uptake and exposure in vertebrates suggested by chemical monitoring, we
investigated whether there was evidence that PAHs were eliciting a biological response.
Several lines of evidence indicate that both fish and birds near Great Lakes tributaries
exhibit physiological responses consistent with exposure to PAHs, and that these
responses were statistically correlated with presence of PAHs. For example, a common
biomarker of exposure to PAHs is expression of the gene cypla which codes for a
protein that modifies many organic chemicals to make them easier for organisms to
excrete. Expression of cypla and the activity of the enzyme it codes for (as measured
by 'EROD activity") increases in response to halogenated aromatic hydrocarbons and
PAHs. In studies of swallows nesting near Great Lakes tributaries, PAHs were found to
be a major contributor (22.8%) to increased cypla activity as well as enzyme activities
commonly triggered when organisms experience oxidative stress which that can
damage tissues (total sulfhydryl activity, 19.2%; Appendix C). Total PAHs were 16 times
greater in the high EROD activity group compared to the normal group.

16


-------
Increased expression of cypla was also found in livers of caged fish exposed at sites in
the Maumee and Detroit River AOCs, suggesting exposures to PAHs and/or
halogenated aromatic hydrocarbons (Appendix F). Changes in global gene expression
in caged fish were also most highly correlated to changes in PAH concentration in
surrounding water, and phenanthrene was highly correlated with levels of cypla
expression. Fluoranthene concentrations were correlated with many changes in gene
expression. This was consistent with fluoranthene being at levels exceeding water
quality benchmarks (Appendix A, Baldwin et al. 2016) and hazard indices for tumor
formation (Appendix A). Resident fish (brown bullhead, white sucker, largemouth and
smallmouth bass) in the Ashtabula, Detroit, Genesee, Milwaukee and St. Louis Rivers
and Conneaut Creek have also been found to have elevated levels of cypla and
oxidative stress-related gene expression consistent with exposure to PAHs (Braham et
al. 2017). Collectively, these findings provide strong support that PAHs are likely
causing biological responses in fish and vertebrate wildlife in the Great Lakes
tributaries, even though there is limited accumulation in tissues.

This is cause for concern as PAHs are known to cause a wide range of adverse effects
including cardiotoxicity, developmental effects, reduced reproduction, DNA damage,
liver and skin tumors, and physical abnormalities in both vertebrates and invertebrates.
Several Great Lakes tributaries have been noted to have beneficial use impairments
related to fish tumors or other deformities, deformities or reproductive problems in birds
or other wildlife, and degraded fish and wildlife populations. Studies of nesting swallows
in the Great Lakes (Appendix C) identified a decrease in reproductive success as PAH
exposure increased (Custer et al. 2018). However, due to the potential effects of other
contaminants (e.g., halogenated aromatic hydrocarbons), as well as ecological
variables such as female age and date within season, a more definitive linkage is hard
to establish (or reject) without additional study. Studies of sunfish (Appendix D) exposed
to Great Lakes tributaries, in cages or as resident fish, suggest that the presence of
PAHs and pharmaceuticals influenced blood glucose concentrations, liver anatomy,
reproductive organ size and overall maturity levels. Likewise, based on existing adverse
outcome pathway (AOP) descriptions (Appendices E, F) linking cypla induction and
liver damage to potential tumor formation, chronic exposure to PAHs is a plausible
contributor to tumors in fish. Investigation of global gene expression in fathead minnows
caged for 4 days at sites in the Detroit and Maumee rivers appear to support that
association, as gene expression changes related to activation of cypla, onset of a fatty
liver condition, and activation of genes linked to tumor promotion were all detected
(Appendix F). This is consistent with observations of internal and external tumors found
in native brown bullhead (Ameiurus nebulosus) from the Trenton channel in the lower
Detroit river (Appendix D; Blazer et al. 2014, Braham et al. 2017). Overall, there is
significant evidence suggesting that PAHs have a strong likelihood to contribute to
adverse biological effects and likely represent ongoing threats to fish and wildlife in the
Great Lakes. Consequently, PAHs were identified as a priority contaminant class for
further investigation under Action Plan II.

Estrogens as a Threat to Great Lakes Fish and Wildlife

17


-------
Chemicals that can bind to and activate vertebrate estrogen receptors, thereby
mimicking effects of this important class of reproductive and developmental hormones,
were another group of contaminants for which an integrated analysis was performed. As
part of the research under Action Plan I, researchers used four complementary
approaches capable of detecting estrogenic chemicals and/or providing insights into
their potential effects. The first approach involved monitoring for chemicals that had
been previously identified as estrogenic. These included both non-steroidal
industrial/commercial chemicals like bisphenol A, alkylphenols and alkylphenol
ethoxylates (Appendices A, B; Baldwin et al. 2016), as well as steroidal compounds
either naturally excreted by vertebrates, including humans, or administered as
therapeutics (Appendix D; Elliott et al. 2017; Jorgenson et al. 2018). Concentrations
detected were multiplied by relative potency factors to express the total expected
estrogenicity in terms of a concentration of a standard reference chemical, 17(3-
estradiol, assuming additivity. Summed and normalized concentrations, expressed as
17(3-estradiol equivalents (EEQchem; ng/L), were then compared with benchmark
concentrations that would be expected to elicit estrogenic responses in fish. This
approach considers only the fraction of estrogenic chemicals that were both detected
analytically, and for which relative potency factors have been previously reported in the
peer-reviewed scientific literature.

A second approach involved the use of cell culture-based assays to screen water
extracts for estrogenic activity (Appendix E; Cavallin et al. 2016; Davis et al. 2016; Li et
al. 2017; Jorgenson et al. 2018). Magnitude of response elicited by each sample extract
was compared to that evoked by different dilutions of the reference chemical (17(3-
estradiol) and regression was used to express the activity of the sample in terms of an
equivalent concentration (EEQbio; ng/L). In contrast with the analytical-based approach,
the cell-based assay method accounts for all estrogenic or anti-estrogenic chemicals
present in the sample extract and does not necessarily assume an additive model.
Rather the cells respond to the integrated effect of the mixture, whether that includes
additive, antagonistic, or potentially even synergistic effects of multiple chemicals.

A third approach examined biochemical and physiological responses in laboratory-
reared fish that were caged in site water for 4 days (Appendix E). For these shorter-term
exposures, induction of the egg yolk precursor protein vitellogenin (VTG) in male fish
was taken as an indicator of exposure to estrogens. Additionally, gene expression
profiling was used, in some instances, to detect estrogenic activity in exposed fish
(Appendix F). Unlike cell-based assays that only consider the fraction of chemicals that
are extracted from a water sample, the caged-fish approach accounts for all compounds
bioavailable in the water column to which the fish are exposed. However, unlike the
previous two methods, one cannot easily estimate the total concentration of EEQs from
the caged fish response.

Finally, the fourth approach involved monitoring for estrogenic responses in resident
fish. In this case, concentrations of VTG in blood plasma of male fish, as well as tissue
damage or abnormalities observed through microscopic analysis were examined as
signs of estrogenic effects (Appendix D). The resident fish approach, while most difficult
to implement from a logistical and statistical perspective, allows one to potentially

18


-------
account for chronic, life-long, exposure to mixtures of estrogenic compounds in the
water column and diet. Consequently, the methods employed were complementary,
with each having its own advantages and disadvantages.

Within the scope of this investigation, these monitoring approaches were not
necessarily brought to bear on the same locations at the same time, which somewhat
limits their comparability. Nevertheless, considering the results in an integrated manner
provides a more comprehensive picture of the state of estrogenic CEC contamination in
Great Lakes tributaries than consideration of any approach alone. For example, based
on analytical and cell-based screening of over 250 samples collected in seven
tributaries (e.g., St Louis River, Detroit River, Maumee River, Fox River, Milwaukee
River, Menominee River, and Kinnickinnic River) distributed across five Great Lakes
AOCs in 2010-2013, it was concluded that although estrogenic activity was broadly
detected across sites, it was rarely above concentrations expected to impact fish
(Appendix E). Additionally, the estrogenic chemicals detected in those samples could
reasonably account for the activity observed (Appendix E). However, broader
surveillance of over 700 samples from 57 different tributaries (Appendix A) suggested
that some of the highest concentrations of EEQchem were detected at locations, such as
the Au Sable, St. Joseph, Rouge, Saginaw, and Raisin Rivers, that were not evaluated
using effects-based approaches like testing water extracts in cell-based bioassays or
measuring VTG in caged or wild-caught male fish. Considering only non-steroidal
estrogens, Baldwin et al. (2016) calculated maximum EEQs ranging from 4-24 ng/L;
concentrations that would be of concern biologically. Likewise, a study of 12 tributaries
in 2013 and 2014 that considered both steroidal and non-steroidal estrogens calculated
maximum EEQ concentrations as high as 28 ng/L, with mean concentrations exceeding
10 ng/L at several sites including the Little Calumet and North Shore Channel (Elliot et
al. 2017) and Chicago River (Thomas et al. 2017). Cell-bioassay data were not
available for either of these broader surveillance efforts, limiting the ability to determine
whether the measured contaminants reasonably account for total biological activity
detected at the other sites. Nonetheless, the levels of EEQchem detected at those sites
are a concern relative to benchmarks proposed for total 17(3-estradiol equivalents in
drinking water or treated wastewaters (see Appendix E for more detail).

Data from biologically-based monitoring of responses in caged or resident fish were
challenging to interpret. Given the ubiquitous occurrence of CECs across the Great
Lakes basin and variation in environmental variables from one site and season to
another, identification of appropriate reference or control populations for statistical
comparisons across sites is neither ideal nor clear cut. For example, both Jorgenson et
al. (2018) and Thomas et al. (2017) detected differing concentrations of VTG in male
fish from different sites. However, without a spatial gradient relative to a source, it was
unclear whether the levels detected were just a slight deviation from the normal mean of
the population, or whether all sites were being impacted to different degrees. To
alleviate this interpretation challenge, Blazer et al. (2018) evaluated their resident fish
responses relative to an expected site-independent baseline concentration (10 |jg
VTG/ml plasma). Using this approach, results suggested estrogenic impacts in a large
majority of resident fish sampled across seven different tributaries in 2010-2011 (Blazer
et al. 2018). If one assumes this to be an appropriate baseline, results suggest that

19


-------
chronic exposures, even at levels of total EEQ estimated to be non-hazardous based on
laboratory studies, may be having long-term endocrine disrupting effects on resident
fish.

Overall, there were multiple lines of evidence to suggest that, in some Great Lakes
tributary locations, estrogenic CECs may be reaching concentrations that may result in
adverse effects in fish. Increased coordination in the application of analytically-based,
cell assay-based, and fish-based approaches for monitoring estrogenic contamination,
deployed as part of the research effort under Action Plan II, is expected to yield further
insight and clarity regarding the following: (1) whether estrogenic CECs currently
monitored via analytical methods reasonably account for the majority of estrogenic
activity measured in cell-based assays, across a broad range of Great Lakes tributaries;
(2) whether EEQ concentrations that exceed certain concentrations reliably and
reproducibly induce biological response in caged fish exposed for a defined duration,
and could be established as an actionable benchmark for management purposes.

20


-------
Overarching Management Implications

It is important for natural resource managers to be aware that CECs are ubiquitous
throughout the Great Lakes Basin. However, that is not unique to the Great Lakes Basin
and CEC prevalence probably exists anywhere there are human influenced landscapes.
A better understanding of CEC presence and possible effects on fish and wildlife will
lead to improved natural resource management and more successful restoration,
conservation, and mitigation efforts. Even where CECs alone may not be causing
obvious adverse effects to Great Lakes fish and wildlife, when combined with other
stressors (e.g., invasive species, habitat loss or alteration, viruses or pathogens, and/or
changing climate conditions), CECs may contribute to overall reductions in population
fitness. Understanding what levels of CECs are in the current landscape helps prepare
natural resource managers to better understand risks and effects to fish and wildlife
populations. CEC presence and effects data can be utilized to further evaluate
strategies and inform best management practices in areas where CECs are present.

CECs are Widespread

CECs were detected in all sampled media (water, sediment, tissues) at all Great Lakes
tributary stream locations. Although this study focused on the Great Lakes Basin,
results could be applicable to nearly any US watershed because of ubiquitous presence
of CECs revealed by a recent nation-wide CEC monitoring study (Bradley et al. 2017).

If natural resource managers are concerned that CECs may be impeding management
goals and objectives in their areas, screening water, sediment, or organisms for the
occurrence of CECs can be a prudent first step to determine if CECs are present at
levels known to cause adverse effects. The surveillance data collected under Action
Plan I may help identify specific classes of CECs to screen for in different locations.
Further, if funding constraints dictate, analysis could be prioritized to first assess the
most commonly detected CECs, gradually expanding the analysis to include other
suites of CECs as time and resources allow. Monitoring of CECs will help natural
resource managers better understand possible risks to natural resources, in relation to
other stressors and pressures on fish and wildlife populations (e.g., climate change,
habitat loss/alterations, invasive species, etc.), and help prioritize restoration projects.

Effects are Often Subtle and Implications for Ecological Fitness Unclear

In nearly all Great Lakes tributary sites evaluated as part of this effort, CECs were not
lethal to fish, mussels, or tree swallows. Biological effects were observed in many
cases, but their severity was mild and their relevance to ecological fitness remains
uncertain. The results indicate that in many cases, CEC exposure may be a factor
contributing to or compounding other stressors, but that CECs are not likely to be the
sole contributing factor to population declines, die-offs, or unsuccessful population
restoration efforts. Therefore, natural resource managers should not evaluate CECs in
isolation from other environmental stressors.

21


-------
For some CECs, benchmarks established by regulatory authorities or screening values
derived from peer-reviewed literature sources are available and can be used for guiding
management objectives for remediation or source reduction.

The effects of long-term exposure of CEC on the reproductive health of organisms is
still unclear. Our investigations provided evidence of potential CEC-related impacts on
tree swallow reproduction. Additionally, there was evidence of physiological stress (e.g.,
altered glucose levels; signs of oxidative stress) in multiple taxa including mussels,
birds, and fish that could influence reproductive fitness. However, more monitoring is
needed to further investigate links of CEC exposure to effects on reproductive fitness.

Identifying and Reducing Sources of PAHs Could Reduce Contaminant-Related
Stress in a Number of Great Lakes Tributaries

PAHs were consistently detected throughout all the studies, frequently at levels
exceeding benchmark standards (Table 2). As a result, monitoring for this chemical
class could be a priority for managers looking to better understand sources, transport
and fate of CECs in their landscape. Because many of the PAH concentrations
exceeded benchmark standards, there is increasing cause for concern regarding their
effects on natural resource management objectives. Management strategies could
include developing a monitoring schedule for PAHs or looking for evidence of effects on
fish and wildlife populations (e.g., reduced survival, reductions or failures in
reproduction, rates of tumors and deformities). If benchmarks and/or custom screening
values standard values are exceeded, source identification and source reduction
strategies actions could take place to reduce PAH levels to, or below, acceptable
standards.

Complementary Methods Show Promise for CECs Surveillance and Monitoring

Multiple methods were used to detect CECs and/or their effects on organisms. Natural
resource managers looking to implement CEC monitoring and surveillance should
choose the method most appropriate to their management objectives. Taking grab
samples of water will yield snapshot results of CEC concentrations at the time of
sampling. Longer term sampling (e.g., on a monthly, seasonal, or yearly basis) will
provide managers with a clearer picture of long-term exposure and CEC prevalence or
occurrence. Sampling of sediments from depositional areas can detect persistent
contaminants deposited over long periods of time (depending on the depth sampled) but
will typically include different contaminants than those found in the water column.
Screening organism tissue samples for the occurrence of CECs will help managers
better understand what CEC types and concentrations are present in biota. Evaluation
of physiological and biological responses to CECs can help natural resource managers
determine if risks to fish and wildlife populations exist for specific class(es) of CECs.
Members of the CEC research team can assist in selecting the appropriate approaches
to address site-specific problems (see https://commynities.geoplatform.gov/glri/ for
technical contacts).

Implementation of Management Practices is Influenced by Contaminant Source

22


-------
CECs have many sources (i.e., point, nonpoint) and pathways throughout environmental
systems. Wastewater treatment plant (WWTP) effluents, combined sewer overflows (CSOs),
illicit discharges, agricultural and urban runoff, and direct inputs can all be possible sources of
CECs. Evaluating the landscape for point source and non-point source pollution sources can
help managers prioritize project locations and evaluate the potential risks CECs may cause to
conservation success. Below we highlight some previous examples where an understanding of
CECs and their sources helped to inform management decisions and subsequent actions:

•	Sampling conducted under EPA's National Urban Runoff Program initiative and Clean
Lakes initiative with the city of Austin, TX, USGS, and Texas State University in 2005
determined coal tar sealant was negatively impacting aquatic communities. The city enacted
a city-wide ban on coal tar-based sealants upon the determination that PAHs from the
sealants were causing harm to wildlife, including the federally endangered Barton Springs
salamander. Follow up studies conducted by USGS indicate that, as of 2014, PAH
concentrations have declined by as much as 58% in some locations (City of Austin, 2005).

•	Limiting jeopardy of endangered species by informing federal actions was illustrated
previously in the use of lampricides in areas of endangered species occurrence. To stop sea
lamprey damage, application of the lampricide 3-triflouromethyl-4-nitrophenol (TFM) was
necessary in Great Lakes streams which also contain the endangered freshwater mussel
snuffbox (Epioblasma triquetra), and their host fish log perch (Percina caprodes), as well as
feeding and nesting habitat of the endangered shorebird piping plover

(Charadrius melodus). Data similar to those generated by this CEC project was used to
assess the potential impact TFM may have on these endangered species (Boogard et al.
2013, 2015). This analysis led to recommendations by USFWS that allowed best
management action plan for TFM application timing and concentration that was protective of
the species of concern as well as ensuring eradication of the sea lamprey.

•	State/federal coordination of herbicide permitting to eliminate aquatic nuisance plants
from lakes and streams illustrates another application for CEC data. When permits were
requested by homeowners to apply herbicides along a chain of lakes in Michigan that
historically hosted the endangered snuffbox mussel, the state with USFWS guidance
developed a recommended herbicide guide utilizing data and principles similar to those
generated by the CEC project. This living guidance document enables state permitting
officers to allow permits based on science for best management practices for the types,
rates, and timing of herbicides to limit jeopardy of snuffbox populations and other aquatic
species.

•	Working with hatcheries and streamside rearing facilities to determine placement of new
facilities is another application of CEC data. Streamside rearing facilities and traditional
hatcheries pump water from adjacent streams for use in the rearing tanks. Some facilities
are located near urban centers or in waterways known to contain CECs that can be
detrimental during the vulnerable life stage of rearing from egg to juvenile (typically May -
Oct.). Monitoring for CECs and prioritizing locations for new facilities with lower CEC
occurrence and concentrations when all other site aspects are equal is a possible solution to
decrease CEC exposure.

•	Plasticizers, antioxidants, detergent metabolites, flame retardants, nonprescription
drugs, flavors/fragrances, dyes/pigments, and human-associated bacteria were all greatest
in watersheds with the most urban influence. Many of these CECs have been shown to
originate from imperfect sanitary conveyance systems that can be rapidly repaired. For
example, in 2007, samples from an outfall to Honey Creek near Miller Park in Milwaukee,
Wisconsin were positive for human-associated bacteria markers. Follow-up dye testing
confirmed a misconnection of a sewage line from Miller Park that was remedied to improve
water quality in Honey Creek. Enhancing illicit discharge detection and elimination programs
(IDDE) would enable identification and remedy of misconnections and faulty sewer pipes

23


-------
around the Great Lakes similar to this incident in the Honey Creek watershed (Behm
2007).

• Other CECs that are commonly particle associated such as PAHs were also most
prevalent in urban areas. Multiple studies have demonstrated that PAHs in runoff can be
removed in common urban stormwater using green infrastructure practices that focus on
settling or filtration of particulate matter. One example at the University of Maryland campus
in 2006-2007 demonstrated that a bioretention cell removed 87% of the PAHs from urban
stormwater that were retained in surficial deposits, allowing for maintenance efforts
to remove and dispose of these contaminants (DiBlasi et al. 2009). Installation of urban
runoff and green infrastructure may similarly reduce PAH loadings to the Great Lakes and
its tributaries.

Data Availability

To assist in natural resource manager decision-making, all data from this investigation
are publicly available in databases, in peer-reviewed open source publications, or by
contacting the corresponding author of the research project summaries (Appendices A-
F) or the technical contacts listed at https://commynities.geoplatform.gov/glri/.

•	Tree swallow data is available through the Environmental Conservation Online
System (ECOS, https://ecos.fws.gov/ecp/), which can be accessed using the
Wildlife & Environmental Contaminants mapper.

•	Data collected by NOAA can be accessed using the DIVER Explorer Application
tool (https://www.diver.orr.noaa.gov/).

•	Data collected by USGS with USFWS can be found in the USGS National Water
Information System (https://waterdata.usgs.gov/nwis?) and the USGS Science
based-catalog online system (https://www.sciencebase.gov/catalog/).

•	Data from EPA is made publicly accessible via ScienceHub and can be accessed
via the EPA Environmental Dataset Gateway (edg.epa.gov) or at Data.gov

•	Transcriptomics data from ACOE is publicly accessible via the Gene Expression
Ominbus site (GEO; https://www.ncbi.nlm.nih.gov/geo/).

Links and/or direct database access to these data sources will also be made available
through https://communities.geoplatform.gov/glri/ as soon as feasible.

24


-------
References

Baldwin, AK, Corsi, SR, De Cicco, LA, Lenaker, PL, Lutz, MA, Sullivan, DJ, & Richards,
KD. 2016. Organic contaminants in Great Lakes tributaries: Prevalence and potential
aquatic toxicity. Science of The Total Environment. 554-555:42-52.

https://doi.orq/10.1016/i.scitotenv.2016 02.137.

Behm D. 2007. Honey creek tainted with human waste. Milwaukee Journal Sentinel.
January 21, 2007. http://www.isonline.com/news/milwaukee/29416799.html (accessed April
2016)

Blackwell BR, Ankley GT, Corsi SR, DeCicco LA, Houck KA, Judson RS, Li S, Martin
MT, Murphy E, Schroeder AL, Smith ER, Swintek J, Villeneuve DL. 2017. An "EAR"
on environmental surveillance and monitoring: A case study on the use of Exposure-
Activity Ratios (EARs) to prioritize sites, chemicals, and bioactivities of concern in
Great Lakes waters. Environ. Sci. Technol. 51 (15):8713-8724.
https://doi.CiM i*1 i*1 i .cs est hM* i

Blazer VS, Mazik PM, Iwanowicz LR, Braham RP, Hahn CM, Walsh HL, Sperry AJ.
2014. Assessment of the fish tumor beneficial use impairment in brown bullhead
(Ameiurus nebulosus) at selected Great Lakes Areas of Concern: US Geological
Survey Open-File Report 2014-1105. 17 p. http://dx.doi.OiM !>' i 3/gti >1 U i .

Blazer VS, Walsh HL, Shaw CH, Iwanowicz LR, Braham RP, Mazik PM. 2018.

Indicators of exposure to estrogenic compounds at Great Lakes Areas of Concern:
Species and site comparisons. Environ Monit Assess. 190(10):577.
https://doi.c	943-5.

Boogaard MA, Hubert TD, Bernardy JA, Kaye CA, Baldwin GA. 2013. Evaluation of 3-
trifluoromethyl-4-nitrophenol (TFM) residues following a lampricide treatment as a
risk assessment to the endangered piping plover. J. Great Lakes Res. 38(2): 362-
367. https://doi.on	ilr 2012 01.009

Boogaard MA, Newton TJ, Hubert TD, Kaye CA, Barnhart MC. 2015. Evaluation of the
short term 12 hour toxicity of 3-trifluoromethyl-4-nitrophenol (TFM) to multiple life
stages of Venustaconcha ellipsiformis and Epioblasma triquetra and its host fish
(Percina caprodes). Environ Toxicol Chem. 34(7): 1634-41. doi: 10.1002/etc.2959.

Bradley PM, Journey CA, Romanok KM, Barber LB, Buxton HT, Foreman WT, Furlong
ET, Glassmeyer ST, Hladik ML, Iwanowicz LR, Jones DK, Kolpin DW, Kuivila KM,
Loftin KA, Mills MA, Meyer MT, Orlando JL, Reilly TJ, Smalling KL, Villeneuve DL.
2017. Expanded target-chemical analysis reveals extensive mixed-organic-
contaminant exposure in U.S. streams. Environ Sci Technol. 51(9):4792-4802.
https://doi c	ics est 7b00012.

25


-------
Braham RP, Blazer VS, Shaw CH, Mazik PM. 2017. Micronuclei and other erythrocyte
nuclear abnormalities in fishes from the Great Lakes Basin, USA. Environ Mol
Mutagen. 58(8):570-581. https://doi.Org/10.1002/em 22123.

Cavallin JE, Jensen KM, Kahl MD, Villeneuve DL, Lee KE, Schroeder AL, Mayasich J,
Eid EP, Nelson KR, Milsk RY, Blackwell BR, Berninger JP, LaLone CA, Blanksma C,
Jicha T, Elonen C, Johnson R, Ankley GT. 2016. Pathway-based approaches for
assessment of real-time exposure to an estrogenic wastewater treatment plant
effluent on fathead minnow reproduction. Environ Toxicol Chem. 35(3):702-16.
https://doi.c	)2/etc 3228.

Choy SJ, Annis ML, Banda JA, Bowman SR, Brigham ME, Elliott SE, Gefell DJ,

Jankowski MD, Jorgenson ZG, Lee KE, Moore JN, Tucker WA. 2017. Contaminants
of emerging concern in the Great Lakes Basin: A report on sediment, water, and
fish tissue chemistry collected in 2010-2012. US Fish and Wildlife Service.

Biological Technical Publication BTP-R3017-2013. 90 p.

City of Austin, TX. 2005. PAHs in Austin, Texas sediments and coal-tar based
pavement sealants polycyclic aromatic hydrocarbons.

(https://www.austintexas.gov/sites/default/files/files/Watershed/coaltar/PAHs_in_Au
stin_2005_final.pdf)

Custer CM, Custer TW, Delaney R, Dummer PM, Schultz S, Karouna-Renier N. 2019.
Perfluoroalkyl contaminant exposure in tree swallows nesting at Clarks Marsh,
Oscoda, Ml. Arch. Environ. Toxicol. Chem. 77:1-13.

Custer CM, Custer TW, Etterson MA, Dummer PM, Goldberg D, and Franson JC. 2018.
Reproductive success and contaminant associations in tree swallows (Tachycineta
bicolor) used to assess a Beneficial Use Impairment in U.S. and Binational Great
Lakes' Areas of Concern. Ecotoxicol. 27:457-476. https://doi.Org/10.1007/s10646-

018-1913-9.

Custer TW, Custer CM, Dummer PM, Goldberg D, Franson JC, and Erickson RA. 2017.
Organic contamination in tree swallow (Tachycineta bicolor) nestlings at United
States and Binational Great Lakes Areas of Concern. Environ. Toxicol. Chem.
36:735-748.

Davis JM, Ekman DR, Teng Q, Ankley GT, Berninger JP, Cavallin JE, Jensen KM, Kahl
MD, Schroeder AL, Villeneuve DL, Jorgenson ZG, Lee KE, Collette TW. 2016.
Linking field-based metabolomics and chemical analyses to prioritize contaminants
of emerging concern in the Great Lakes basin. Environ Toxicol Chem. 35(10):2493-
2502. https://doi c	tc 3409.

Diblasi CJ, Li H, Davis AP, Ghosh U. 2009. Removal and fate of polycyclic aromatic
hydrocarbon pollutants in an urban stormwater bioretention facility. Environ. Sci.
Technol. 43: 494-502. https://doi.c	!l/es802090g

26


-------
Dila, DK, Corsi SR, Lenaker PL, Baldwin AK Bootsma MJ, McLellan SL. 2018. Patterns
of host-associated fecal indicators driven by hydrology, precipitation, and land use
attributes in Great Lakes watersheds. Environ Sci Technol. 52(20): 11500-11509.

https://doi c	M/acs est 8b01945.

Elliott SM, Brigham ME, Lee KE, Banda JA, Choy SJ, Gefell DJ, MinarikTA, Moore JN,
Jorgenson ZG. 2017. Contaminants of emerging concern in tributaries to the
Laurentian Great Lakes: I. Patterns of occurrence. PLOS ONE 12(9): e0182868.

https://doi.c	3umal.pone.0182868.

Gefell DJ, Banda J, Moore JN, Secord AL, Tucker WA. 2019a. Ecological Hazard
Assessment of Contaminants of Emerging Concern in the U.S. Great Lakes Basin:
Part B - Supplemental Document: Technical Resources for Ecological Hazard
Assessments of Contaminants of Emerging Concern in Freshwater Fish. U.S. Fish
and Wildlife Service, Biological Technical Publication BTP-R3018-2019. .

Gefell DJ, Annis M, Banda J, Bellamy A, Choy S, Hummel S, Jorgenson Z, Moore JN,
Secord AL, Tucker WA. 2019b. Ecological Hazard Assessment of Contaminants of
Emerging Concern in the U.S. Great Lakes Basin: Part A - Screening Assessment
of Relative Hazard to Fish from Surface Water Exposures to Fourteen Contaminants
of Emerging Concern in the Great Lakes Basin. U.S. Fish and Wildlife Service,
Biological Technical Publication BTP-R3019-2021.

Great Lakes Restoration Initiative (GLRI). 2010. Action Plan I, FY2010-FY2014. 41 p.

https://www.glri.us/sites/default/files/glri actionplan pdf.

Jorgenson ZG, Thomas L, Elliott SM, Cavallin JE, Randolph EC, Choy SJ, Alvarez DA,
Banda JA, Gefell DJ, Lee KE, Furlong ET, Schoenfuss HL. 2018. Contaminants of
emerging concern presence and adverse effects in fish: a case study in the
Laurentian Great Lakes. Environ Pollut. 236:718-733.
https://doi.c	envpol 2018 01.070.

Kavlock R, Chandler K, Houck K, Hunter S, Judson R, Kleinstreuer N, Knudsen T,
Martin M, Padilla S, Reif D, Richard A, Rotroff D, Sipes N, Dix D. 2012. Update on
EPA's ToxCast program: Providing high throughput decision support tools for
chemical risk management. Chem Res Toxicol. 25(7): 1287-302.

https://doi.c	00939.

Krewski D, Acosta D Jr, Andersen M, Anderson H, Bailar JC 3rd, Boekelheide K, Brent
R, Charnley G, Cheung VG, Green S Jr, Kelsey KT, Kerkvliet Nl, Li AA, McCray L,
Meyer O, Patterson RD, Pennie W, Scala RA, Solomon GM, Stephens M, Yager J,
Zeise L. 2010. Toxicity testing in the 21st century: A vision and a strategy. J Toxicol
Environ Health B Crit Rev. 13(2-4):51-138.
https://doi.cu s s ^0/10937404 i I i

Li S, Villeneuve DL, Berninger JP, Blackwell BR, Cavallin JE, Hughes MN, Jensen KM,
Jorgenson Z, Kahl MD, Schroeder AL, Stevens KE, Thomas LM, Weberg MA,

Ankley GT. 2017. An integrated approach for identifying priority contaminant in the

27


-------
Great Lakes Basin - Investigations in the Lower Green Bay/Fox River and
Milwaukee Estuary Areas of Concern. Sci Total Environ. 579:825-837.

https://doi.CiM mm* \ scitotenv : *M* n *>21.

Nakayama SF, Strynar MJ, Reiner JL, Delinsky AD, Lindstrom AB. 2010. Determination
of perfluorinated compounds in the upper Mississippi River basin. Environ. Toxicol.
Chem. 44:4103-4109.,

Thomas LM, Jorgenson ZG, Brigham ME, Choy SJ, Moore JN, Banda JA, Gefell DJ,
MinarikTA, Schoenfuss HL. 2017. Contaminants of emerging concern in tributaries
to the Laurentian Great Lakes: II. Biological consequences of exposure. PLOS ONE
12(9): e0184725. https://doi.on	/journal pone 0184725.

28


-------
Appendix A

Organic Contaminants, Microplastics,
Waterborne Pathogens, and Host-
Associated Bacteria Surveillance and
Potential Biological Effects in Great
Lakes Tributaries

Lead Organization: U.S. Geological Survey, Upper Midwest Water Science Center,
Middleton, Wl, USA

Contributing Authors: Steven R. Corsi, Michelle A. Lutz, Austin K. Baldwin, Peter L.
Lenaker

Contributing Investigators: Baldwin AK, Bootsma MJ, Borchardt MA, Corsi SR, De
Cicco LA, Dila DK, Lenaker PL, Lutz MA, Mason, SA, McLellan SL, Richards KD,
Spencer SK, Sullivan DJ

Corresponding Author Contact: srcorsi@usgs.gov
A.1. Introduction

Anthropogenic activities related to industrial, agricultural, domestic, and urban water
uses introduce an untold number of contaminants into the Great Lakes and their
tributaries on a daily basis (Bennie et al. 1997; Blair et al. 2013; Venier et al. 2014).
Flame retardants, drugs, herbicides, plasticizers, polycyclic aromatic hydrocarbons
(PAHs), and other organic compounds (collectively referred to as organic waste
compounds or OWCs) as well as plastic debris, and microbiological organisms enter
waterways through wastewater treatment plant (WWTP) discharges, combined sewer
overflows, leaking septic and municipal sewer systems, urban and agricultural runoff,
industrial discharges, and atmospheric deposition, among others (Barber et al. 2015;
Kolpin et al. 2002).

A surveillance program was conducted in Great Lakes tributaries from 2010-2014 to: 1).
define general occurrence and magnitude of multiple classes of OWCs, plastic debris,
and microbiological organisms, 2). to prioritize OWCs based on prevalence and
potential for effects on ecological species, 3). to define the frequency of occurrence and
magnitude of microplastics contamination, and 4). to use microbiological genetic
markers to define contamination from sewage and cattle manure sources, and 5). to
define the various conditions under which each of these contaminants are most
prevalent.

29


-------
A.2. Organic Contaminants

Tributaries of the Great Lakes are impacted by a diverse set of influences that can
introduce numerous contaminants into waterways. Exposure to many of these
compounds can result in adverse ecological effects, with some that can be serious
enough to lead towards a decline or collapse in populations (Collier et al. 2013; Ingersoll
et al. 2002; Johnson et al. 2013). The ability of some of these compounds to
bioaccumulate (Ismail et al. 2014; Jenkins et al. 2014) creates a risk to organisms
higher up the food chain including mink, river otter, bald eagles, osprey, and humans
(Hinck et al. 2009; Jenkins et al. 2014). Most drinking water treatment plants do not fully
remove many of these compounds from the water supply, creating another exposure
route for humans (Kingsbury et al. 2008; Stackelberg et al. 2004).

A number of factors may influence the occurrence of contaminants in environmental
waters including land use, hydrologic condition and season. As a result, there is a need
for management strategies that consider many different factors including a multitude of
chemical contaminants from different sources to establish priorities for resource
allocation.

A.2.1 Objectives

The overall objective of the study was to assess the occurrence and possible adverse
biological effects of organic contaminants in Great Lakes tributaries. Specific objectives
were as follows: 1) identify occurrence and magnitude of monitored compounds, 2)
define how presence and magnitude of these compounds vary by land cover, flow
regime, and season, 3) prioritize compounds based on screening techniques that
estimate potential for biological effects, 4) prioritize tributaries with respect to potential
biological effects, and 5) develop techniques for use of high-throughput screening data
(ToxCast: https://www.epa.gov/chemical-research/toxicity-forecasting) to expand
capabilities for assessing potential biological effects from chemical concentrations and
6) develop techniques to link the chemical exposures to potential adverse outcomes for
ecological species using the adverse outcome pathway wiki (AOP-Wiki; aopwiki.org). In
this first phase of GLRI, objectives 1 and 2 were completed, a traditional method of
achieving objectives 3 and 4 was completed, and the groundwork was established for
achieving objectives 3-6 using the ToxCast database and the AOP-Wiki. The current
section on organic contaminants in Great Lakes tributaries provides a summary of
findings from the first phase of GLRI; additional detail and supporting data are published
elsewhere (Baldwin et al. 2016a).

A.2.2 Methods

Study sites included 57 Great Lakes tributaries (Figure A.1). Flow from these sampling
sites accounted for approximately 41 % of the total tributary inflow to the Great Lakes.
Watershed drainage areas ranged from 101 - 16,400 square kilometers (km2), with
mean annual flows from 2.58 - 219 cubic meters per second (2010- 2013). Watershed
land cover varied from dominantly urban (up to 92% of watershed) to agricultural (84%)
to forest and wetland (93%). Watershed population densities ranged from 3.3 - 2,498

30


-------
people/ml2. The portion of river flow from WWTP effluent ranged from less than 1% -
47%.

Figure. A.l. Sampling locations, watershed boundaries, and watershed land-uses. Map IDs
are defined in Table A.l. (Instituto National de Estadistica Geografia e Informatica, 2006a,
Instituto Nacional de Estadistica Geografia e Informatica, 2006b; National Atlas of the
United States, 2005, U.S. Department of Agriculture-Natural Resources Conservation
Service, 2009).

Of the 709 water samples collected between September 2010 and September 2013,
thirty-eight sites were sampled 1-2 times each, generally during low-flow and medium-
flow periods, and the remaining 19 sites were sampled more frequently, with 7- 64
samples each, during runoff and low-flow conditions and throughout different seasons.
Flow composite samples were collected by permanently stationed automatic samplers
at eight of the 19 frequently-sampled sites (Table A.1). Whole water samples were
analyzed for 69 organic waste compounds (OWCs) (USGS National Water Quality
Laboratory schedule 4433; (Zaugg et al. 2006)).

OWCs were aggregated into 15 classes: antioxidants, dyes and pigments, fire
retardants, PAHs, plasticizers, fuels, solvents, herbicides, insecticides, antimicrobial
disinfectants, detergent metabolites, flavors and fragrances, nonprescription drugs,
sterols, and miscellaneous (Baldwin et al. 2016a).

Mean contaminant concentrations by site were analyzed for relationships with land
cover attributes, streamflow condition (low-flow versus runoff), and season. Sample
results were also analyzed for potential adverse biological impact by comparison with
established water quality benchmarks from institutions such as U.S. Environmental
Protection Agency and Canadian Council of Ministers of the Environment (27
compounds) and by using 17(3-estradiol equivalency factors (8 compounds). Toxicity

31


-------
quotients (TQs) were computed for each site by dividing the greatest measured
concentration of a compound at a particular site by the lowest known water quality
benchmark for that compound. EEQ's were computed by multiplying measured
concentrations of eight nonsteroidal compounds (bisphenol A, p-dichlorobenzene, 4-
nonylphenol, 4-nonylphenol monoethoxylate, 4-nonylphenol diethoxylate, 4-tert-
octylphenol, 4-tert-octylphenol monoethoxylate, and 4-tert-octylphenol diethoxylate) by
their respective estradiol equivalency factors (EEF) and summing for each sample
(Vajda et al. 2008).

Table A.l. Site characteristics and types of samples collected, 2010-2013. Drainage area
from NWIS unless unavailable, then GIS computed. [A, water samples collected using an
autosampler; n, number of samples; ID, identification, AgMix, agricultural mix of
pasture/hay and crops].

Site Name

Map
ID

Dominant
land
cover

n

Site Name

Map
ID

Dominant
land cover

n

StLouis, MN

SI

Wetland

31

Cheboygan, Ml

H2

Wetland

2

Nemadji, Wl

S2

Wetland

1

Thunder Bay, Ml

H3

Wetland

2

Bad, Wl

S3

Forest

1

AuSable, Ml

H4

Forest

26

White, Wl

S4

Forest

1

Rifle, Ml

H5

Forest

2

Montreal, Wl

S5

Wetland

1

Saginaw, Ml

H6

Crops

31

Presque Isle, Ml

S6

Wetland

1

Black, Ml

El

Crops

2

Ontonagon, Ml

S7

Forest

30

Clinton, MIA

E2

Urban

43

Sturgeon, Ml

S8

Forest

1

Rouge, Ml A

E3

Urban

43

Tahquamenon, Ml

S9

Wetland

1

Huron, Ml

E4

Urban

2

Manistique, Ml

Ml

Wetland

1

Raisin, MIA

E5

AgMix

44

Escanaba, Ml

M2

Wetland

2

Maumee, OHA

E6

Crops

64

Ford, Ml

M3

Wetland

2

Portage, OHA

E7

Crops

64

Menominee, WIA

M4

Wetland

40

Sandusky, OH

E8

Crops

2

Peshtigo, Wl

M5

Wetland

1

Huron, OH

E9

Crops

2

Oconto, Wl

M6

Crops

1

Vermilion, OH

E10

Crops

2

Fox, Wl

M7

Crops

7

Black, OH

Ell

AgMix

2

Manitowoc, Wl A

M8

AgMix

43

Rocky, OH

E12

Urban

2

Milwaukee, WIA

M9

Urban

45

Cuyahoga, OH

E13

Urban

28

IndianaHC, IN

M10

Urban

2

Grand, OH

E14

Crops

2

Burns, IN

Mil

Urban

31

Cattaraugus, NY

E15

AgMix

1

StJoseph, Ml

M12

Crops

25

Tonawanda, NY

01

AgMix

1

Paw Paw, Ml

M13

Crops

1

Genesee, NY

02

AgMix

14

Kalamazoo, Ml

M14

AgMix

1

Oswego, NY

03

AgMix

26

Grand, Ml

M15

AgMix

2

Black, NY

04

Forest

1

Muskegon, Ml

M16

Forest

2

Oswegatchie, NY

LI

Forest

1

White, Ml

M17

Crops

2

Grass, NY

L2

Forest

1

Pere Marquette, Ml

M18

Forest

2

Raquette, NY

L3

Forest

1

Manistee, Ml

M19

Forest

2

StRegis, NY

L4

Forest

16

Indian, Ml

HI

Forest

2









A.2.3 Key findings

One or more compounds were detected in 92.5% of the 709 samples. Six sites were
absent of detections, including White (S4), Tahquamenon (S9), Manistique (M1), Black
(04), Oswegatchie (L1), and Grass (L2), all of which had only one sample collected
(Table A.1). Mixtures of 10 or more compounds were detected at 35% of sites, with a
maximum of 53 compounds detected in a single sample. The most frequently detected
class of compounds was the insecticides, with an overall occurrence rate of 60%. The

32


-------
majority of the insecticide class detections were for two compounds: DEET and
carbazole. Other frequently detected classes include the PAHs (43%), herbicides (37%,
primarily metolachlor and atrazine), and flavors/fragrances (31%, primarily HHCB and
benzophenone). All other compound classes were detected in less than 25% of
samples. The solvents, miscellaneous, and antimicrobial disinfectant classes were the
least frequently detected, with occurrence rates of less than 5%.

Watershed land cover was related to occurrence and concentration for many of the
compound classes. A pattern of relatively low concentrations in forest- and wetland-
dominated watersheds, moderate concentrations in agriculture-dominated watersheds,
and higher concentrations in urban-dominated watersheds was observed for the classes
insecticides, PAHs, plasticizers, antioxidants, detergent metabolites, fire retardants,
nonprescription drugs, sterols, flavors/fragrances, and dyes/pigments (Figure A.2). The
only class with frequent detections which did not follow this pattern was herbicides, with
concentrations in agriculture-dominated watersheds comparable to or greater than
those in urban-dominated watersheds.

Seasonal differences in compound concentrations were observed for four compounds,
all of which had relatively high detection frequencies. Compounds that had significantly
greater concentrations in warm weather months included Metolachlor, Atrazine, and
DEET, consistent with common use patterns of these chemicals. HHCB concentrations
were significantly greater in winter. Seasonal differences varied by site. For example, a
clear seasonal pattern was observed for atrazine and metolachlor in samples from the
highly agricultural Maumee and Portage Rivers in Ohio, with summertime
concentrations 1-2 orders of magnitude greater than wintertime concentrations.

One or more water quality benchmarks were exceeded in samples from 20 sites (Figure
A.3; (Baldwin et al. 2016a)). Many of the sites with regular exceedances were those
dominated by urban land cover. Water quality benchmarks were exceeded by a factor of
10 (TQ >10) at seven sites: Rouge, Indiana Harbor Canal, Clinton, Cuyahoga,
Milwaukee, St. Joseph, and Portage. The Clinton River had the most compounds with
exceedances (9), followed by the Rouge (8), St. Joseph (7), and Milwaukee (6) rivers.
Compounds with the most frequent water quality benchmark exceedances were the
PAHs benzo[a]pyrene, pyrene, fluoranthene, and anthracene, the detergent metabolite
4-nonylphenol, and the herbicide atrazine. Water quality benchmarks were exceeded by
a factor of 10 or more (up to a factor of 117) for six compounds: pyrene,
benzo[a]pyrene, fluoranthene, dichlorvos, atrazine, and anthracene.

33


-------
herbicides

fuels

insecticides

miscellaneous

1.0
0.10
ND

1.0 H

0.10

0)

<7)

O IB

AB LJ

AiL

J_ !£_ V o a_a^_a_Z_4LQU e

A A A A A

NDH

XI
CO

c

03
CD

E

c
o

CD

o ND
c
o
O

PAHs

C ABC

A°_ ^

detergent metabolites

1.0-

0.10-

o

o '
A

_ n _

sterols

1.0 -
0.10
ND

&u

T

~r

&
CD

~o

C

03

a5

5

-r

x co	c

9-	ro

•S o	_o

®	±z

< O	D

plasticizers

solvents

antioxidants

IB

° AB AB

A?_ A_ D	 Q

—

AB

A A A A

A_ A_ ~ M

fire retardants

antimicrobial dis.

non-prescription drugs

A B



~

o
A°

o
AO

AB

~

AB

flavors/fragrances

dye/pigments

6

~i	r-

"55 "a
S n

o

T-

X

CD
<

—i	r~

to c
Q. CO

2 £

O 3

S3

AB

AB

(/)
CD

i_

O
LL

"D
C

a3

-r~

x

n—

CO
Q.
O

< o

o
-T-

C
CO
.Q

(/)
0

-I-

"O

-T-

X

TO

D)

0 <

-1	

C/D
CL
O

i_

o

c

03
_Q

Figure A.2. Dominant land cover and site mean concentrations of compound classes.
Number of sites per land use category: forest, 15; wetland, 12; AgMix, 9; Crops, 13; Urban,
8. Boxplot labels A, B, and C indicate which groups of samples are statistically similar
(those sharing a common letter) and statistically different (those not sharing a common
letter) using the Kruskal-Wallis multiple comparisons test (p-values < 0.05). [ND, not
detected; fig/'L, micrograms per liter; antimicrobial dis. antimicrobial disinfectants;

AgMix, agricultural mix of pasture/hay and crops].

34


-------
"O

c

3

o 4

~~~~u

0

z

TQmax >10
1.0 < TQmax < 10
0.5 < TQmax < 1.0



~~r~

—i	r

o ^

	1	1	T"

' ^ ^ ^
•£" ^ ^

—i	1	r

f f i" # ^

£ £
c7 ^

f* Sb rs

¦ £

f £
<£ «?

<2

¦?

«f /

" J?



*

*r*

&



$

/ *
/ o*

/

¦t ^

 0.5 (most to least urban, left to right)

Figure A.3. Sites with compound concentrations approaching (0.5 < TQmax < 1.0) or
exceeding (TQmax > 1.0) water quality benchmarks. Total number of samples at each site
shown in parentheses. [TQmax, maximum toxicity quotient].

Sixty-eight percent of sites had detections of nonsteroidal endocrine-disrupting
compounds (EDCs; (The Endocrine Disruption Exchange, Inc. 2012). Mixtures of EDCs
(detectable concentrations of two or more) were observed in samples from 61 % of sites-
-highlighting the importance of chemical mixtures. Twenty-three percent of sites had ten
or more EDCs (maximum = 21 at River Rouge) in a single sample. Computed EEQs
indicated medium to high risk (greater than the no observable effect concentration or
lowest observable effect concentration, respectively) of estrogenic effects for intersex or
vitellogenin induction at 10 sites. Steroidal estrogens were not measured, and therefore,
these estrogenic effects are likely considerably underestimated.

This study highlights the complexity of compound mixtures in streams, especially
streams with urban influences. There was an approximately four-fold difference in the
mean number of detected compounds per sample and the mean total sample
concentration between sites with greater than 15% urban land cover and those with less
than 15% urban land cover. Along with other urban-associated factors such as
increased stream flashiness, OWCs have potential to stress stream ecosystems and
contribute to degraded populations offish, invertebrates, and other organisms (Bell et
al. 2012).

A.3. Microplastics

Concern surrounding plastics, and especially microplastics particles (less than 5 mm in
diameter), in aquatic environments has been growing in recent years. Microplastics are
introduced into aquatic environments from a variety of sources: spillage of production
materials; atmospheric deposition; wastewater treatment plant (WWTP) effluent and
sludge; and degradation of larger items, such as Styrofoam, plastic bags, bottles,
wrappers, cigarette butts, and tires.

35


-------
The ecological consequences of microplastic contamination is an active area of
research and includes impacts at multiple trophic levels, uptake, accumulation,
associated adverse impacts on reproduction, metabolism, liver physiology, and other
effects (Anbumani et al. 2018). In addition, ingested microplastics can also serve as a
vector for exposure to harmful chemicals including components of plastic formulations,
chemicals sorbed to plastic particles and pathogens (Anbumani et al. 2018; McCormick
et al. 2014).

In the Great Lakes, microplastics concentrations as high as 466,000 particles/km2 have
been observed (Eriksen et al. 2013). Although tributaries were assumed to be the major
source of microplastics, published characterization of microplastics in rivers was scarce.
The available studies provided a valuable foundation but lacked sufficient scope and
scale to provide insight into the influence of key watershed attributes and hydrology
(Wagner et al. 2014).

A.3.1 Objectives

The objectives of this study were to (1) determine occurrence and concentrations of
microplastics in Great Lakes tributaries, (2) determine relations between microplastics
and watershed attributes such as land cover, population density, and wastewater
effluent contribution, and (3) explore the role of hydrology on microplastic occurrence. A
summary of study findings is provided here; additional detail and supporting data are
published elsewhere (Baldwin et al. 2016c, 2016d).

A.3.2 Methods

Study sites included 29 Great Lakes tributaries in 6 states (Figure A.4). Watershed
drainage areas of the tributaries varied from 101 - 16,400 square kilometers (km2), with
2.9 - 92% urban land cover, and 0 - 44% wastewater effluent as a percentage of
streamflow.

A total of 107 samples were collected from April 2014 to April 2015. Each tributary was
sampled three or four times, capturing low-flow and runoff-event conditions. Samples
were collected using a 1.5 m long, 333 |jm neuston net with an opening 100 cm wide x
40 cm high (Sea-Gear Corp. Miami, Florida, USA). The net skimmed the surface and
upper 20-35 cm of the water column, keeping a portion of the net opening above water.
Samples were collected by boat, from a bridge, or by wading (Figure A.5A-C).

Samples were processed and sieved into 3 size classifications (Baldwin et al. 2016b).
The sample from each sieve class was visually observed using a dissection
microscope, and microplastic particles were thereby enumerated and categorized
according to morphology as: fragments (broken down pieces of larger debris such as
plastic bottles), pellets/beads (preproduction pellets, microbeads from personal care
products and bead blasting, and other spheroids), lines/fibers (particles of fishing line
and nets, and fibers from synthetic textiles), films (plastic bags and wrappers), or foams
(foam cups, take-out containers, packaging) (Figure A.5D,E). Plastic particle
concentrations were reported in particles per cubic meter (p/m3).

36


-------
Figure A.4. Sampling locations, watershed boundaries, and land cover for 29 tributaries of
the Great Lakes. Map comprised of various spatial datasets (Instituto Nacional de
Estadistica Geografia e Informatica et al. 2006a, 2006b; National Atlas of the United States,
2005; U.S. Department of Agriculture-Natural Resources Conservation Service et al. 2009).

Concentration differences between nonurban low-flow, nonurban runoff-event, urban
low-flow, and urban runoff-event samples were evaluated, with urban samples defined
as those from watersheds with greater than 15% urban land cover.

Land cover

Lake Superior

Open water
~ Developed, open and low intensity
| Developed, medium and high intensity
I Forest, shrubland, herbaceous, and barren
] Planted/cultivated
| | Wetland

	 National boundary

	State boundary

	 Watershed boundary

O Stream sampling sites

>t" Louis'
Nemadji

initowoc

Ontario

Sheboygan

Tonawand
Buffalo*

Grand,

Milwaukee

Kalamazoo'

Lake!

Ashtabula

Grand

Paw Pawrt

\ St Joseph
Indiana Harbor Canal^ -Ryrns i

Huron
bRaisin

Maumee<

luyahoga

Huron

Sandusky

Portage'

Kilometers

37


-------
Figure A.5. Sample collection using (A) a bridge crane and (B) by wading; (C), washing
particles from the net into the cod end using a backpack sprayer; (D, E), microscopic
images of assorted microplastic particles.

A.3.3 Key findings

Plastic particles were found in all 107 samples analyzed (complete sample results
published previously, Baldwin et al. 2016c). Sample concentrations ranged from 0.05 -
32 p/m3 (median 1.9 p/m3 mean 4.2 p/m3). Seventy-two percent of particles were in the
smallest size range sampled (0.355-0.99 mm), 26% were in the 1.0-4.75 mm size
range, and 2% were > 4.75 mm. Seventy-one percent of plastic particle types were
lines/fibers (mostly fibers), 17% were fragments, and the remaining particles were
foams, films, and pellets/beads, accounting for 8%, 3%, and 2% of all particles,
respectively.

Concentrations of fragments, pellets/beads, films, and foams were positively correlated
with watershed attributes related to urban development, including total urban land cover
(Figure A.6), population density, and (films excepted) percent impervious cover.
Hydrology also appeared to influence concentrations of these particle types: in urban
and nonurban watersheds, concentrations of fragments, films, and foams were greater
during runoff-events than during low-flow conditions when normalized by mean
concentration for the sampling site (Figure A.7). These litter-related plastics can be
transported efficiently in urban conveyance systems from impervious areas to receiving
waters.

38


-------
Fibers/lines were ubiquitous across all land use types (Figure A.6); concentrations were
not correlated with any of the tested watershed attributes, nor were they affected by
hydrology (Figure A.7). Recent research indicates that atmospheric deposition may be
one important source of fibers (Dris et al. 2016) that could subsequently enter streams
via direct deposition or washoff from surfaces throughout the watershed. None of the
plastic types were significantly correlated with the contribution of wastewater effluent to
streamflow; however, land application of WWTP sludge may be a significant source of
fibers in agricultural areas (Kang et al. 2018). Further work is necessary to
comprehensively evaluate sources of fibers to streams.

Given the abundance of microplastics of less than 333 |jm reported previously (Dris et
al. 2015; OSPAR, 2009), the current study likely underrepresents true microplastic
concentrations. This can be ecologically significant since such particles can be taken up
into cells and can translocate from the gut into the circulatory system (Browne et al.
2008), and their larger surface area to volume ratio enhances potential as vectors for
sorbed contaminants.

The relative proportion of particle types in the current tributary study differs greatly from
previous findings in the Great Lakes themselves. Most notably, the difference in the
proportion of fibers/lines in the current study (71%) is much greater than the proportion
of fibers in surface samples from the lakes (up to 14%). Hydraulics within the river
systems as compared to the Great Lakes together with the physical properties of the
plastics may explain this difference in abundance of fibers. Negatively-buoyant fibers
made of polymers such as polyester, rayon, nylon, and cellulose acetate may remain in
suspension in the turbulent flow of a river (allowing them to be captured by surface
sampling), but likely settle out upon reaching the more quiescent lakes. Accumulations
in sediment may have important effects on benthic organisms, as well as higher trophic
level organisms reliant on these benthic organisms.

This study provides an important baseline for future studies. The number and diversity
of sampling locations, the regional scale, and the incorporation of varying hydrologic
conditions provided a multifaceted approach that allowed for the exploration of many
factors potentially influencing the prevalence of plastic debris in rivers. The results have
advanced our currently limited understanding of the sources, transport, and fate of
plastics in fluvial systems.

39


-------
A. all particle types

Sampling locations

Figure A.6. Average concentrations of plastic particles (A-F) and watershed land cover (G)
at sampled Great Lakes tributaries, 2014-15.

40


-------
0)

a>
E

|1

to °
0 Q-

O C/5

£ CD

o £

03
Q_

10

0.1

ND

-r-

it
o

-i—

it
o

foams

films

pellets/beads

fibers/lines

B,C

B

0

A,B AB __ o A A,B

i I iiil

AAA

-1—

it
o

-1—

£
o

-1—

it
o

it
o

—r~

it
o

—i—

it
o

—i—

it
o

~i—

o

-1—

it
o

T | T
^ " $
-2 § ©

5= C 5=

t

O

S " 5
° § °

5^5
° § °

— CD —

^ C q=
I

<:
o

i

$

— 03 —

Figure A.7. Plastic concentrations in nonurban low-flow (n = 40), nonurban runoff (n = 35),
urban low-flow (n = 17), and urban runoff (n = 15) samples. Urban watersheds are those
with greater than 15% urban land cover. Boxplot labels A, B, and C indicate which groups
of samples are statistically similar (those sharing a common letter) and statistically
different (those not sharing a common letter) using the Kruskal-Wallis multiple
comparisons test (p-values < 0.05). [boxes, 25th to 75th percentiles; dark line, median;
whiskers, 1.5 x the interquartile range (IQR); circles, values outside 1.5 x the IQR; ND, not
detected].

A.4. Microorganisms

Human and livestock waste are two substantial sources of contamination that enter
Great Lakes waterways through a variety of pathways. Human waste sources include
degraded public and private sanitary sewer lines and improper connections, sanitary
and combined sewer overflows, treated wastewater effluent, properly functioning and
defective septic systems, and land application of waste effluent. Livestock waste enter
waterways through direct access to streams, overland flow from barnyards, pastures,
and manure application, and through subsurface drain tiles. These waste streams serve
as substantial sources of waterborne pathogens and chemical contaminants to
waterways where they can pose risks to human and/or ecological health. However,
since their occurrence in wastes vary hydrologically, temporally, and seasonally (for
example, human viruses are present only when the host population is infected and
varies seasonally and temporally), these contaminants may not always be detectable
when human and agricultural wastewater contamination is present. Therefore, testing
for non-pathogenic, host-associated indicators can provide considerable value since
they are abundantly present in the original fecal sources and remain at detectable levels
even after substantial dilution in receiving waters (Lenaker et al. 2018).

A.4.1 Objectives

41


-------
Building upon the investigation of organic waste compounds reported in section 3.2, the
overall objective for the current study was to characterize variability of microbiological
contaminants and potential hazard from waterborne pathogens in tributaries of the
Great Lakes. Specific objectives were to 1) provide information on the prevalence of
human and livestock waste contamination in tributaries of the Great Lakes using host-
associated indicators, 2) investigate the occurrence of waterborne pathogens as a result
of human and livestock contamination, and 3) characterize variability of host-associated
indicators by hydrology (low-flow and periods of increased runoff), season, and
watershed attributes. The current section on microorganisms provides a summary of
findings from the first phase of GLRI; additional detail and supporting data are published
elsewhere (Corsi et al. 2018; Dila et al. 2018; Lenaker et al. 2017).

A.4.2 Methods

Study sites included eight Great Lakes tributaries selected to represent a gradient of
urban and agricultural land covers (Figure A.8). Flow-weighted composite water
samples were collected during low-flow and runoff event periods from February 2011 to
June 2013 and analyzed for waterborne pathogens (290 samples) and host-associated
bacteria (214 samples). Host-associated bacteria analyses included human Bacteroides
(HB), Lachnospiraceae (Lachno2, human associated) and ruminant Bacteroides
(BacR). Waterborne pathogen analyses included eight human viruses, eight bovine
viruses, two protozoa, and four bacteria. Detailed sampling, analytical methods,
resulting data, and data analysis methods have been previously described (Corsi et al.
2014; Dila et al. 2018; Lenaker et al. 2017).

42


-------
EXPLANATION
Land Cover
¦ Water

~ Developed (open and
low intensity)

¦ Developed (medium and
high intensity)

Forest, shrubland,
herbaceous, and barren
I I Planted/cultivated
r I Wetland
	State/province boundary

	Great Lakes watershed

boundary

	Site watershed boundary

A Sampling site

Base composited from North American Atlas
political boundaries, 1:10,000,000,2006; North
American Atlas hydrography, 1:10,000,000,2006;
U.S. Geological Survey hydrologic unit maps,
1:250,000,1994; Canada Land Inventory Level-I
watershed maps, 1:2,000,000; National Land Cover
Database 2011,30-m resolution, 2015. Michigan
GeoRef, referenced to North American Datum
of 1927.

Figure A.8. Sampling locations and land cover in eight tributaries of the Great Lakes. Map
comprised of various spatial datasets.

A.4.3 Key findings

Overall, five of the eight human viruses and four of the eight bovine viruses analyzed
were detected at least once. Human viruses (n=290) were present in 16% of samples,
and the human bacterial markers HB and Lachno2 (n=219) were present in 94% and
87% of samples respectively. Bovine viruses and pathogenic bacteria were present in
14% and 1.4% of samples (n=290), respectively, and the ruminant marker, BacR, was
present in 47% of samples (n=219). Protozoa were not detected during the study
period.

Evidence of human and bovine fecal pollution was present in all eight watersheds
(Figure A.9). Occurrence of human markers was generally higher in watersheds with the
most urban influence. Likewise, occurrence of bovine markers was generally highest in
watersheds with higher densities of cattle and pasture land. The Portage River was the

43


-------
most notable exception to human and bovine bacterial marker trends; the exact cause
of these differences is unknown but is thought to be attributable to a variety of factors
including those that have efficient conduits directly to the stream such as combined
sewer overflows or drain tiles. The comparatively lower bovine and human virus
occurrences at this site highlights the utility of using multiple parameters in evaluating
waste contamination.

Ill

Human

~ Bacteria
¦ Viruses





^ * Ss/S s



,>0 J

 Low Urban

Bovine/Ruminant

U.D.



s y ~ ~







& <^p

High -> Low Cattle x Pasture

Figure A.9. Occurrence of microbiological indicators of human and cattle waste in samples
collected at eight Great Lakes tributaries, 2011-2013. The sum of human bacteria markers
and the sum of human viruses are used as indicators of human waste (A), and ruminant
Bacteroides and the sum of bovine viruses are used as indicators of cattle waste (B).

Mean concentrations of human bacterial markers in urban and mixed land use
watersheds (Clinton River, Rouge River, Milwaukee River) were ~10 to 30-fold greater
during runoff-events compared with low-flow periods; however, mean concentrations in
agricultural watersheds did not differ with hydrologic condition (Figure A.10A). In
contrast, mean concentrations of BacR were greater during runoff-events (compared
with low-flow periods) across all sites (Figure A.10B); highest mean concentrations were
observed in runoff-event samples from the watershed with the highest cattle density
(Manitowoc River). Mean concentrations of human and bovine viruses were not
significantly different with respect to flow conditions at individual sampling locations nor
when data were grouped by land cover or cattle density characteristics (Figures A.10C
and D).

44


-------
Human

n = 18 12 21 12 13 11 12 12 18 12 17 10 17 10 14 10

Bovine and Ruminant

17 10 13 11 12 12 18 12 21 12 14 10 17 10 18 12

_l

106-

E



C )



o





10° -

CD



-Q



E

"D

104 -

:z:



>>



Q.
O

103 -

o





ND -

Event
Low-flow

.~

0

CO

1	30

O

o 10
O _j

7?)

CD
Q_

° 1 H
o ' "

o

E
o

5 0.1 H
e>

B

I

I.

T

I

ND J

oJ

^ c/V ^	^ ^ *

<# ^ aO° .	<& ,0

, Low Cattle x Pasture

High -> Low Urban

CD

i_

Q)
O
CD
CQ

n = 12 13 24 22 11 11 23 22 13 11 22 23 12 10 17 17	12 13 24 22 11 11 23 22 13 11 22 23 12 10 17 17

100

10

c/)
CD
cn
=j

c

CD
0

Figure A.10. Boxplots of host-associated bacteria and bar plots of mean virus
concentrations at eight Great Lakes watersheds from 2011-2013, categorized by hvdrologic
condition. Host-associated bacteria include the sum of human Bacteroiiles and
Lachnospiraceae (A) and ruminant Bacteroides (B). Viruses include mean of the sum of
human viruses (C) and mean of the sum of bovine viruses (D).

Human- and cattle-associated bacterial markers and viruses were all present to varying
degrees in all seasons, with greater concentrations observed during the cooler months
of December through April than other months (Figure A.11). This same pattern was also
true when considering runoff-event periods and low-flow periods separately with only
one exception: the mean sum of human viruses had no significant difference between
the two seasonal periods during runoff-events. This seasonal difference could be
influenced by multiple possible factors including: increased survival in cold weather
resulting from decreased sun exposure due to ice cover and shorter daylight periods
more efficient transport during periods with saturated soils, winter spreading of manure
on frozen ground, and less disinfection of treated wastewater compared to warm
weather periods.

Multiple regression analysis was used to explore factors that explain variability in host-
associated marker flux (marker quantity per unit time per unit watershed drainage area)

45


-------
for samples collected during rainfall periods. For each of the three host-associated
markers investigated, one regression equation representing all eight sites was effective
at describing variability of flux using four predictor variables: population density (human
for HB and Lachno2, and cattle for BacR) as an indicator of source, and season, rainfall
depth, and percent drain tile coverage which all influence hydrology (Dila et al. 2018).
Coefficients for the seasonal variables reflected a peak seasonal contribution in late
winter and early spring for human and ruminant indicators. This is consistent with direct
analysis of seasonal data described above. These cooler months are typically the time
of year in the Great Lakes region when the ground is saturated or frozen and low-flow
levels in streams are greatest, leading to efficient runoff mechanisms. Tile drainage also
increases efficiency of watershed hydraulics, which leads to efficiency in contaminant
transport, including microorganisms (Wang et al. 2010).

Figure A.ll. Boxplots of host-associated bacteria and bar plots of mean virus
concentrations at eight Great Lakes watersheds from 2011-2013, categorized by Seasonal
grouping and hydrologic condition. Host-associated bacteria include the sum of human
bacteria markers (A) and ruminant Bacteroides (B). Viruses include mean of the sum of
human viruses (C) and mean of the sum of bovine viruses (D).

46


-------
Collectively, three different classes of parameters were measured in the first phase of
GLRI that indicate presence of human waste: human-associated bacteria markers (Dila
et al. 2018), human viruses (Lenaker et al. 2017), and chemicals associated with human
waste, including 3-non-prescription drugs and 10-flavors and fragrances (Chapter 3,
section 3.2; Baldwin et al. 2016). The prevalence of each of these three classes of
human waste indicators is dependent on multiple factors that rely on source
concentration as well as fate and transport properties. Even with the inherent
differences in these factors, occurrence and concentration of human viruses and
human-waste chemicals both increased with increasing human bacterial marker
concentration (Figure A. 12). Together, these results indicate that the human-associated
bacteria markers can be used as indicators of a potential health hazard from waterborne
pathogens and presence of toxic chemicals associated with human and cattle waste.
Given the relatively low cost and ubiquitous presence in the host, host-associated
bacterial markers are a reasonable choice for screening for the presence and
magnitude of human and cattle waste in surface water.

1.0 -i

o) 0.5 -

0.0

Cy

Wastewater Compounds


o

0.0 C
a>

1.0 H

0.5

0.0

0-225 225- 10J 10 -104

>10=

Sum of Human Bacteria (Copy Number/100 ml_)

Figure A.12. Comparison of human-associated bacteria markers with the concentration
and occurrence frequency of human wastewater associated compounds and human virus
samples collected from eight Great Lakes Tributaries, 2011-2013.

47


-------
A.5. Knowledge gaps

•	To fully understand potential for biological effects of CECs, there is a need to
expand this assessment from the limited number of compounds in the current
evaluation to a broader suite of compounds in specific chemical classes including
parent and degradation products.

•	Results indicated that a number of compounds at numerous sites occur at
concentrations exceeding water quality benchmarks and exceeding levels at
which EEQs would indicate potential for reproductive issues. Even so, water
quality benchmarks and estradiol equivalency information are available for fewer
than half of the sampled compounds. There is a need to expand the evaluation of
potential biological effects to include a larger proportion of the monitored
compounds. Evaluation that leverages the growing database of chemical-specific
high throughput in vitro biological activity data being generated via the ToxCast
program (https://www.epa.gov/chemical-research/toxicity-forecasting) has been
initiated for this purpose (see Appendix E).

•	There is a need to expand the evaluation of potential biological effects to include
consideration of monitoring results with chemical mixtures. Use of information
from the ToxCast database to estimate cumulative biological activity for
chemicals that influence common biological pathways has been initiated for this
purpose (see Appendix E).

•	Comparison of chemical monitoring results to water quality benchmarks provides
information on potential for adverse effects but does not indicate hazard for
specific biological functions. To give resource managers confidence in
monitoring-based evaluation results, the findings must be validated. To provide
enough information to design experiments for validating evaluations based on
chemical monitoring, it is important to identify the specific biological pathways of
concern. One way to achieve this would be to link results from a ToxCast-based
evaluation with information captured in the adverse outcome pathway wiki
(Society for the Advancement of Adverse Outcome Pathways, 2018).

•	Objectives of the continued research on this program are to take advantage of
information in ToxCast to prioritize chemicals, identify sites at which these
chemicals occur, and screen for potential adverse biological effects that may be
associated with those chemicals. To achieve these objectives, software tools are
needed to efficiently facilitate this type of evaluation of complex multidimensional
data sets containing multiple samples per site and numerous chemicals per
sample at many sites.

•	Short-term variability of CECs may play an important role in acute effects on
biological health, however most efforts for evaluation of CECs rely on a limited
number of observations at a low collection frequency. Additional data is needed
to define sub-daily variability in exposure concentrations (due to hydrologic
condition, etc.).

48


-------
•	The fate of plastic fibers once delivered from tributaries to the Great Lakes is in
need of investigation. From the current study, it was determined that the
distribution of different microplastic particle morphologies was different in
tributaries compared to the Great Lakes due to the large proportion of fibers/lines
in tributary samples. To investigate whether deposition of negatively-buoyant
fibers contributes to this result, a study of bed sediment samples from Lakes
Michigan and Erie has been undertaken.

•	Previous water sampling methods capture only particles on the surface.

Additional work is needed to quantify the abundance and morphologies of
plastics vertically throughout the water column to better characterize
microplastics in the aquatic environment.

•	The size range of particles captured is dependent on the mesh size of sampling
nets. Additional work is needed to define particles in smaller size ranges that
have greater likelihood of uptake and potential adverse impact on organisms in
the lower trophic levels.

•	Further work is necessary to thoroughly evaluate sources of microplastics to
receiving waters and potential for management options.

•	A relation with human-associated bacteria markers and wastewater-associated
chemicals was established. Current methods to evaluate the many wastewater
compounds with potential adverse ecological effects are expensive and time
consuming. There is a need for an efficient and cost-effective method to screen
for potential biological effects for which host-specific markers may play an
important role.

•	There is a critical need for a more comprehensive characterization of the fate and
transport characteristics of contaminants and pathogens moving from tributaries
into the nearshore zone of the Great Lakes, in order to more accurately evaluate
their potential ecological and human health consequences from exposure during
recreational activities.

A.6. Management Implications

Resulting information concerning OWCs, microplastics, and microorganisms provide
characterization of these contaminants that can be used to help identify the most likely
sources and scenarios for which they are most prevalent. Land use, hydrologic, and
seasonal associations with specific contaminant classes can be used to evaluate the
practicality of designing management scenarios for controlling specific contaminants
that have been identified to be of greatest potential concern:

Urban land use:

•	The greatest concentrations of PAHs were present in urban watersheds and did
not vary with season. PAH contamination can be introduced from individual point

49


-------
sources such as coal-fired power plants or other industrial activities or from more
diffuse non-point sources such as coal-tar pavement sealcoat or vehicle exhaust.
Remedies for individual sources must be considered on a case-by-case basis,
but control of diffuse sources would require community-level considerations or by
implementation of urban stormwater runoff management and green infrastructure
practices.

•	Human-associated bacterial marker concentrations and flux increased as urban
influence in the watershed increased. They were also greater during cool
weather months and during periods of increased runoff. Illicit discharge detection
and elimination programs (IDDE) in the Great Lakes region are likely to
underestimate the severity of sewage contamination because the most common
activity periods for IDDE programs include dry weather periods during the ice-
free months. Shifting a portion of IDDE efforts to cool weather months and during
runoff periods would increase the likelihood of capturing the periods of greatest
contamination.

•	Insecticides, plasticizers, antioxidants, detergent metabolites, fire retardants,
nonprescription drugs, sterols, flavors/fragrances, and dyes/pigments were all
greatest in watersheds with the most urban influence. Control options for these
contaminants include IDDE programs for those originating from sewage, or urban
stormwater runoff management practices for other sources. Urban stormwater
runoff practices are variably effective for these classes of contaminants given
that many of them are in solution, and many management practices rely on
particulate deposition.

Agricultural land use:

•	Herbicides (Atrazine and metolachlor) were identified as a high priority with the
greatest concentrations coinciding with application periods during late spring and
early summer. Strategies to reduce the potential biological impact of pesticides
would be most effective if designed to reduce pesticide runoff during application
with consideration of the variable hydrologic and vegetation cover conditions
during this elevated concentration period.

•	Concentrations and flux of cattle-associated microorganisms increased with
cattle density and percent pasture in the watershed. They were also greatest
during winter and early spring, consistent with other common agricultural
pollutants such as nutrients and sediment. Multiple options for runoff
management are available for these land uses, and study results indicated that
there is potential for substantial reductions if chosen agricultural management
practices are implemented to be effective during these cool weather months.

Microplastics:

•	Concentrations of most types of microplastics increased with increasing urban
influence in the watershed but were present in nonurban watersheds as well.

50


-------
Contamination of several morphologies of microplastics were more prevalent
during runoff periods in urban and nonurban watersheds. This indicates that
agricultural and urban runoff control measures that rely on filtration or infiltration
would likely be effective for removal of many microplastics, but microplastics with
positive buoyancy may not be captured in control measures that rely on particle
deposition.

• Microplastics have not been studied sufficiently within watersheds to understand
sources well. Control measures would have a greater likelihood of success if
watershed management efforts first focused on source area identification efforts
and developed estimations of the relative quantity of microplastics originating
from specific sources.

Modeling results for host-associated bacterial markers can enhance watershed
management activities: One regression model per host-associated bacterial marker was
effective at describing variability in all eight watersheds. This result indicates that there
is potential for transferability of this model to additional watersheds to evaluate human
and cattle waste contamination for watershed management activities such as estimation
of total maximum daily loads.

A.7. Acknowledgements

The authors gratefully acknowledge the many individuals at the USGS involved in
sample collection from the Water Science Centers in Minnesota, Wisconsin, Indiana,
Michigan, Ohio, and New York. We also thank SUNY Fredonia students Rachel Ricotta,
Joylyn Kovachev, Katie Donnelly, and Evan Miller for the many hours spent analyzing
microplastics samples in the laboratory, the U.S. Geological Survey National Water
Quality Laboratory for providing sample analysis and technical input, the U.S.
Department of Agriculture/U.S. Geological Survey Laboratory for Infectious Diseases in
the Environment and the University of Wisconsin-Milwaukee School of Freshwater
Sciences McLellan Laboratory for microorganism analysis. Support for this project was
provided by the Great Lakes Restoration Initiative through the U.S. Environmental
Protection Agency's Great Lakes National Program Office, the Graham Sustainability
Institute Water Center at University of Michigan, and National Institutes of Health. Any
use of trade, product, or firm names is for descriptive purposes only and does not imply
endorsement by the U.S. Government.

A.8. Products

Baldwin, A. K. Corsi, S. R. De Cicco, L. A. Lenaker, P. L. Lutz, M. A. Sullivan, D. J. &
Richards, K. D. (2016b). Organic contaminants in Great Lakes tributaries: Prevalence
and potential aquatic toxicity. Science of The Total Environment, 554-555, 42-52.
https://doi.Org/10.1016/j.scitotenv.2016.02.137

Baldwin, A. K. Corsi, S. R. & Mason, S. A. (2016c). Microplastics in 29 Great Lakes
tributaries (2014-15): U.S. Geological Survey data release.
http://dx.doi.org/10.5066/F7ZC80ZP

51


-------
Baldwin, A. K. Corsi, S. R. & Mason, S. A. (2016d). Plastic Debris in 29 Great Lakes
Tributaries: Relations to Watershed Attributes and Hydrology. Environmental Science &
Technology, 50(19), 10377-10385. https://doi.org/10.1021/acs.est.6b02917

Corsi, S. R. Dila, D. K. Lenaker, P. L. Baldwin, A. K. Bootsma, M. J. & McLellan, S. L.
(2018). Regression models and associated data for describing variability of host specific
bacteria fluxes in eight Great Lakes tributaries, 2011-2013 [Data set], U.S. Geological
Survey. https://doi.Org/10.5066/F7VX0DRH

Dila, D. K. Corsi, S. R. Lenaker, P. L. Baldwin, A. K. Bootsma, M. J. & McLellan, S. L.
(2018). Patterns of Host-Associated Fecal Indicators Driven by Hydrology, Precipitation,
and Land Use Attributes in Great Lakes Watersheds. Environmental Science &
Technology. https://doi.Org/10.1021/acs.est.8b01945

Lenaker, P. L. Corsi, S. R. Borchardt, M. A. Spencer, S. K. Baldwin, A. K. & Lutz, M. A.
(2017). Hydrologic, land cover, and seasonal patterns of waterborne pathogens in Great
Lakes tributaries. Water Research, 113, 11-21.
https://doi.Org/10.1016/j.watres.2017.01.060

A.9. References

Anbumani, S. & Kakkar, P. (2018). Ecotoxicological effects of microplastics on biota: a
review. Environmental Science and Pollution Research, 25(15), 14373-14396.
https://doi.Org/10.1007/sl 1356-018-1999-x

Baldwin, A. K. Corsi, S. R. De Cicco, L. A. Lenaker, P. L. Lutz, M. A. Sullivan, D. J. &
Richards, K. D. (2016a). Organic contaminants in Great Lakes tributaries:

Prevalence and potential aquatic toxicity. Science of The Total Environment, 554-
555, 42-52. https://doi.Org/10.1016/j.scitotenv.2016.02.137

Baldwin, A. K. Corsi, S. R. De Cicco, L. A. Lenaker, P. L. Lutz, M. A. Sullivan, D. J. &
Richards, K. D. (2016b). Organic contaminants in Great Lakes tributaries:

Prevalence and potential aquatic toxicity. Science of The Total Environment, 554-
555, 42-52. https://doi.Org/10.1016/j.scitotenv.2016.02.137

Baldwin, A. K. Corsi, S. R. & Mason, S. A. (2016c). Microplastics in 29 Great Lakes
tributaries (2014-15): U.S. Geological Survey data release.
http://dx.doi.org/10.5066/F7ZC80ZP

Baldwin, A. K. Corsi, S. R. & Mason, S. A. (2016d). Plastic Debris in 29 Great Lakes
Tributaries: Relations to Watershed Attributes and Hydrology. Environmental
Science & Technology, 50(19), 10377-10385.
https://doi.Org/10.1021 /acs.est.6b02917

Barber, L. B. Loyo-Rosales, J. E. Rice, C. P. Minarik, T. A. & Oskouie, A. K. (2015).
Endocrine disrupting alkylphenolic chemicals and other contaminants in wastewater
treatment plant effluents, urban streams, and fish in the Great Lakes and Upper

52


-------
Mississippi River Regions. Science of The Total Environment, 517, 195-206.
https://doi.Org/10.1016/j.scitotenv.2015.02.035

Bell, A. H. Coles, J. F. McMahon, G. & Woodside, M. D. (2012). Urban development
results in stressors that degrade stream ecosystems (USGS Numbered Series No.
2012-3071). Reston, VA: U.S. Geological Survey. Retrieved from
http://pubs.er.usgs.gov/publication/fs20123071

Bennie, D. T. Sullivan, C. A. Lee, H.-B. Peart, T. E. & Maguire, R. J. (1997). Occurrence
of alkylphenols and alkylphenol mono- and diethoxylates in natural waters of the
Laurentian Great Lakes basin and the upper St. Lawrence River. Science of The
Total Environment, 193(3), 263-275. https://d0i.0rg/l 0.1016/S0048-9697(96)05386-
7

Blair, B. D. Crago, J. P. Hedman, C. J. Treguer, R. J. F. Magruder, C. Royer, L. S. &
Klaper, R. D. (2013). Evaluation of a model for the removal of pharmaceuticals,
personal care products, and hormones from wastewater. Science of The Total
Environment, 444, 515-521. https://doi.Org/10.1016/j.scitotenv.2012.11.103

Browne, M. A. Dissanayake, A. Galloway, T. S. Lowe, D. M. & Thompson, R. C. (2008).
Ingested microscopic plastic translocates to the circulatory system of the mussel,
Mytilus edulis (L). Environmental Science & Technology, 42(13), 5026-5031.

Collier, T. K. Anulacion, B. F. Arkoosh, M. R. Dietrich, J. P. Incardona, J. P. Johnson, L.
L. ... Myers, M. S. (2013). 4 - Effects on Fish of Polycyclic Aromatic Hydrocarbons
(PAHS) and Naphthenic Acid Exposures. In A. P. F. and C. J. B. Keith B. Tierney
(Ed.), Fish Physiology {Vol. 33, pp. 195-255). Academic Press. Retrieved from
http://www.sciencedirect.com/science/article/pii/B9780123982544000042

Corsi, S. R. Borchardt, M. A. Spencer, S. K. Hughes, P. E. & Baldwin, A. K. (2014).
Human and bovine viruses in the Milwaukee River watershed: Hydrologically
relevant representation and relations with environmental variables. Science of The
Total Environment, 490, 849-860. https://doi.Org/10.1016/j.scitotenv.2014.05.072

Corsi, S. R. Dila, D. K. Lenaker, P. L. Baldwin, A. K. Bootsma, M. J. & McLellan, S. L.
(2018). Regression models and associated data for describing variability of host
specific bacteria fluxes in eight Great Lakes tributaries, 2011-2013 [Data set], U.S.
Geological Survey. https://doi.org/10.5066/F7VX0DRH

Dila, D. K. Corsi, S. R. Lenaker, P. L. Baldwin, A. K. Bootsma, M. J. & McLellan, S. L.
(2018). Patterns of Host-Associated Fecal Indicators Driven by Hydrology,
Precipitation, and Land Use Attributes in Great Lakes Watersheds. Environmental
Science & Technology, https://doi.org/10.1021/acs.est.8b01945

Dris, R. Gasperi, J. Rocher, V. Saad, M. Renault, N. &Tassin, B. (2015). Microplastic
contamination in an urban area: a case study in Greater Paris. Environmental
Chemistry, 12(5), 592-599.

53


-------
Dris, R. Gasperi, J. Saad, M. Mirande, C. & Tassin, B. (2016). Synthetic fibers in
atmospheric fallout: A source of microplastics in the environment? Marine Pollution
Bulletin. https://doi.Org/10.1016/j.marpolbul.2016.01.006

Eriksen, M. Mason, S. Wilson, S. Box, C. Zellers, A. Edwards, W. ... Amato, S. (2013).
Microplastic pollution in the surface waters of the Laurentian Great Lakes. Marine
Pollution Bulletin, 77(1-2), 177-182. https://doi.Org/10.1016/j.marpolbul.2013.10.007

Hinck, J. E. Schmitt, C. J. Chojnacki, K. A. & Tillitt, D. E. (2009). Environmental
contaminants in freshwater fish and their risk to piscivorous wildlife based on a
national monitoring program. Environmental Monitoring and Assessment, 752(1-4),
469-494. https://doi.Org/10.1007/sl 0661 -008-0331 -5

Ingersoll, C. G. MacDonald, D. D. Brumbaugh, W. G. Johnson, B. T. Kemble, N. E.
Kunz, J. L. ... Ireland, D. S. (2002). Toxicity Assessment of Sediments from the
Grand Calumet River and Indiana Harbor Canal in Northwestern Indiana, USA.
Archives of Environmental Contamination and Toxicology, 43(2), 156-167.
https://doi.Org/10.1007/S00244-001 -0051 -0

Instituto Nacional de Estadfstica Geograffa e Informatica, The Atlas of Canada, & U.S.
Geological Survey. (2006a). North American Atlas - Political Boundaries (Vector
digital data). Reston, Virginia, USA: U.S. Geological Survey. Retrieved from
https://nationalmap.gov/small_scale/mld/boundOm.html

Instituto Nacional de Estadfstica Geograffa e Informatica, The Atlas of Canada, & U.S.
Geological Survey. (2006b). North American Atlas - Political Boundaries:
Government of Canada, Ottawa, Ontario, Canada. Instituto Nacional de Estadfstica,
Geograffa e Informatica, Aguascalientes, Aguascalientes, Mexico;U.S. Geological
Survey, Reston, Virginia, USA. Retrieved from http://nationalatlas.gov/atlasftp-
na.html

Ismail, N. S. Muller, C. E. Morgan, R. R. & Luthy, R. G. (2014). Uptake of contaminants
of emerging concern by the bivalves anodonta californiensis and corbicula fluminea.
Environmental Science and Technology, 48(16), 9211-9219.
https://doi.Org/10.1021 /es5011576

Jenkins, J. A. Olivier, H. M. Draugelis-Dale, R. O. Eilts, B. E. Torres, L. Patino, R. ...
Goodbred, S. L. (2014). Assessing reproductive and endocrine parameters in male
largescale suckers (Catostomus macrocheilus) along a contaminant gradient in the
lower Columbia River, USA. Science of the Total Environment, 484(1), 365-378.
https://doi.Org/10.1016/j.scitotenv.2013.09.097

Johnson, L. L. Anulacion, B. F. Arkoosh, M. R. Burrows, D. G. da Silva, D. A. M.

Dietrich, J. P. ... Ylitalo, G. M. (2013). 2 - Effects of Legacy Persistent Organic
Pollutants (POPs) in Fish—Current and Future Challenges. In A. P. F. and C. J. B.
Keith B. Tierney (Ed.), Fish Physiology (Vol. Volume 33, pp. 53-140). Academic
Press. Retrieved from

http://www.sciencedirect.com/science/article/pii/B9780123982544000029

54


-------
Kang, H.-J. Park, H.-J. Kwon, O.-K. Lee, W.-S. Jeong, D.-H. Ju, B.-K. & Kwon, J.-H.
(2018). Occurrence of microplastics in municipal sewage treatment plants: a review.
Environmental Health and Toxicology, 33(3), e2018013.
https://doi.Org/10.5620/eht.e2018013

Kingsbury, J. A. Delzer, G. C. & Hopple, J. A. (2008). Anthropogenic Organic

Compounds in Source Water of Nine Community Water Systems that Withdraw from
Streams, 2002-05 (No. SIR-2008-5208). United States Geological Survey. Retrieved
from http://pubs.usgs.gov/sir/2008/5208/

Kolpin, D. W. Furlong, E. T. Meyer, M. T. Thurman, E. M. Zaugg, S. D. Barber, L. B. &
Buxton, H. T. (2002). Pharmaceuticals, Hormones, and Other Organic Wastewater
Contaminants in U.S. Streams, 1999-2000: A National Reconnaissance.
Environmental Science & Technology, 36(6), 1202-1211.
https://doi.Org/10.1021 /es011055j

Lenaker, P. L. Corsi, S. R. Borchardt, M. A. Spencer, S. K. Baldwin, A. K. & Lutz, M. A.
(2017). Hydrologic, land cover, and seasonal patterns of waterborne pathogens in
Great Lakes tributaries. Water Research, 113, 11-21.
https://doi.Org/10.1016/j.watres.2017.01.060

Lenaker, P. L. Corsi, S. R. McLellan, S. L. Borchardt, M. A. Olds, H. T. Dila, D. K. ...
Baldwin, A. K. (2018). Human-Associated Indicator Bacteria and Human-Specific
Viruses in Surface Water: A Spatial Assessment with Implications on Fate and
Transport. Environmental Science & Technology.
https://doi.org/10.1021/acs.est.8b03481

McCormick, A. Hoellein, T. J. Mason, S. A. Schluep, J. & Kelly, J. J. (2014). Microplastic
is an abundant and distinct microbial habitat in an urban river. Environmental
Science and Technology, 48(20), 11863-11871. https://doi.org/10.1021/es503610r

National Atlas of the United States. (2005). National Atlas of the United States, County
Boundaries of the United States, 2001 (National Atlas of the United States). Reston,
VA: National Atlas of the United States. Retrieved from
http://nationalatlas.gov/atlasftp. htm I

OSPAR. (2009). Marine litter in the North-East Atlantic Region: Assessment and
priorities for response. London, United Kingdom.

Society for the Advancement of Adverse Outcome Pathways. (2018). Aopwiki.

Retrieved September 21, 2018, from https://aopwiki.org/

Stackelberg, P. E. Furlong, E. T. Meyer, M. T. Zaugg, S. D. Henderson, A. K. &
Reissman, D. B. (2004). Persistence of pharmaceutical compounds and other
organic wastewater contaminants in a conventional drinking-water-treatment plant.
Science of The Total Environment, 329(1-3), 99-113.
https://doi.Org/10.1016/j.scitotenv.2004.03.015

55


-------
The Endocrine Disruption Exchange, Inc. (2012). TEDX list of potential endocrine
disruptors. Retrieved February 28, 2012, from

http://www.endocrinedisruption.com/endocrine.TEDXList.overview.php.

U.S. Department of Agriculture-Natural Resources Conservation Service, U. S.

Geological Survey, & U.S Environmental Protection Agency. (2009). The Watershed
Boundary Dataset (WBD) (Vector digital data). Fort Worth, Texas: U.S. Department
of Agriculture, Natural Resources Conservation Service, National Cartography and
Geospatial Center. Retrieved from http://datagateway.nrcs.usda.gov

Vajda, A. M. Barber, L. B. Gray, J. L. Lopez, E. M. Woodling, J. D. & Norris, D. 0.
(2008). Reproductive disruption in fish downstream from an estrogenic wastewater
effluent. Environmental Science and Technology, 42(9), 3407-3414.
https://doi.org/10.1021/es0720661

Venier, M. Dove, A. Romanak, K. Backus, S. & Hites, R. (2014). Flame retardants and
legacy chemicals in Great Lakes' water. Environmental Science and Technology,
48(16), 9563-9572. https://doi.org/10.1021/es501509r

Wagner, M. Scherer, C. Alvarez-Munoz, D. Brennholt, N. Bourrain, X. Buchinger, S. ...
Reifferscheid, G. (2014). Microplastics in freshwater ecosystems: what we know and
what we need to know. Environmental Sciences Europe, 26(1), 12.
https://doi.Org/10.1186/sl 2302-014-0012-7

Wang, D. & Cai, X. (2010). Comparative study of climate and human impacts on
seasonal baseflow in urban and agricultural watersheds - Wang - 2010 -
Geophysical Research Letters - Wiley Online Library, 37.
https://doi.org/doi: 10.1029/2009GL041879

Zaugg, S. D. Smith, S. G. & Schroeder, M. P. (2006). Determination of Wastewater
Compounds in Whole Water by Continuous Liquid-Liquid Extraction and Capillary-
Column Gas Chromatography/Mass Spectrometry (No. TM-5-B4). United States
Geological Survey. Retrieved from http://pubs.usgs.gov/tm/2006/05B04/

56


-------
Appendix B

Monitoring of Contaminants of Emerging
Concern by Great Lakes Mussel Watch

Lead Organization: National Oceanic and Atmospheric Administration, National
Centers for Coastal Ocean Science, Silver Spring, MD, USA

Contributing Authors: Ed Johnson, Kimani Kimbrough, Erik Davenport, Michael
Edwards and Annie Jacob

Contributing Investigators: Jaruga Pawel, Dan Bearden, Tracey Schock and Amy
Ringwood

Corresponding Author Contact: ed.johnson@noaa.gov
B.1 Introduction

NOAA's Mussel Watch Program (MWP), a longstanding national contaminant
monitoring program that utilizes resident bivalves as bioindicators of water quality,
launched its monitoring activities in the Great Lakes in 1992. The then recently
established non-indigenous population of Ponto-Caspian dreissenid mussels in the
Great Lakes (except Lake Superior) with attributes such as high filtering capacity, ability
to bioaccumulate chemical contaminants with limited ability to metabolize them,
sedentary habits and widespread distribution on hard substrates was identified as a
prospective tool for contaminant monitoring. The slew of research on dreissenid
mussels, following their introduction in the Great Lakes, revealed their role in
contaminant cycling via trophic transfer from the base of the food web to top predators,
thus bolstering the value of using dreissenid mussels as a bioindicator by MWP. The
program established 23 long-term monitoring sites within the Great Lakes region at near
shore sites away from known outfalls and hotspots mirroring the national program
directive to sample from areas intended to represent general conditions of broad coastal
areas for water quality assessment. MWP monitored approximately 150 chemicals,
including trace metals and persistent, bioaccumulative and toxic chemicals, the so-
called "legacy contaminants" that were prevalent during the industrial revolution and
have been banned since the 1970s following the implementation of many environmental
regulations.

With increased attention on the much larger group of chemicals that remain unregulated
and unmonitored in the aquatic environment known as the contaminants of emerging
concern (CEC), MWP began assessing how best to incorporate these new classes of
chemicals in its monitoring protocol in the early 2000s. Polybrominated diphenyl ethers
(PBDE), a group of flame retardant chemicals, was the first CEC to be added to the
monitoring list in 2004. MWP conducted pilot CEC studies in the Chesapeake Bay,
Biscayne Bay, and the Gulf of Farallones in 2006 and in the Southern California Bight

57


-------
led by the Southern California Coastal Water Research Project in 2011. Concurrently,
under the Phase 1 Action Plan of the Great Lakes Restoration Initiative, MWP
expanded, adding sites in Areas of Concern (AOC) in 2009/2010 (Kimbrough et al.,
2014) and conducting place-based contamination assessments by adopting new
approaches and techniques including use of caged mussels and effects-based tools.
Mussels can be caged and strategically relocated for precise place-based assessments
(monitoring along a pollution gradient, pre and post remediation/restoration
assessments and/or contaminant source tracking) for both legacy and CEC monitoring.

Though the focus of Phase 1 monitoring was to provide data for historic contamination,
MWP obtained limited tissue data for CECs by coupling field efforts with retrospective
analyses of tissue samples. The main objectives of this opportunistic CEC monitoring
during Phasel was to determine the feasibility of using dreissenid mussels for CEC
monitoring and explore the use of effects-based monitoring tools in mussels. MWP
intended this Phase 1 subsidiary activity to be a testing ground before committing to full-
fledged CEC work in Phase 2 and to answer two fundamental questions: 1) What is the
occurrence, frequency and spatial distribution of CECs in dreissenid mussel tissue? 2)
Can we identify mussel health metrics to link exposure to biological effects through
collaborative partnership with academia and federal partners?

B.2 Methods

Dreissenid mussel samples from basin-wide monitoring (2009/2010), place-based
contamination assessments in specific AOCs (2011-2014), and offshore sampling
(obtained in partnership with EPA's Great Lakes Fish Monitoring and Surveillance
Program; 2012-2014) were chosen for CEC analysis. Given the natural variability and
heterogeneity in chemical contamination sources and processes in different zones of
the lake, mussel sampling sites were categorized as 1) Harbor-River-Tributary sites 2)
Nearshore sites and 3) Offshore sites (Figure B.1A). Samples are a combination of in
situ mussels collected from hard substrates and caged mussels deployed in rivers and
tributaries. The CECs included in this report are polycyclic aromatic hydrocarbons
(PAH), polybrominated diphenyl ethers (PBDE), pharmaceutical and personal care
products (PPCPs) and alkylphenols (AP). A comprehensive report of CECs in mussels
from the Great Lakes collected between 2013- 2016 can be found in Kimbrough et al.
2018. To increase the likelihood of finding CECs, samples from rivers and harbors
(mainly Milwaukee Estuary AOC and Niagara River AOC (Figure B.1B-C)) were
selected preferentially over samples collected from relatively less polluted nearshore
and offshore sites. The methods of in situ mussel collection, caged mussel deployment
and collection, and analytical methods for CECs and effects-based tools can be found in
detail in the Quality Assurance Project Plan (QAPP available at
https://coastalscience.noaa.gov/proiect/great-lakes-mussel-watch-supports-presidents-
great-lakes-restoration-initiative/).

58


-------
1A.

Lake Superior

[LMarys River

Minnesota Poi

lanistique Rivet

Rocky IslancTl
Menominee River
Bayshore Park£
Green Bay/Fox River *=

BockportiLhiFEj # ~

Lincoln Bawftgfr
Thunder BayA Lake Huron

.awrence

/	Sand

\Saginaw River & Batff	\

Muskegon	A' \

^¦Breakwater Black River Canal
T Muskegon St. Clair River^
A Lake Clinton River t
Hq/lland	Rouge River^i/

Breakwater Detroit River/T
River Raisin
Mauftiee River /VA

Sheboygan River|

Eighteenmih
Niagara ^

North Chicago^ J
I Hammod Marina -
Calumet Breakwater*-

Canada

Monitoring Sites

#	Harbor - River - Tributary Sites
~ Nearshore Sites

#	Offshore Sites
C3 Great Lakes States

	 	

Figure B.l: The general location of mussel sampling sites in the Great Lakes for
characterization of CECs from 2009- 2014. Most locations have 1-3 sites with the exception
of Milwaukee Estuary (13 sites; Fig IB) and Niagara River AOC (28 sites; Fig 1C). PAHs
were monitored at all sites and PBDEs at a smaller subset of sites between 2009- 2014.
PPCPs were retrospectively analyzed in mussel tissue samples collected from Milwaukee
Estuary in 2013, Niagara River, Presque Isle Bay, Ashtabula River, Cuyahoga River and
Black River AOCs along eastern shore of Lake Erie, and Thunder Bay National Marine
Sanctuary in Lake Huron in 2014.

59


-------
All chemical concentrations in mussel tissue were blank corrected and values below
detection limit were assigned zero. PAH concentration is reported as the sum of 54
compounds, PBDE as the sum of 52 congeners and both PAH and PBDE sums were
expressed in ng/g dry weight. PPCP concentrations are expressed in ng/g wet wt.
Hierarchical Ward's cluster analysis was used to cluster the PAH and PBDE
concentrations into high, medium and low concentrations. Non-parametric Wilcoxon text
was used to test significant differences in concentrations across zone types.

B.3 Results and discussion

B.3.1. Polycyclic Aromatic Hydrocarbons (PAHs)

PAHs have been monitored by MWP from the inception of the program in 1992 and its
national status and trends have been summarized for the nation (Kimbrough et al.
2008). As PAH compounds have current and ongoing sources and many compounds
are toxic, it is necessary to dissociate the negative impacts of PAH from those of
hydrophilic, less persistent CECs and hence are included in the CEC analyses.

Two hundred and twenty-one dreissenid tissue samples collected from around the
Great Lakes were analyzed for PAHs from 2009-2014. PAHs were ubiquitous in
distribution and the total concentration (sum of 54 compounds) ranged from 15.8-
112638.1 ng/g dry wt in mussels from 2009/2010 basin-wide monitoring (Figure B.2).

Total PAH Concentration (ng/g dry wt)

~ 15.8 - 1866.3
WM 1866 4-8420 7
M 8420 8 - 112638 1
Q3 Great Lakes States

60


-------
Figure B.2: The total PAH concentration (sum of 54 compounds; ng/g dry wt) in mussel
tissue from harbor-river-tributaries (O) and nearshore sites (A) collected in 2009/2010
and from offshore sites (O) sampled from 2010-2014 in partnership with EPA Great Lakes
Fish Monitoring and Surveillance Program.

Following this basin-wide monitoring effort, MWP conducted place-based contamination
assessments at several AOCs using both caged and in situ mussels. For example, the
PAH distribution in the main stem of the Niagara River and seven of its tributaries were
assessed to identify the contamination hotspots (Figure B.3) within the AOC boundary.

©Gill Creek

Q

af

q mCayuga Creek
"TJiagara-8

Niagara-9

Gill Creek

¦



Cayuga Creek





•

0;



Two Mile Creek

•



Smokes Creek



•

•

Niagara-6

o

gon

onawanda Creek

Two Mile Creek Ellicott Creek <

o •

Niagara-5

Tonawanda/Elliot Creek

O

Scajaquada Creek

M



NFTA Park/Times Beach

"	o f; ••

¦

v' *7

Total PAH Cone, (ng/g dry wt)
O 696.7 -9815.2

•	9815.3 -30098.9

•	30099.0 -49858.4

•	49858.5 -68168.4

^)Scajaquada Creek

ONiagara-4
Niagara-2

o

Niagara-1 Q ®Times Beach

Niagara-3 O ® NFTA Park

Smokes Creek <

61


-------
Figure B.3: PAH characterization using caged and in situ mussels at Niagara River AOC
conducted in 2014.

When all the PAH data were pooled and analyzed, we found that the total PAH
concentration from harbor-river-tributary sites were significantly higher than mussels
from offshore and nearshore sites (Wilcoxon test, p < 0.05; Figure B.4 excluding
extreme outliers). The harbor-river-tributary site samples were mainly from Milwaukee
Estuary and Niagara River AOCs. The large number of samples from different zone
types in the Great Lakes across years allows MWP to examine the pattern of PAH
composition and distribution in mussels (Kimbrough et al., In Preparation). Unlike
vertebrates, mussels possess only limited ability to metabolize PAHs and hence are
historically regarded as a good indicator for PAH monitoring.

40000-

"O

CJ)
c

c
o


-------
samples collected during basin-wide monitoring in 2009-2011 and the concentration
ranged from 3,2-126.2 ng/g dry wt (Figure B.5),

Figure B.5: The total PBDE concentration (sum of 51 congeners) in mussels from Harbor-
river- tributaries (O) and nearshore sites (A) collected from 2009-2011 and from offshore
sites (O) sampled from 2010-2014 in partnership with EPA Great Lakes Fish Monitoring
and Surveillance Program.

Additional samples analyzed over the years (2009-2014) show that PBDEs in offshore
and nearshore mussel sites were significantly lower than mussels from harbor-river-
tributary sites (Wilcoxon test, p < 0.05; Figure B.6).

63


-------
200-

¦o 150-

O)

O)
c

c
o

J? 100-

c
cu
o
c
o

o

50-

o-









&



(J®1

^8'





<0

cf



«*

4®

Site Classification

Figure B.6: Whisker plot for total PBDE mussel tissue concentrations obtained from 2009-
2014 across three zone types in the Great Lakes.

The harbor-river-tributary site samples were mainly from Milwaukee Estuary and
Niagara River AOCs. The Niagara tributary sites had the highest concentration of
PBDEs (Figure B.7). Of the 51 congeners analyzed, only 16 congeners were detected
in more than 20 percent of the samples. PBDE 47, 99, 154 and 206 were the most
dominant congeners with more than 75% detection.

64


-------
1 Gill Creek

q ^Cayuga Creek
Niagara-8

^Niagara-9

Gill Creek

Cayuga Creek

Two Mile

Creek^

Smokes Creek

... „ iTonawanda Creek
Niagara-6Q f ^	^

Two Mile Creek^|Ellicott Creek 0

o c

Niagara-5

Tonawanda/Elliot Creek

Scajaquada Creek

NFTA Park/Times Beach

1 L

Total PBDE Cone, (ng/g dry wt)
O 6.5-39.1

•	39.2 -69.1

•	69.2 -236.2

(Scajaquada Creek

O Niagara-4
^Niagara-2
Niagara-IQ A Times Beach

Niagara-30 A NFTA Park

Smokes Creek!

Figure B.7: PBDE characterization using caged and in situ mussels at Niagara River AOC
conducted in 2014.

B.3.3. Pharmaceuticals and Personal Care Products (PPCPs)

PPCPs analyzed included over-the-counter, elicit and prescription drugs, synthetic
musks, antimicrobials, antibiotics and insect repellents (Kimbrough et al., 2018). 141
PPCP compounds were analyzed in 41 tissue samples, of which only 40 compounds
were detected. Amitryptyline (anti-depressant), DEET (insect repellant) and Sertraline

65


-------
(anti-depressant) were the most commonly detected PPCPs in mussel samples (Figure
B.8).

PPCP Compounds

Cuyahoga River-5-(0H)





¦¦



m

u





¦

¦







¦





U

¦

¦



¦¦¦

¦





¦

¦¦

Two Mile Creek-OOA-(NY)







¦





u



¦









¦



¦



¦

¦



¦¦



¦







Milwaukee Bay-4B-(WI)





¦

¦







¦



















¦

















Two Mile Creek-OIA-(NY)













m









































Smokes Creek-OIA-(NY)













¦





¦



¦



¦



























Milwaukee Bridge-6-(WI)

































¦



¦

¦











¦¦



Cuyahoga River-(OH)













¦







¦







¦













¦¦







¦¦



Milwaukee Bay-1-(WI)

































¦

¦¦



¦



¦¦











Milwaukee River-8-(WI)































¦



¦¦

¦





¦¦







¦¦



Tonawanda Creek-OOA-(NY)
Milwaukee River-7-(Wi)















¦



¦















¦¦







¦







¦





















































¦



Milwaukee Bay-O-(WI)

































¦

¦















¦¦



Menomonee River-9-(WI)































¦





¦¦



¦

¦¦











Kinnickinnic River-13-(WI)

¦











¦





















¦ ¦













¦¦



Milwaukee Bay-4A-(WI)

































¦







¦¦







¦



Cayuga Creek-OIA-(NY)































¦ ¦





¦¦¦











Tonawanda Creek-OIB-(NY)







¦





¦











¦¦







¦



¦









¦¦

Milwaukee Beach-5A-(WI)



































¦¦



¦¦







¦



Black River-1-(OH)



¦





































¦¦







¦

¦

Tonawanda Creek-02A-(NY)



























¦

¦







¦







¦



Scajaquada Creek-OOA-(NY)









¦

















¦



















Milwaukee Bay-IO-(WI)





¦







¦























¦







¦¦



Menomonee River-12-(WI)

¦



























¦



¦

¦





¦











Scajaquada Creek-OIA-(NY)



























¦





¦











¦¦¦



Niagara River-4-(NY)
Gill Creek-OIA-(NY)
Black River-O-(OH)

























¦

¦¦

















¦

¦





































¦¦



¦



¦

































¦













¦

¦

Times Beach-OIB-(NY)





































¦



¦



¦



Presque lsle-7-(PA)
Presque lsle-5-(PA)





































¦ ¦









¦





































¦









¦

NFTA Boat Harbor-OIA-(NY)































¦1



¦



¦



¦



Milwaukee Beach-5B-(WI)





































¦







¦¦¦



Ellicott Creek-OIA-(NY)









¦































¦



¦¦

¦

Ashtabula Harbor-3-(OH)











m

























¦



_



¦

¦

Thunder Bay-(MI)







¦





























¦¦











Gill Creek-02A-(NY)

















































Ashtabula Harbor-I-(OH)

















































Niagara River-9-(NY)















¦

































Niagara River-1, rep 1-(NY)
Niagara River-I-(NY)



















































































¦



¦

Gill Creek-03A-(NY)

I

¦

¦¦¦









¦



¦















¦¦¦







¦¦¦¦¦

C

'0

X

0

TJ

C C -r-

O O 2

c a>

•2 ¦—

!


c

c p a>
a) £ c

••r 5".

C  ®

In

0
0

a) (D ~ a) c |— a)

c E £ <° HI .£

— "7Z If)

° y =!

*- ^ m

x XX ~
o-go

! O O £ £ 00

3

(0	(0

X	<1> c

>*	£ 

O)

!>

E
Q

x .a _

s-111 <

o ^ O-O
6 a> °

E

v>
a>

O	£	CL 3

q=	O	3 r

®	— C

E	1	I

Q- Q.
ZJ

o ® &

£ J)

I"

^	Mo	®	^ i:	O-E

2	¦ -	^	>* ™	O CO

c	O E	2	^ E	r

w	P	£ ?	»-

a 5> -t= o
LU > £ -
< h=

¦5,

LU

al

Figure B.8: Presence (¦) and absence ( ) of pharmaceutical and personal care products
found in mussel tissue at various locations. The locations in red are nearsliore sites and the
rest are harbor-river-tributaries sites.

66


-------
At least 10 compounds were detected in nearshore sites (situated away from known
outfalls) that can be considered as reference sites for mussels. This finding underscores
the need for spatially robust monitoring of CECs. Figure B.9 provides perspective on the
relative concentration of the PPCPs that were above three times the detection limits and
occurred at least at five sites.

PPCP Compounds

Verapamil -

Triclocarban-

Sertraline-

Diphenhydramine-
DEET-

Citalopram-

Azithromycin -

Amitriptyline-

Concentration (ng/g (wet weight basis))

Figure B.9. Concentrations of selected pharmaceutical and personal care products. Only
chemical compounds found at more than five sites that had concentrations above 3 x
detection limit are included. This plot provides perspective to the relative concentrations of
the most commonly found PPCPs in dreissenid mussel tissue.

B.3.4. Alkylphenols and Alkyl Ethoxylates

Alkylphenols are used to make alkylphenol ethoxylates (APEO), which are widely used
as industrial surfactants. Two AP compounds- 4-nonylphenol (4-NP), 4-n-octylphenol
(4n-OP) and two APEO compounds- 4-nonylphenol monoethoxylate (NP1EO) and 4-
nonylphenyl diethoxylate (NP2EO) were analyzed in 32 tissue samples. All four
compounds were detected in dreissenid mussels and three of the four compounds were
detected at all sites including nearshore sites that can be considered as reference sites
(Figure B.10).

67

\

I









b

i







•

1-



1

•



















§j. .

















Hp

o

o

o

o

o


-------
Phenol Compounds

Two Mile Creek-OIA-(NY)
Two Mile Creek-OOA-(NY)
Tonawanda Creek-02A-(NY)
Tonawanda Creek-OIB-(NY)
Tonawanda Creek-OOA-(NY)
Times Beach-OIB-(NY)
Smokes Creek-OIA-(NY)
Scajaquada Creek-OIA-(NY)
Scajaquada Creek-OOA-(NY)
Niagara River-9-(NY)
Niagara River-4-(NY)
Niagara River-1, rep 1-(NY)
Niagara River-1-(NY)
NFTA Boat Harbor-OIA-(NY)
Milwaukee River-8-(WI)
Milwaukee River-7-(WI)
Milwaukee Bridge-S-(WI)
Milwaukee Beach-5B-(WI)
Milwaukee Beach-5A-(WI)
Milwaukee Bay-4B-(WI)
Milwaukee Bay-4A-(WI)
Milwaukee Bay-10-(WI)
Milwaukee Bay-1-(WI)
Milwaukee Bay-O-(WI)
Menomonee River-9-(WI)
Menomonee River-12-(WI)
Kinnickinnic River-13-(WI)
Gill Creek-03A-(NY)
Gill Creek-02A-(NY)
Gill Creek-OIA-(NY)
Ellicott Creek-OIA-(NY)
Cayuga Creek-OIA-(NY)

M

r

Q.

o

o o

111 UJ

Q_ CL
2 2

Figure B.10: Presence (¦) and absence ( ) of pharmaceutical and personal care products
found in mussel tissue at various locations. The locations in red are nearshore sites and the
rest are harbor-river-tributaries sites

Of the three most commonly detected compounds, 4-NP registered the highest
concentration in dreissenid mussels (Figure B.11).

68


-------
Phenol Compounds

NP2EO-

NP1EO-

4-NP-

o
o

o
o

CM

o
o

CO

o
o

¦*3"

Concentration (ng/g (wet weight basis))

Figure B.ll: Concentrations of alkylphenols (4-NP) and alkyl phenol ethoxylates (NP1EO
and NP2EO) in dreissenid mussel tissue. Only chemical compounds found at more than
five sites that had concentrations above 3 x detection limit are included. This plot provides
perspective to the relative concentrations of most commonly found alkylphenols and alkyl
phenolethoxylates in dreissenid mussel tissue.

B.3.5. Bivalve Health

The continued presence of legacy organic contaminants coupled with the threat of
contaminants of emerging concern in the Great Lakes necessitates the incorporation of
newer monitoring approaches, particularly effects-based tools to the traditional
chemical-based contaminant monitoring. During Phase I monitoring, MWP conducted
pilot studies to determine the feasibility of using DNA damage assays, metabolomics
and cellular biomarkers in dreissenid mussels.

B.3.5.1. DNA Damage

Tissues of bivalves exposed to xenobiotic substances are threatened by the production
of elevated levels of reactive oxygen and nitrogen species as a result of the metabolism
or direct reactions of pollutants. Overproduction of free radicals can cause damage to
biomolecules including DNA. Biomarkers of oxidatively induced damage, i.e., modified

69


-------
DNA bases and nucleosides in DNA of dreissenid mussels can be used as bioindicators
for environmental genotoxicity.

MWP in collaboration with National Institute of Standards and Technology (NIST)
conducted a pilot project to study the applicability of quantitative mass spectrometric
assessment of oxidatively induced DNA base damage in dreissenid mussels. The aim
of this pilot was to see whether oxidatively induced DNA lesions could serve as early
warning biomarkers for pollution and specifically, to determine whether samples from
polluted sites can be differentiated from those collected from reference sites based on
DNA damage. Mussel samples from two sites in the outer Ashtabula Harbor, a
historically polluted harbor and a reference site in Lake Erie, approximately 6.5 km east
of the Ashtabula River mouth were analyzed for DNA damage. Results show that the
mussels from one site in the outer harbor had significantly greater levels of seven of the
eight measured oxidatively induced DNA bases and nucleosides than those from the
reference site (Table B.1; Jaruga et al., 2017). Further evaluation of this monitoring tool
is planned in Phase 2 with mussels collected from other Great Lakes harbors in
agricultural and industrial watersheds for comparison with reference sites as part of a
larger strategic plan to identify and assess adverse impacts of CECs in Great Lakes
tributaries.

70


-------
Table B.l: The mean and SE of DNA bases and nucleosides measured in mussels from the
harbor site (LEAR-1). All but one were significantly different from those measured at the
reference site (LEAB).



LEAB

LEAR1



DNA base

mean

SE

mean

SE

significant

FapyAde

3.26

0.41

7.15

0.49

*

FapyGua

7.49

1.05

15.7

0.78

*

8-OH-Gua

1.35

0.06

2.14

0.08

*

ThyGly

5.86

0.89

9.94

0.57

*

5-OH-5-
MeHyd

6.65

0.33

8.38

0.42

*

5,6-diOH-
Ura

8.21

0.37

8.41

0.57



S-cdA

0.039

0.003

0.216

0.033

*

R-cdG

0.954

0.173

2.19

0.105

*

S-cdG

2.32

0.40

5.89

0.29

*

B.3.5.2. Metabolomics

Metabolomics is the systematic study of concentration profiles of endogenous
metabolites in biofluids and tissues of a given biological system and has found
applications in many fields including medicine, pharmacology, and more recently in
environmental toxicology. Metabolomics can be used to investigate metabolic changes
within an organism in response to toxicant exposure in laboratory conditions as well as
in natural habitats. However, metabolomics data with respect to dreissenid mussels was
lacking. During Phasel, MWP conducted a feasibility study to determine whether
mussel metabolomics could augment standard practices for evaluating ecosystem
impairment (Watanabe et al., 2015).

MWP partnered with NIST to study the application of NMR-based metabolomics to the
analysis of the whole-body metabolome of dreissenid mussels collected from the three
inner harbor sites of Milwaukee Estuary AOC and a reference site in Lake Michigan in
2012. One of the objectives of this pilot study was to examine whether there were

71


-------
differences in metabolite profiles between impacted sites and the reference site. A total
of 26 altered metabolites with significant differences were successfully identified in a
comparison of dreissenid mussels from an inner harbor site and the reference site
(Figure B.12; Watanabe et al., 2015). This study has demonstrated the feasibility of
NMR-based metabolomics approach to assess whole-body metabolomics of dreissenid
mussels and are being explored further in Phase 2 activities.

>
111

s®
00

CM

O

CL

0.25
0.2
0.15
0.1
0.05
0

-0.05
-0.1
-0.15
-0.2

•0.6

\ LMMB1

-0.1

PC1 (72.26%EV)

0.4

Figure B.12: PCA scores plot of the processed NMR spectra data obtained dreissenid
mussels from impacted site LMMB1 (A), and the reference site LMMB5 (•). The ovals
indicate the 95% Hotelling's confidence interval. The solid line represents the projection of
the scores onto the hybrid scores axis connecting the centers of each group. A Student's t-
test for these projected points shows a significant difference between the groups (p=
5.65xl06).

B.3.5.3. Cellular Biomarkers

Biomarkers are quantitative measurements of biochemical and or physiological changes
in organisms induced either by exposure to and or the adverse effect of xenobiotic
substances. Bivalves respond to xenobiotics exposure by inducing enzymatic

72


-------
antioxidant defense mechanisms to detoxify excess reactive oxygen species. Therefore,
antioxidant response and cellular damage can be useful biomarkers of oxidative stress
in bivalves for environmental monitoring.

MWP ran a pilot project to examine the feasibility of using two cellular biomarkers, total
glutathione (GSH) and lipid peroxidation (LPx), in wild populations of dreissenid
mussels collected from around the Great Lakes to help identify highly impacted sites.
The methods for analysis of these biomarkers in dreissenid mussels were optimized
and tested for both whole body samples and hepatopancreas tissue. Our results
indicated that the GSH and LPx biomarker responses were inversely correlated
(ANCOVA, r2 = 0.40; Figure B.13) and the pattern of response was identical for whole
body samples and hepatopancreas tissue samples suggesting that the laborious work
of isolating organ tissue can be avoided. Based on known biochemical mechanisms,
high LPx and low GSH in animals are typically indicative of stress and in this study, we
were able to rank the sites as 'normal', 'intermediate stressed' and 'highly stressed'.
Additional results of analyses linking biomarker data and mussel tissue burden data will
be summarized in Ringwood et al. (Manuscript in Preparation).

O

o
in

O)

-------
B.4. Key Findings

•	Dreissenid mussels are a valuable bioindicator in the Great Lakes because of
their basin-wide distribution and sessile nature, which allows us to characterize
chemical contamination in three zones: offshore (open-lake), nearshore
(shallow), and river-harbors. Mussels can be caged and strategically relocated
for precise place-based assessments (monitoring along a pollution gradient, pre
and post remediation/restoration assessments and/or contaminant source
tracking) for both legacy and CEC monitoring.

•	Our monitoring data indicates that dreissenid mussels accumulate both legacy
contaminants and a broad suite of CECs, which were previously assumed to
have low or no bioaccumulation potential. Furthermore, dreissenid mussels
collected from hard substrates and hard natural rock bottom in nearshore and
offshore locations reflect what chemical contaminants are bioavailable in the
water column unlike other benthic organism that live in the sediment.

•	Our preliminary work on mussel health metrics such as cellular biomarkers (GSH
and lipid peroxidation), DNA damage, and metabolomics show that mussels can
be utilized for effects-based monitoring and are able to discriminate impacted
sites. The utility of these methods as practical and feasible tools for temporally
and spatially robust CEC monitoring will be explored further during Phase 2
activities.

B.5. Management Implications

•	Dreissenid mussels can be effectively monitored in all three zones of the Great
Lakes for status and trends in chemical contamination, biological effects, and
improve understanding of tropic transfer of contaminants. Compared to fish,
mussels are known to tolerate high levels of pollution and have minimal ability to
metabolize organic contaminants. Further, understanding the trophic
relationships among mussels, fish and birds will likely lead to better predictive
modelling of the distribution and effects of contaminants in the Great Lakes.

•	The ubiquitous piers, jetties, revetments, and breakwaters that improve
navigation into Great Lakes river-harbors support robust colonies of dreissenid
mussels and provide for a standardized sampling approach for comparability of
data across river-harbors and other stressed and polluted habitats in the Great
Lakes.

•	Mussel tissue and sediment data are available through the online database: Data
Integration Visualization Exploration and Reporting (DIVER). For Great Lakes
Mussel Watch chemical contaminant data 2009-2014.

https://www.d iver. orr. noaa.gov/

74


-------
B.6. Knowledge Gaps

•	Can the composition of CEC mixtures in mussels be predicted based on adjacent
land use and point source discharges?

•	How can the contribution of ubiquitous PAHs to bivalve health metrics be
differentiated from other contaminants of emerging concern?

•	How do multiple stressors interact to affect the biological effects in mussels?

•	Is mussel monitoring data (contaminants and health metrics) predictive of
conditions in other species (native mussels, fish and birds)?

B.7. Acknowledgements

The authors gratefully acknowledge those individuals from the National Oceanic and
Atmospheric Administration and our collaborating partners U.S. EPA, U.S. ACE-Buffalo
District, National Institute of Standards and Technology, University of Wisconsin-
Milwaukee, University of North Carolina-Charlotte involved in sample collection,
processing, and analysis. We thank Gunnar Lauenstein and Greg Piniak for technical
input and Edwin Smith and Elizabeth Murphy for program coordination. Support for this
project was provided by the Great Lakes Restoration Initiative through the U.S. EPA's
Great Lakes National Program Office. The views expressed in this work are those of the
authors and do not necessarily reflect the views or policies of NOAA. Any use of trade,
product, or firm names is for descriptive purposes only and does not imply endorsement
by the U.S. Government.

B.8. References

Jaruga P, Coskun E, Kimbrough K, Jacob A, Johnson WE, Dizdaroglu M. 2017.
Biomarkers of oxidatively induced DNA damage in dreissenid mussels: A
genotoxicity assessment tool for the Laurentian Great Lakes. Environmental
Toxicology 32:2144-2153. https://doi.orG M 1 ftox.22427

Kimbrough KL, Johnson WE, Lauenstein, GG, Christensen, JD, and Apeti, DA. 2008.
An Assessment of Two Decades of Contaminant Monitoring in the Nation's Coastal
Zone. Silver Spring, MD. NOAA Technical Memorandum NOS NCCOS 74. 105pp.

Kimbrough KL, Johnson WE, Lauenstein, GG, Christensen, JD, and Apeti, DA. 2009.
An Assessment of Polybrominated Diphenyl Ethers (PBDEs) in Sediments and
Bivalves of the U.S. Coastal Zone. Silver Spring, MD. NOAA Technical
Memorandum NOS NCCOS 94, 87pp.

Kimbrough KL, Johnson WE, Jacob A, Edwards M, Davenport E, Lauenstein G, Nalepa
T, Fulton M, Pait A. 2014. Mussel Watch Great Lakes Contaminant Monitoring and
Assessment: Phase 1. Silver Spring, MD. NOAA Technical Memorandum NOS
NCCOS 180, 113 pp.

75


-------
Kimbrough KL, Johnson WE, Jacob A, Edwards M and Davenport E. 2018. Great Lakes
Mussel Watch: Assessment of Contaminants of Emerging Concern. Silver Spring,
MD. NOAA Technical Memorandum NOS NCCOS 249, 66 pp.

Kimbrough K, Jacob A, Davenport E, Johnson WE and Edwards M. 2019

Characterization of Polycyclic Aromatic Hydrocarbon Characterization in the Great
Lakes Basin using Dreissenid Mussels. Submitted to Environmental Monitoring and
Assessment.

Watanabe M, Meyer KA, Jackson TM, Schock TB, Johnson WE, Bearden DW. 2015.
Application of NMR-based metabolomics for environmental assessment in the Great
Lakes using zebra mussel (Dreissena polymorpha). Metabolomics 11(5): 1302-1315.
doi: 10.1007/sl 1306-015-0789-4.

Online data base. Data Integration Visualization Exploration and Reporting (DIVER).
For Great Lakes Mussel Watch chemical contaminant data 2009-2014.

https://www.diver.orr.iioaa.gov/

Ringwood A, Phan T, Johnson WE, Davenport E, Jacob A, Kimbrough K, and Edwards
M. 2019. Utilization of Cellular Biomarkers in Dreissenid Mussels for the Great Lakes
Mussel Watch. Manuscript in Preparation.

76


-------
Appendix C

Exposure and Effects of
Bioaccumulative Contaminants of
Emerging Concern in Tree Swallows
Nesting across the Laurentian Great
Lakes

Lead Organization: U.S. Geological Survey, Upper Midwest Environmental Sciences
Center, La Crosse, Wl, USA

Contributing Authors: Christine M. Custer, Thomas W. Custer, Paul M. Dummer

Contributing Investigators: Christine M. Custer, Thomas W. Custer, Paul M. Dummer,
J. Christian Franson, Diana Goldberg, Natalie Karouna-Renier, Cole Matson

Corresponding Author Contact: ccuster@usgs.gov

C.1. Problem Statement and Study Overview

Contaminants of emerging concern (CECs) are a loosely defined group of chemicals
whose wide-spread usage or presence in the environment has occurred more recently
or for which there has been relatively little research done until recently. Many of these
CECs are not currently regulated. The National Toxicology Program within the U.S.
Department of Health and Human Services estimates that about 2000 CECs are
introduced each year (https://ntp.niehs.nih.gov/about/). An unknown number may pose
a risk to human or animal health. The Phase 1 (2010 - 2014) CEC work in birds, which
is the subject of this report, assessed exposure across the Great Lakes to
polybrominated diphenyl ethers (PBDEs), perfluorinated compounds (PFASs), and
polycyclic aromatic hydrocarbons (PAHs), and put those exposures into context with
data from biologically relevant endpoints such as reproductive success, as well as,
physiological response indicators (bioindicators) to assess possible effects. The group
of chemicals included in Phase 1 were mainly those chemicals that bioaccumulate in
tissues. Phase 2 (2015 - 2019) CEC work with tree swallows was expanded to include
CECs whose occurrence in the environment is more temporary or seasonal, and that do
not necessarily bioaccumulate. These are often called pseudo-persistent, because,
while they are not long-lived in the environment, there are often daily inputs via waste
water treatment plants, and run-off from farm fields and storm drainages, thereby
making them available to biota year-round. These include pharmaceuticals, personal
care products, and newer pesticides including herbicides. Tree swallow work on these

77


-------
less persistent CECs will be reported in the future, however see other Appendices in
this report for information on some of these types of CECs (Appendices A, B, D).

C.2. Introduction

Tree swallows (Tachycineta bicolor) are an avian species that is now being widely used
to assess contaminant exposure and effects. Their natural history traits, such as
willingness to use artificial nest boxes, their tolerance of human activity in the vicinity of
their nest box, and their food habits (emergent phase of benthic aquatic insects), make
them an ideal study species for examining contaminant exposure and effects.
Additionally, there is now a wealth of information on exposure and concentrations at
which adverse biological effects are likely (effect thresholds) for legacy contaminants,
such as polychlorinated biphenyls (PCBs), trace elements such as mercury, lead, and
cadmium, and organochlorine pesticides. More recently, effect thresholds are becoming
available for some CECs as well.

For the study of CECs in birds, the lack of analytical chemistry methods to meaningfully
quantify exposure in tissues often delays our ability to study them in the field. Biological
matrices are complex and require considerable effort and time before new chemicals
can be reliably quantified in biotic tissues. Another issue regarding the study of CECs in
biota is that some CECs can cause adverse outcomes, but not be accumulated in
tissues in a traditional dose-response manner, so different approaches to study
exposure and effects are required. The choice of sampling matrices and methods to be
used are dictated by the constraints mentioned above. The objectives of Phase 1 were
to quantify exposure to selected bioaccumulative CECs and polycyclic aromatic
hydrocarbons (PAHs), which are bioaccumulative in only some taxa, and to determine if
there were any effects, at different levels of biological organization, associated with
these CECs. Study sites were concentrated in Areas of Concern (AOCs), highly
contaminated areas designated as such by the Great Lakes Water Quality Agreement
(2012), but study locations also included 11 sites for comparative purposes that were
not designated as AOCs (Figure C. 1).

C.3. Methodology

Approximately twenty nest boxes were erected at each of ~70 sites across the Great
Lakes (Figure C.1), protected from ground predators with metal cylinders, and then
checked weekly to follow reproduction including the number of eggs laid, the number
that hatched and the number of young that reached 12 days of age. Because tree
swallows will readily nest in human-made boxes, this allowed data to be collected at all
sites even in highly industrialized and urbanized landscapes where few other bird
species readily nest. Tree swallows arrive in the Great Lakes region in April, lay their
eggs in May, and rear their young in June. By early July, most have fledged from the
nest boxes and have begun to disperse. Each site was studied for at least one year
between 2010 and 2014; a few sites had multiple years of data when the level of
chemical exposure was high, or remediation and restoration actions were in progress or
planned. Egg, blood plasma, and other tissue samples were collected for chemical and
physiological analyses at designated times each year. The types of CECs chemically

78


-------
analyzed in Phase 1 included polybrominated diphenyl ether flame retardants (PBDEs),
perfluorinated substances (PFASs), and polycyclic aromatic hydrocarbons (PAHs), in
addition to the standard suite of legacy and trace element chemicals. Contaminant
analyses followed standard EPA methods with blanks, duplicates, and certified
reference material analyzed with each batch. It is important when assessing effects that
as many chemicals as possible be analyzed, especially those known to cause adverse
effects in wildlife, to avoid drawing incorrect inferences. For this reason, a broad suite of
legacy contaminants was also analyzed. The bioindicator analyses included enzyme
analyses (EROD), oxidative stress measurements (GSH, GSSG, PBSH, TBARS, TSH),
and somatic cell DNA damage. These bioindicators were analyzed by collaborators
(Natalie Karouna-Renier, USGS and Cole Matson, Baylor Univ.) using published
methods for each assay and incorporated standard assay-specific quality assurance
methods. After the data had been assembled, univariate, multivariate, and multistate
modelling statistical analyses were performed to compare among sites, between AOCs
and nearby non-AOCs, and to assess whether there were adverse effects associated
with the chemical contamination. The term non-AOC was used here, rather than
'reference' or 'control' sites, because the 11 sites were chosen to be nearby and
indicative of sites in the Laurentian Great Lakes basin rather than choosing sites that
were known to be pristine or lightly contaminated. Sites not designated as AOCs by the
International Joint Commission under that program (Great Lakes Water Quality
Agreement) could also be contaminated as evidenced by the area around Midland, Ml
which is highly contaminated with dioxins and furans, or Oscoda, Ml which is highly
contaminated by perfluorinated substances. The information presented here is
published as indicated in the literature cited section; data are available in Science Base
of USGS and easily visualized and accessed via a StoryMap

(http://usqs.maps.arcqis.com/apps/MapSeries/index. html?appid=820ce23a0cb04dadb6

525ace6ae4edc7).

79


-------
Wild Rice L

a

St. Louis R.
(4 sites)

Apostle Islands Torch L
Nat'l Lakeshore m

o

Deer L. ¦

Star L.O Manistiquej

St. Mary's (x 2 sites)

D

100

km

o

Green Mtn.

Menominee R. ~
Little Tail Point ^
Green Bay (2 sites) ¦

Sheboygan R. (3 sites) ^

Milwaukee h
(5 sites)

Waukegan Harbor I

Oscoda

o

White L.

a

zoo R

Indian Ridge Ay^Burns

®Ditch
Grand Calumet R.

Midland & Bay City

oa

.. .	St. Clair R.

Muskegon (x 2 sites)

Q	Clinton RjJfl

Kalama- Rouge R.rfB'Detroit R
Raisin R«_ (4 sites)

St. Lawrence R. &
ChaumontR.

<*>

Rochester
(2 sites)

Niagara R.

(2 sites)

Buffalo R. (2 sites)
® Presque Isle Bay (2 sites)

Maumee R QQ Cuyahoga R.

(x 6 sites) Black
R.

| Area of Concern (AOC) |_ = |ake
O Non-AOC	R- = river

Figure C.l Study locations for contaminants of emerging concern (CECs) in tree swallows across the Laurentian Great Lakes
basin, 2010-2014.

80


-------
C.4. Key Findings

This compendium of work is one of the most extensive and exhaustive studies of both
legacy and CECs in birds from such a large geographic area (100,000 sq. miles),
collected within a comparable timeframe, and using the same species. Contaminants of
emerging concern were present in tree swallows nesting at these 70 sites across the
Great Lakes, and at concentrations that varied by location and land use. These
differential exposure patterns allowed for a rigorous assessment of both reproductive
effects and physiological responses. Two of the CECs (PBDEs and PFASs) do not
seem to be at exposure levels that affect either reproduction or physiological responses
in nesting birds, however, one class of CECs, the PAHs, were associated with adverse
reproductive, as well as, physiological responses.

C.4.1 Polycyclic Aromatic Hydrocarbons (PAHs)

In the context of contaminants in birds, PAHs are a CEC, not because they are a
recently discovered contaminant, but because they have been little studied in vertebrate
biota. This is because PAHs do not accumulate in a clear dose-response manner in
birds, like some other CECs and most legacy organic contaminants do, so different
approaches are needed to conduct field assessments in vertebrate biota. Because there
is only limited metabolism of PAHs in aquatic insects and mussels (see Appendix B),
both of these animal groups can serve as a measure of exposure to vertebrate biota
that consume them. Therefore, PAHs were assessed in pooled samples of diet,
primarily the aerial stage of benthic aquatic insects, from stomachs of nestling tree
swallows, and then compared, on a site basis to background exposure levels and
reproductive and physiological effects in swallows.

There are several hundred individual PAHs which differ in the number of benzene rings
and attached methyl groups (see review in Abdel-Shafy and Monsour 2016). In this
study we quantified 48 individual PAHs (detection limit = 0.25 ng/g wet wt., 20 parent
and 28 alkylated PAHs, Custer et al. 2017a). Polycyclic aromatic hydrocarbons can be
produced naturally, through burning of biomass in forest fires or internally in biota, but
also through the actions of humans. Industrial production of PAHs has occurred since
the Industrial Revolution, including in manufactured gas plants and through intensive
fossil fuel burning. The PAHs which were the most prevalent and at the highest
concentration were in the more highly industrial watersheds such as the Detroit River
corridor and the Milwaukee Estuary, Wl, and to a lesser extent in the Chicago, IL,

Duluth, MN, and Toledo, OH areas (Figure C.2, Custer et al. 2017a). At many of the
sites, the source of PAHs was a mixture of pyrogenic (resulting from combustion) and
petrogenic (petroleum sources) based on the ratio of phenanthrene to anthracene.
Pyrogenic PAHs tend to have more benzene rings (4 - 6) and are usually not alkylated
(i.e. with few methyl groups attached). Petrogenic PAHs tend to be smaller molecules, 2
- 4 benzene rings, and have attached methyl groups. The sites with higher total PAH
concentrations, such as the Rouge and Detroit Rivers, Ml, and the Milwaukee, Wl and
Duluth, MN areas, had predominately pyrogenic sources (Custer et al. 2017a). These
sites had an abundance of historic heavy industry including steel mills and other
industries with a heavy reliance on fossil fuels. The Raisin and Maumee Rivers, both

81


-------
sites with moderately high PAHs exposure, had petrogenic-based sources
predominantly. Most sites (>70%) had a mixture of PAH sources. Adding to the
uncertainty of apportioning sources of PAHs is that there can be transformation of PAHs
in sediments (Lei et al. 2005) and in biota (Naf et al. 1992).

There was sufficient exposure to PAHs, as measured in the diet, at some sites to illicit
both a physiological response and a reproductive effect in the swallows. Whereas the
physiological responses were anticipated based on previous avian work, the association
with adverse reproduction was a surprise, in part because there have been very few
studies that have tried to test for an association in a field situation. On the physiological
level, both the alkylated and parent PAHs were the primary drivers separating high total
sulfhydryl (TSH) from normal TSH levels (Custer et al. 2017b). Total sulfhydryl is a
measure of cellular oxidative stress which is an imbalance created when there is an
increased amount of reactive oxygen (free radicals) that exceeds the body's ability to
detoxify them. This imbalance can lead to oxidative damage to proteins, molecules, and
genes. Total PAHs in the elevated TSH group were 8 times higher compared to the
normal TSH group. Protein bound sulfhydryl (PBSH), another measure of oxidative
stress, followed this same pattern, likely due to the high degree of correlation between
these 2 measures of oxidative stress. Ethoxyresorufin-O-dealkylase (EROD) activity, a
liver enzyme used as an indicator of exposure to PAHs and other organic contaminants,
was also higher in birds exposed to high concentrations of PAHs in their diets. This
enzyme is activated when a toxin reaches sufficient levels to necessitate a physiological
response by the organism to detoxify the contaminant. Geometric mean concentrations
of total PAHs were 16 times higher in the highly-active EROD group compared to the
normal group. Toxic equivalency (TEQ) is a method to quantify the toxicity of a group of
related chemicals each of which may have a different level of toxicity. The TEQ value
for the sum of the PAHs was 39 times higher in the EROD induced group compared to
the normal EROD group. The coefficient of variation for DNA (DNA-CV), a measure of
disruption of DNA division processes in somatic cells, in this case red blood cells, did
not clearly associate with PAH exposure even though PAHs have been found to be
associated with DNA-CV in other studies. Statistical analyses in these other studies
were primarily single-variable analyses not multivariate analyses as was done in the
current work (Custer et al. 2017b), which may possibly explain these differing results.

The association of PAHs and reproductive success in this study is one of the first such
association found. While the detrimental effects of oil spills either on birds or their eggs
is well known and documented, and the effects of injected petroleum in the laboratory
on the survival of avian embryos is also well documented, the effects of ingested PAHs
are much more difficult to study. This difficulty results because PAHs cannot be
meaningfully measured in vertebrate tissue, because they are quickly metabolized and
removed from the body. Because of this, a different approach was needed, namely to
measure PAHs in the invertebrate food items that are being consumed. Getting
adequate sample mass for chemical analyses of the food consumed is difficult and time
consuming.

82


-------
2000

0

5

¦5?

c

(n

1

£

ra
,o

1500

1000

500

0

I Alkylated PAHs

~ Parent PAHs

Mean non-AOCs (132 ng/g)



/

ffffTTTTTT

d> CD Jit

ro ro > > >
-j -i cncria:

 £ X C CD

®°oroa)

>%'F

o E

=i < -0
^ >> <0 -

^ CO -c 0)
CD CO ^
c
CD
CD

<2

aj

(U O C 0)

> .Q CD > c

in ro § cn o
£ ro I .S-^

mfiiii1;

3^ (dt:
® win
<* CD ~1

-* JZ *

o
o

tr

¦— >, i—	i— * i—

a) ro cd	^ ^

> m >	> c >

qi id cn	cn 2 cn

CD C C	CO

m < O	O) CD CO

SScOSu

® > 73	ro CD J)

Z OH 0	>>* }

5	o

ro	u
c

D)

CD
CO



CD ro -o ^ CD ^ ^
— "O CO r- ¦— ^ ¦— ¦—

^ocDroC^cDQrir
!= ^ ^	O ^

CD
CD

S 7* ~
O ® 2

* cr*
w
CD

i_

CL

_	O

>(D	O 0

ro y	N ro

•> cn	2 co

£t}	c

CO =	ro

>	CD

7	^

££
o

Area of Concern (AOC) and non-AOC locations (with asterisks)

Figure C.2 Concentrations of total poiycyclic aromatic hydrocarbons (PAHs) in the diet of tree swallows at study sites across
the Laurentian Great Lakes basin, 2010-2014. Graph adapted from Custer et al. 2017a. Environ. Toxicol. Chem. 36: 735-748.

83


-------
There was a decrease in reproductive success as PAH exposure, as measured in their
food, increased. Reproductive success was quantified as the daily probability of egg
failure (Custer et al. 2018) which accounted for adverse hatching outcomes, such as
embryo death or infertility, and the timing of those adverse effects. As the probability of
egg failure increased, the exposure to PAHs via their diet also increased (Figure C.3;
Custer et al. 2018). Because of other factors known to affect reproductive success,
including other contaminants such as the dioxin and furan TEQs, as well as ecological
variables such as female age and date within season, the association with total PAHs
was not strong, but warrants additional work.

cc

L—

o
E

CD
CD
0

4—

o

-I—'

15

03

_Q

O
!_

CL

"to
~

0.08

0.07

0.06

Maumee

R = 0.28
P = 0.036

L Erie Metro Park
(Detroit R.) Rouge R.

plant ~

Zug Island
(Rouge R.)
~

Lakeshore Park
(MilwauJ^e Estuary)

Three Bridges (Milwaukee Estuary)
^Trenton Channel (Detroit R.)

~ Connor Creek (Detroit R.)

1000

2000 3000

4000

5000 6000

Total PAHs (ng/g wet wt.) in swallow diet

Figure C.3 Association of egg mortality and exposure of tree swallows to polycyclic
aromatic hydrocarbons (PAHs) at sites across the Laurentian Great Lakes basin, 2010-
2014. Area of Concern name in parentheses where appropriate. Graph adapted from
Custer et al. 2018. Ecotoxicol. 27: 457-476.

84


-------
C.4.2 Polybrominated Diphenyl Ethers (PBDEs)

The polybrominated flame retardants (PBDEs) were detected in all egg samples
(detection limit = 0.004 ng/g wet wt.; 40 congeners analyzed, Figure C.4; Custer et al.
2016, 2017a) with the highest concentrations found in the highly industrialized,
urbanized corridor along the Detroit River, Ml and in Cleveland, OH. Another site with
elevated exposure to PBDEs, and the only site that was statistically above background
exposure in eggs, was near Midland, Ml. The geometric mean concentrations at the rest
of the sites were at or below the mean background concentration (96 ng/g wet wt.)
which was recently established in a tree swallow study in Canada (Gilchrist et al. 2014).
Sites not associated with AOCs had a mean concentration of 48.6 ng/g which can now
also be considered a background value as well (Custer et al. 2016). There are few other
studies that have established background concentrations in eggs of a passerine bird
species. In both nestlings and diet, the site with the highest exposure was at Torch
Lake, Ml (Custer et al. 2017a) followed by other highly industrialized rivers such as the
St. Clair and Niagara Rivers, and the Detroit River corridor. It is unclear what the source
might be for the PBDEs in Torch Lake. A similarly small amount of data exists for effect
thresholds for PBDEs, either reproductive effects or physiological responses. The
lowest observed effect level (LOEL) for hatching success was estimated to be ~1000
ng/g in osprey eggs (Pandion haliaetus, Henny et al. 2009). Consistent with that LOEL
value, and because the highest mean egg concentrations were >7 times lower than that
threshold, we found no association of PBDE exposure with the daily probability of egg
failure (Custer et al. 2018). We also did not find any biomarker responses associated
with PBDEs (Custer et al. 2017b) indicating that exposure did not rise to the level that
prompted a physiological response in the swallows.

85


-------
1038

400

300

Geometric mean (top of bar) and 95% CI (vertical line)

Lower limit for hatching
effect threshold > 1000 ng/g

background (Canada, 96 ng/g)
l/

 > > >

or en a: a:

®	Si	m	^	«	

ro	.2	(»	m m >	as

-j	cm	c	v -o a:	-j

0 0>	"Jrr	V)	CO £

«	11)^	ur

-j	t=	3 a>	en >

-=	co	cr >	co >

X * £ £ z

Si to
CL "5
o
—I

•M

CO

 a) >

CH E 01

 >> ¦*-
=J CO O O

cr -Q ro 0

ts E m h

¦!2 LLI

C i-

co £
cu

JZ

o
o
a:

x;
c

9 a3 o3  _ ...
w > > > > -p

oa a: ai ai 5

X C (U trt is

y o jj 0

g-i j o

_ o £ ^
< -Q o ^
> J! c1"

£
m co

c

(D
CD

O

-------
C.4.3 Perfluorinated Substances (PFASsJ

Perfluorinated substances were detected in all tree swallow plasma samples (detection
limit = 0.56 ng/ml_;13 PFAS congeners separately analyzed), and at concentrations that
varied among sites (Figure C.5). As with other CECs, there are limited data on
background exposure levels, but in tree swallows background levels in plasma are <75
ng/mL (Custer et al. 2014). Only slightly higher was the mean exposure of 111 ng/mL at
7 non-AOCs across the Great Lakes (Figure C.5, Custer et al. 2017a). Sixty three
percent of the AOCs studied had mean PFAS concentrations less than the mean
concentrations at non-AOCs. There were 2 hot-spots of PFAS exposure, one near
Duluth, MN (581.9 ng/mL) and the other near Oscoda, Ml (1649.3 ng/mL); both
locations were associated with airfields where film-forming fire-fighting foams were
extensively used in fire-suppression training exercises. Like PBDEs, highly urban and
industrial sites, including locations along the Detroit, Rouge, and Raisin Rivers in
Michigan, the Chicago, IL area, and along the Niagara River, NY (Custer et al. 2017a),
tended to have higher exposures to PFASs than less urban/industrial sites (Figure C.5).
One PFAS in particular, perfluorooctane sulfonate (PFOS), dominated the suite of
PFASs found in blood plasma. This result is similar to other studies world-wide that
found that PFOS was the prevalent PFAS congener, especially in urban and industrial
locations.

Similar to PBDEs, exposure to PFASs, except at one location, was below exposure
levels that cause hatching effects in laboratory studies (Newsted et al. 2005). A
predicted no-effect concentration in that study was set at 1000 ng/mL in serum, which
was 5 - 10 times higher than mean exposure at many of the 70 sites across the
Laurentian Great Lakes. Our field data were therefore consistent with the laboratory
data. Even at the site near Oscoda, Ml which had extremely high exposures,
reproductive success of swallows was above average (Figure C.6; Custer et al. 2018);
Oscoda had the 5th lowest egg failure rate among the 37 locations. There were also no
biomarker responses associated with PFAS exposures (Custer et al. 2017b) indicating
the lack of a physiological response induced by this class of contaminants.

87


-------
CD

w
CO
<

TO

*4—'

O

2000

1500

1000

500

0

Geometric mean (top of bar) and 95% CI (vertical line)



Mean non-AOCs (111 ng/ml)
\

co	a>
u

o	co

o	_i

U5 cd
O .y
* en

"O

UJ
CO

¥
cd
>

tr

CD

CD

ii—

CD

*

CD
O)
"O

>

>

>

in

cn

in



£Z

CD

in

0

O

CD

"55

•*—>
c

=3
O

sr
CO

Q

0

m "O

£Z

1— *

a> ^

E ^

co

CO
CD
* CO

CD a) CD
> > >

n in in

co o .b
o co

N TX

co O

E co
_cg
co

c
CD

E
>>
CO
_Q

E

LU

CD
-*—>

V)

0

sz

o

o

cn

CD

>	C

cn	o

CD	Q-

O" CO

I 1	I

 js;

E ^

is

c
O)

co
CO

[ilDddndintgDidtbifaeiA ^J3 1 C3.

*

>> *
CO c
a CD
to <->
LU .E
CD >
CD 0
CL

« 03
§ o

-kr

CO
_1

I	

CO
-*—'

CO

*

CD ^ <2
^ CO
CC -a
X d

o <

U_ ^

T3 ^

c >

< a:

>> U5
CO =

m =

CD :d
CD CO

CD

CD CD <_¦
> > C

a: a: 2

CD C C
O) CO CD
O CD CD
r >, >-
co o O

>,jQ *
J CD

o -«=

CO

— ® 5 ffl ® ®
c > > CD > CO

0	a: a: cd a: -d

1	o ® " >, ^

ilpp

C—' (0 (/) <-.
* O CD CO

c Ql

CO

CD

CO
CD

y CQ

CD CD

CO	CO

_J	_J

CD	-C

.-•=:	o

If

CO

5

Area of Concern (AOC) and non-AOC locations (with asterisks)

Figure C.5 Concentrations of total perfluorinated substances (PFASs) in tree swallow nestling plasma (n=566) at study sites
across the Laurentian Great Lakes basin, 2010-2014. Graph adapted from Custer et al. 2017a. Environ. Toxicol. Chem. 36:
735-748.

88


-------
3 0.080

2?

a

U)
CD
0

o

>.

TO
-Q
O

Q.

>

TO

Q

0.060

0.040

0.020

L

95% CI band for m€
for 10 non-AOCs

0.000

-0.020 J

u

irt

>. *

CO
CD
_0
w
0

O" *

 > > > E
"o bd bd bd bd o

E > iS as

CD <
=3 >.
03
00

C
0
0

5

o
od

£
=3

m

Oi CL

.03

o

CO

f\ £. ft

Q) i

.55 "0

0
_ >
bd be

C £=
03 O

Si
£ o

0
-£=

CO

>, 0
J*.

co
LU
0
0

03

"D
C

C 2

o i2

CD c

0 o
S.

o
o

0
-*£
03

O

j*:

co

=3

= CO

0

>

q:
0
E

0!

o

-a
s=

2

6

T

*

O) ~

¦g

b£

c
ro

tj
c

s

il iiiiUMiiLit D

0

>

bd

0
0
c

o
c
0

c

(O
03

cd

0

>

bd

>. 0

c

0
o
c

>
0

03
CD
~o
c
<
0

03
_l

V	

0
0

Q

Cl >

« a:

O to
*

o

_l

CO

0 0

> >

bd bd
o +=:
o o

N
03

E

_03
03

O
J3

0

Q

0

>

bd	oi

03	m

O)	C

o	CO

t	®

O	03

0

0

bd ro ad

in —i
>. 0

|sm

S "O

(f) ~

I j [	Vf —' —'	i— 1—

o'i!	d w	O 3

_o? 0	g~0 o	0 m

CD "co	tr! * -n	^ n ^

>.
03
-Q

E

LU

i—	*	*

>	-g	0

Dd	o	E

0	O	=J £=

Z3	CO	— "

CO
0
sz
o
o
Cd

« * o o

c 03 * *

03

O

® 0

03

a>

03

0 0
.-!=: >
_S= 03

s8
<:
c
o

Areas of Concern and nearby non-AOCs (with asterisks)

Figure C.6. Daily probability of egg failure (me) at study sites across the Laurentian Great Lakes basin, 2010-2014. Vertical
lines are ± 1 standard error. Graph adapted from: Custer et al. 2018. Ecotoxicol. 27: 457-476.

89


-------
C.5. Management Implications

The quantity of 2 of the 3 CECs documented in this Phase 1 work in swallows have
either declined or stabilized in biotic tissues in the Upper Midwest including the
Laurentian Great Lakes, generally because of either voluntary change by the
manufacturer (PFASs), regulatory actions (PBDEs), or a combination of both. For
example, PFOS concentrations in great blue heron (Ardea herodias) eggs declined by
>50% between 1993 and 2010 and 2011 (Custer et al. 2013) along the Mississippi
River. There has been a decline in PFOS in herring gull (Larus argentatus) eggs
between 1990 and 2010 (Gebbink et al. 2011) in the Great Lakes. While PFOS declined
in both studies cited above, some of the perfluoroalkyl carboxylates have tended to
increase over those same timeframes. Continued monitoring of PFAS congeners,
especially new ones entering the environment, as well as development of analytical
methods for PFAS congeners that are currently not widely available seems warranted,
as do focused studies on effects of those congeners that are increasing. The picture for
the PBDEs is less clear (Gauthier et al. 2008), but should be resolved with analysis of
the ongoing data collected as part of the Canadian Wildlife Services herring gull
monitoring program (http://iic.org/greatlakesconnection/en/2018/04/herring-gulls-are-
sentinels-of-the-skies/; Hebertetal. 1999). There seems to have been a decrease of
PBDEs in avian tissues once usage was restricted, but more current information is
needed to confirm trends. The tree swallow data set presented here can become a
baseline to access future trends in PBDE and other CEC exposure. Polycyclic aromatic
hydrocarbons remain an underappreciated issue because PAHs do not bioaccumulate
in vertebrates in the typical dose-response manner which makes that class of CECs
difficult to study. Development of alternative methods should continue to more fully
understand the possible effects of current exposure to PAHs, along with the inclusion of
additional species in the effects studies (mussels and fish), and the addition of study
sites where PAH exposure is high should be completed.

C.6. Knowledge gaps

While many of these bioaccumulative CECs are now being more widely studied in birds,
CECs such as pharmaceuticals and personal care products, and next generation
pesticides including herbicides and new types of insecticides, such as the
neonicotinoids, need more research including new methods, in some cases, to quantify
and study exposure and effects. These pseudo-persistent CECs are often more short-
lived in the environment, but because they are continually entering the environment via
non-point source run-off and passing through waste water treatment facilities, there is
continual, but low-level exposure. Phase 2 CEC work on tree swallows is providing new
data to fill these gaps for birds, including not only on exposure, but also to add
information on possible effects of CECs on the activity of the thyroid hormone system
(T3 and T4 concentrations) as well as the other commonly-used biomarkers. The use of
metabolomics and transcriptomics to quantify perturbations in various physiological
pathways and genes that may be impacted by these non-accumulative CECs is also
ongoing.

C.7 Disclaimer

90


-------
The USGS information has been peer reviewed and approved for publication consistent
with USGS Fundamental Science Practices processes. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.

C.8 Acknowledgements

We thank the over 75 landowners who allowed us access to their properties during the
course of our field work and also thank the 30 field technicians and data entry support
personnel for their help. The authors thank Heiko L. Schoenfuss and Ed Johnson for
helpful comments on earlier drafts of this report. Funding was provided by the Great
Lakes Restoration Initiative, as well as, the Environmental Health mission Area of U.S.
Geological Survey.

C.9 References

Abdel-Shafy, H.I., and M.S.M. Mansour. 2016. A review of polycyclic aromatic
hydrocarbons: source, environmental impact, effect on human health and
remediation. Egyptian J. Petrol. 25:107-123.

Custer, T.W., P.M. Dummer, C.M. Custer, Q. Wu, K. Kannan, and A. Trowbridge. 2013.
Perfluorinated compound concentrations in great blue heron eggs near St. Paul,
Minnesota USA in 1993 and 2010-2011. Environ. Toxicol. Chem. 32(5): 1077-1083.

Custer, C.M., T.W. Custer, P.M. Dummer, M.A. Etterson, W.E. Thogmartin, Q. Wu, K.
Kannan, A. Trowbridge, and P.C. McKann. 2014. Exposure and effects of
perfluoroalkyl substances in tree swallows nesting in Minnesota and Wisconsin,
USA. Arch. Environ. Contam. Toxicol. 66:120-138.

Custer, C.M., T.W. Custer, P.M. Dummer, D. Goldberg, and J.C. Franson. 2016.
Concentrations and spatial patterns of organic contaminants in tree swallow
('Tachycineta bicolor) eggs at United States and binational Great Lakes Areas of
Concern, 2010-2015. Environ. Toxicol. Chem. 35:3071-3092.

Custer T.W, C.M. Custer, P.M. Dummer, D. Goldberg, J.C. Franson, and R.A. Erickson,
2017a. Organic contamination in tree swallow (Tachycineta bicolor) nestlings at
United States and Binational Great Lakes Areas of Concern. Environ. Toxicol.

Chem. 36:735-748.

Custer, T.W., C.M. Custer, P.M. Dummer, E. Bigorgne, E. Oziolor, N. Karouna-Renier,
S. Schultz, R.A. Erickson, K.A. Aagard, and C.W. Matson, 2017b. EROD activity,
chromosomal damage, and oxidative stress in response to contaminant exposure in
tree swallows (Tachycineta bicolor) nestling from Great Lakes Areas of Concern.
Ecotoxicol. 26:1392-1407.

Custer, C.M., T.W. Custer, M.A. Etterson, P.M. Dummer, D. Goldberg, and J.C.
Franson. 2018. Reproductive success and contaminant associations in tree
swallows (Tachycineta bicolor) used to assess a Beneficial Use Impairment in U.S.
and Binational Great Lakes' Areas of Concern. Ecotoxicol. 27:457-476.

91


-------
Gauthier, L.T., C.E. Hebert, D.V. Weseloh, and R.J. Letcher. 2008. Dramatic changes in
the temporal trends of polybrominated diphenyl ethers (PBDEs) in herring gull eggs
from the Laurentian Great Lakes: 1982-2006. Environ. Sci. Technol. 42:1524-1530.

Gebbink, W.A., R.J. Letcher, C.E. Hebert, and D.V. Weseloh. 2011. Twenty years of
temporal change in perfluoroalkyl sulfonate and carboxylate contaminants in herring
full eggs from the Laurentian Great Lakes. J. Environ. Monit. 13:3365—3372.

Gilchrist T.T., R.J. Letcher, P. Thomas, and K.J. Fernie. 2014. Polybrominated diphenyl
ethers and multiple stressor influence the reproduction of free ranging tree swallows
('Tachycineta bicolor) nesting at wastewater treatment plants. Sci. Total Environ.
472:63-71.

Great Lakes Water Quality Agreement. 2012. The Government of Canada and The
Government of the United States of America. Available from

http://www.epa.gOv/greatlakes/.glwqa/.

Hebert, C.E., R.J. Norstrom, and D.V. Chip Weseloh. 1999. A quarter century of

environmental surveillance: the Canadian Wildlife Service's Great Lakes herring gull
monitoring program. Environ. Rev. 7:127-166.

Henny C.J., J.L. Kaiser, R.A. Grove, B.L. Johnson, and R.J. Letcher. 2009.

Polybrominated diphenyl ether flame retardants in eggs may reduce reproductive
success of ospreys in Oregon and Washington, USA. Ecotoxicol. 18:802-813.

Lei L., A.P. Khodadoust, M.T. Suidan, and H.H. Tabak. 2005. Biodegradation of
sediment-bound PAHs in field-contaminated sediment. Water Res. 39:349-361.

Naf, C., D. Broman, and B. Brunstrom. 1992. Distribution and metabolism of polycyclic
aromatic hydrocarbons (PAHs) injected into eggs of chicken (Gallus domesticus)
and common eider duck (Somateria mollissima). Environ. Toxicol. Chem. 11:1653-
1660.

Newsted, J.J., P.O. Jones, K. Coady, and J.P. Giesy. 2005. Avian toxicity reference
values for perfluorooctane sulfonate. Environ. Sci. Technol. 39:9357—9362.

StoryMap: Utilizing Tree Swallows as Indicators for Contaminants in the Great Lakes
Area. U.S. Geological Survey. February 7th,

2018. http://usqs.maps.arcqis.com/apps/MapSeries/index.htmi?appid=820ce23a0cb
04dadb6525ace6ae4edc7. Date accessed: July 20, 2018.

92


-------
Appendix D

Survey of Contaminants of Emerging
Concern and Their Effects to Fish and
Wildlife in Great Lakes Tributaries

Lead Organization: U.S. Fish and Wildlife Service, Ecological Services, Region 3,
Bloomington, MN, USA

Contributing Authors: Stephanie L. Hummel and Heiko L. Schoenfuss

Contributing Investigators: Mandy Annis, Jo A. Banda, Amber Bellamy, Vicky Blazer,
Mark E. Brigham, Steven J. Choy, Sarah M. Elliott, Daniel J. Gefell, Stephanie L.
Hummel, Zachary G. Jorgenson, Kathy E. Lee, Jeremy N. Moore, Heiko L. Schoenfuss,
Linnea M. Thomas, Annette Trowbridge, and William A. Tucker

Corresponding Authors Contact: stephanie_hummel@fws.gov;
hshoenfuss@stcloudstate.edu

D.1. Problem Statement and Scope

The Laurentian Great Lakes are a critical economic and environmental resource in
North America. These lakes support an abundance of wildlife refuges; fish hatcheries,
and a thriving commercial and recreational fishery (USFWS 2017, USFWS 2018b). The
Great Lakes Basin is also home to 21 federally endangered species, 14 federally
threatened species and many more species are threatened or endangered at the state
level making restoration and conservation of the Great Lakes Basin's natural resources
important for the continuing benefit of current and future generations (USFWS 2018a).

The Great Lakes and their tributaries have been subject to historical and recent
degradation associated with human development. Legacy pollutants from early
industrial development of the Great Lakes Basin have been researched, regulated and,
in some instances, mitigated. More recently, however, new classes of chemicals have
been identified as being of emerging concern to fish and wildlife and human health.
These contaminants of emerging concern (CECs) are a loosely defined group of
chemicals whose wide-spread usage or presence in the environment has occurred
more recently. Additionally, due to analytical chemistry limitations, cost, or other
impediments relative little research has been completed on CECs, until recently. Little is
known about the occurrence of CECs and their effects on wildlife.

Therefore the overall objective of this study was to assess the presence and
biological consequences of CECs in US tributaries of the Great Lakes.

93


-------
0 26 50 ^5 1»M
»,2550751«fKilomet|

EXPLANATION
Land Cover
H Water
~ Developed (open and low intensity)
I Developed (medium and high intensity)
I Forest, shrubland, herbaceous, and barren
I I Planted/cultivated
I I Wetland

	State/province boundary

	Great Lakes watershed boundary

Water and sediment sampling site

Base composited from North American Atlas political boundaries, 1:10,000,000,2006;
North American Atlas hydrography, 1:10,000,000,2006; U.S. Geological Survey
hydrologic unit maps, 1:250,000,1994; Canada Land Inventory Level-I watershed maps,
1:2,000,000; National Land Cover Database 2011,30-m resolution, 2015. Universal
Transverse MercatorZone 16 North, referenced to North American Datum of 1983.

Figure D.l: Map of the 24 water bodies sampled within the Great Lakes Basin. Water and
sediment were sampled at all locations. Sampling occurred from 2010-2014 during spring,
summer, and fall months. Each white triangle indicates a tributary sampled, but within
each tributary 5-36 individual samples were collected. Sampling location names from west
to east are as follows: St. Louis River, Waupauca Chain ()' Lakes, Little Lakes Butte Des
Mortes, Fox River, Kewaunee River, Milwaukee River, Chicago (North Shore Channel and
Little Calumet), Grand/Maple River, Saginaw River, Swan Creek, Maumee River, Raisin
River, Detroit River, Clinton River, St. Clair River, Cuyahoga River, Tinkers Creek,
Ashtabula River, Long Pond, Genesee River, Irondequoit Bay, Oswegatchie River, and
Raquette River.

As the human population has grown, so has the market for new and innovative
products. With human population growth and industry projected to increase in the Great
Lakes Region for the next two decades (Pendall et al. 2017) it can be assumed the
threat of CECs is not easily understood or quickly solved. This population growth in turn
causes increasing use of new products and chemicals and therefore new and

94


-------
increasing inputs into water sources from household use, direct point source inputs from
industry, and run off from agricultural and urban areas. The overall presence and
occurrence of CECs in the Great Lakes Basin and their biological effects remain largely
unknown. Given the geographic scale (>260,000 km2) of the study area and the
diversity of aquatic habitats and surrounding land use, this study proceeded
sequentially. A discovery phase (2010-12) focused on determining presence and
concentrations of CECs in water, sediment and fish tissues in US tributaries across all
five Great Lakes. Biological investigations of native fish health accompanied this focus
on chemical detection. In 2013, the study progressed to a two-year focused assessment
of the linkage between CEC presence, land use and biological effects. When integrated,
these efforts produced the most in-depth and comprehensive assessment of the CEC
exposure on native fish and wildlife ever conducted in the North American Great Lakes.

D.2. Methodology

At the onset of this study, only scant and discontinuous knowledge of the presence and
concentrations of CECs in Great Lakes tributaries and its native fauna existed. To meet
the project objective, it was necessary to select representative US tributaries to the
Great Lakes across all five lakes and encompassing a variety of land uses. This list
included 24 rivers, impoundments and embayments across the Great Lakes Basin
(Figure D.1).

To assure consistency and quality of sampling, all samples were collected using peer-
reviewed standard sampling protocols for water, sediment and fish tissue (Lee et al.
2012; Elliott et al. 2015, 2016). These protocols required the collection of duplicate
samples, split samples, and blank samples to assess variability and detect
contamination of samples from other sources. All chemical concentration data reported
here are the result of these rigorous quality assurance practices and exclude any data
not meeting minimum quality standards.

As the investigations moved beyond the survey of CEC presence and concentration to
the assessment of biological effects in native fishes, similar quality assurance protocols
were implemented (Blazer et al. 2014; Thomas et al. 2017; Jorgenson et al. 2018). The
biological analysis between 2010 and 2012 focused on small numbers offish in a few
selected sites (Blazer et al. 2014) and expanded to a broader assessment of native and
caged fish from six rivers over two years in 2013 and 2014 (Thomas et al. 2017;
Jorgenson et al. 2018). Similar to the selection of rivers for chemical analysis, rivers for
biological assessment offish were chosen to represent a continuum of land use from
forested reference streams (Raquette River, NY), to intensely agricultural watersheds
(Fox River, Wl), and densely urban watersheds (Chicago River, IL; Clinton River, Ml). In
each river multiple sites along the rivers' length were chosen to bracket known point
sources (specifically treated wastewater effluent discharge) and transitions in land use.
Using existing protocols established by Environment Canada (Munkittrick, 1992) to
assess fish health in effluent affected streams, an effort was made to collect bass,
suckers and sunfish at each stream site. These collections were augmented with two-
week caging of hatchery-reared sunfish at five to six sites in each river to provide a
more controlled biological integrator of exposure and biological effect.

95


-------
Once fish were processed, rigorous quality assurance and quality control protocols were
implemented (Lee et al. 2012; Blazer et al. 2014c; Elliott et al. 2015, 2016) to avoid
sampling or analysis bias. These measures included duplicate and triplicate sample
analysis, randomizing of sample analysis and "blinded" analysis of samples to assure
that the researcher is unaware of the source of each sample. Lastly rigorous statistical
methodology was applied to reduce the likelihood of "false positive" results.

D.3. Key Findings

D.3.1 CECs Are Ubiquitous in Great Lakes Waterbodies

The objective of Phase 1 of the study was to characterize CECs across the Great Lakes
basin to better understand their presence and concentration. CECs were found in
surface water and sediment of all 24 waterbodies sampled from 2010 to 2014 (Figure
D.1). At each of the 24 waterbodies between two and 36 different sampling sites were
established. Samples were analyzed for over 200 chemicals which were then grouped
into nine chemical classes:

1.	Alkylphenols, building block materials and additive chemicals used to make other
chemicals and products such as resins, fuels, fragrances, lubricants and
adhesives to name a few

2.	Flavors and fragrances, found in perfumes, soaps, and detergents, and food
products with artificial flavors

3.	Hormones, naturally occurring and man-made pharmaceutical products for
human and animal use

4.	PAHs, stands for polycyclic aromatic hydrocarbon, and are found in coal tar
residues and are also from incomplete organic matter combustion such as
vehicle exhaust, wood burning, and coal-tar pavement sealcoat. Although not an
emerging contaminant this chemical class was included in analyses because
PAHs remain an environmental concern. New levels or their interactions with
other chemicals could be a cause for concern and warranted investigation

5.	Pesticides, found in commercial, agricultural, and household products

6.	Pharmaceuticals, prescription and over the counter medications for humans and
animals such as anti-depressants, anti-seizure, blood thinners, and pain relievers

7.	Plasticizers and flame retardants, additive chemicals in products which produce
flexibility in plastics and reduce flammability in commercial and household
products such as carpets and upholstery

8.	Sterols, naturally occurring in plants and animals, and can be an indicator of
waste water

9.	Other, miscellaneous category for chemicals which did not fit the above
descriptions

Although CECs were detected in all waterbodies sampled, the concentrations and
individual chemicals detected differed substantially. Despite these differences, sampling
sites within each tributary tended to have similar chemical profiles (Choy et al. 2017,
Elliott et al. 2017). This indicates other drivers such as land use, human population,

96


-------
number or types of point sources, size of watershed, or other factors, are influencing the
presence and concentration of CECs within each waterbody.

- 1 0 25 50 (5 lOOiM

\ |
• mL25 50 75iPoKilometj

EXPLANATION
Land Cover
¦ Water
Z] Developed (open and low intensity)
| Developed (medium and high intensity)
| Forest, shrubland, herbaceous, and barren
I I Planted/cultivated
I I Wetland

	State/province boundary

	Great Lakes watershed boundary

@ Wild fish sampling site
^ Caged fish sampling site

Base composited from North American Atlas political boundaries, 1:10,000,000,2006;
North American Atlas hydrography, 1:10,000,000,2006; U.S. Geological Survey
hydrologic unit maps, 1:250,000,1994; Canada Land Inventory Level-I watershed maps,
1:2,000,000; National Land Cover Database 2011,30-m resolution, 2015. Universal
Transverse MercatorZone 16 North, referenced to North American Datum of 1983.

Figure D.2: Sampling locations for wild and caged fish. Wild and caged fish were both
sampled for at the following locations from west to east: Fox River, Little Lake Butte Des
Mortes, Chicago (North Shore Channel and Little Calumet), Detroit River, Clinton River,
Cuyahoga River, and Raquette River. Only wild fish were sampled at (from west to east):
St. Louis River, Milwaukee River, River Raisin, Swan Creek, Ashtabula River, Maumee
River, Genesee River, Irondequoit Bay, Long Pond, Raquette River.

Types of chemicals frequently detected in sediment varied from those frequently
detected in water. Pesticides, pharmaceuticals, and plasticizers/flame retardants were
detected more frequently in water than sediment. Commonly detected CECs in
sediment included diphenhydramine (pharmaceutical sleep aid found in common over-
the-counter medications), indole (a fragrance and fecal indicator), estrone (hormone),
and carbazole (pesticide). The St. Louis River (MN) had the greatest frequency of

97


-------
pharmaceuticals detected in water and sediment, while Irondequoit Bay (NY) had the
lowest frequency of pharmaceuticals detected. The Maumee River (OH), Fox River
(Wl), River Raisin (Ml), and Long Pond (NY) had the highest frequency of pesticides
detected in the water samples while the Detroit River (Ml) had the least. Detailed
explanations of findings and site descriptions can be found in Choy et al. (2017) and
Elliott et al. (2017).

Even though concentrations of detected CECs were low (detections were in ng/L or
parts per trillion), some exceeded concentrations previously shown to impact fish and
wildlife as did measured concentrations of some PAHs that exceeded water quality
benchmarks (Elliott et al. 2017). The highest detected concentrations of Metformin (a
diabetes medication) has been shown by previous studies to alter fish physiology which
can lead to population declines (Crago et al. 2016). Much of what is known about the
effects of CECs on fish and other wildlife are the result of laboratory studies. However, it
is still unclear how these chemicals impact fish and wildlife when they are present in
complex mixtures and under fluctuating environmental conditions. Therefore, the
impacts to aquatic life could be more (or less) severe in natural habitats than in
laboratory investigations. This knowledge gap requires further study.

The observations and findings described above are not unique to this study or the Great
Lakes Basin. CECs have been found in water and sediment across the globe, in
freshwater and marine systems (Herberer et al. 2002; Maruya et al. 2012; Pritz et al.
2014; Yager et al. 2014; Odendall et al. 2015; Morales-Caselles et al. 2017; Edwards et
al. 2018). Since CECs are widely found at detectable concentrations and many studies
show even low concentrations of CECs impact aquatic life, it is possible that the CECs
and the associated observed concentrations are negatively impacting fish and wildlife
residing in the Great Lakes' tributaries. CECs have also been detected in plant and
animal tissues (Maruya et al. 2016, Sengupta et al. 2016) raising the specter of food-
web interactions (reviewed in Nilsen et al. in press). The observed ubiquitous presence
of CECs justified the next step in this investigation: an assessment of the biological
effects of CECs in fish residing in Great Lakes Tributaries in order to provide the basis
for informed management decisions for the continuing benefit of Great Lakes natural
resources.

D.3.2 Biological Effects Are Subtle and Widespread

Fish Tumors in Native Fish

Assessment offish health began in 2010 with the collection of brown bullheads
(Ameiurus nebulosus), white sucker {Catostomus commersonii), and large or small
mouth bass (Micropterus salmoides or dolomieu), in federally designated Great Lakes
Areas of Concern (AOCs). Histopathological investigations revealed that fish in AOCs
had a higher prevalence of skin tumors than fish at reference sites. The prevalence of
liver tumors matched or was incrementally higher at AOC sites (Blazer et al. 2014b),
especially those in highly urbanized tributaries. Similar, subtle changes in tumor
prevalence were found in white suckers (Catostomus commersonii) collected from an
AOC in a Lake Superior tributary (St. Louis River) when compared to fish outside the

98


-------
AOC (Mazik et al. 2015; Blazer et al. 2014a). These histopathological findings provided
some evidence that pollutant exposure in Great Lakes tributaries has adverse effects on
exposed fish populations. Further evaluation of specific chemicals and comparison to
reference sites is still needed to draw concrete conclusions.

Biological Effects

To provide greater linkage between biological effects and CEC presence in Great Lakes
tributaries, a concerted effort was made in 2013 and 2014 to collect up to 40 sunfish
(.Lepomis spp.) at multiple sites in each of six Great Lake tributaries. These resident fish
collections were paired with two-week long caging of hatchery-reared bluegill sunfish
(Lepomis macrochirus) to provide a more controlled exposure scenario of known length
to a group of fish of similar age and exposure history. Resident as well as caged sunfish
were found to have increased blood glucose concentrations at sites with higher CEC
presence and concentrations (Figure. 6.3). The increased glucose concentrations are
likely the result of pollutant-induced metabolic stress, and were associated with
increased liver size and cellular changes in liver tissue (Thomas et al. 2017). These
indicators of pollutant exposure were inversely correlated (had opposite trends) with
indicators of reproductive potential. In fish with higher blood glucose concentrations and
altered liver anatomy, reproductive organs (testis, ovary) were smaller and less mature
(Thomas et al. 2017). These altered fish health parameters were more commonly
associated with sites containing greater CEC presence and concentrations (Thomas et
al. 2017).

-i	1	1	1	1	1	1	 	1	1	1	1	1	1	r

-3-2-10123	-3-2-10123

Figure D.3. Relative changes in vital functions in caged female (left) and male (right)
sunfish are associated with locations in Great Lakes tributaries that experience greater
(red) or lesser presence of Contaminants of Emerging Concern. Greater arrow length
indicates more pronounced effects. Arrows pointing similar directions indicate positive
associations while arrows pointing in opposite directions indicate negative associations. For
example, larger glucose values are associated with less maturity. Associations along the
horizonal axis are stronger than along the vertical axis. Comparisons are between fish
caged at sites in the study area.

99


-------
The analysis of CECs at sites where fish were collected and caged also suggested that
the presence of PAHs and pharmaceuticals can explain most of the variation in blood
glucose concentrations, altered liver anatomy, reduced reproductive organ size and
lesser maturity. These findings were consistent for resident and caged fish of both
sexes. Unfortunately, the data set was not sufficient to determine which classes of
pharmaceuticals (for example, antidepressants, antibiotics, NSAIDs) were most
responsible for the observed changes. However, the strength of a negative association
between blood glucose concentrations and reproductive endpoints (maturity and
ovary/testis size) suggested that there is an energetic cost associated with exposure to
CECs (likely as a result of the recruitment of detoxification pathways by the exposed
organisms) that may divert energetic resources from reproduction.

D.3.3 Effects of CECs Differ between Species

In addition to the resident and caged sunfish described above (3.2), largemouth bass
and white suckers were collected across 16 sites (Jorgenson et al. 2018). These fish
were analyzed for many of the same biological indicators of exposure as the sunfish
described above. Similar to findings in sunfish, liver size and liver structure were often
altered in fish living at sites downstream from point source pollution were more likely
contaminated by CECs. In addition, this analysis identified different response patterns in
different fish species. Patterns of biological response to CEC exposure clustered around
three biological processes (Figure D.4): (i) organismal health as assessed by the overall
nutritional status of the fish (condition factor) and the presence of pathologies such as
parasite infestation; (ii) liver health as indicated by plasma glucose concentrations,
changes to liver cellular structure and liver size; and (iii) reproductive health as indicated
by differences between sites in the size and maturity of the reproductive organs as well
as the presence of the egg yolk pre-cursor protein vitellogenin in male fish. Changes in
organismal health indices were more closely associated with female sunfish. Changes
in liver health indices were more closely associated with female largemouth bass. Male
white suckers were also closely associated with changes in liver health, while both
sexes of white suckers were found to have close associations between CEC presence
and changes in reproductive health indices (Jorgenson et al. 2018). These results
suggest that subtle changes in fish health could be a response to CEC exposure may
differ between species. Other factors such as chemicals not measured, altered stream
characteristics, and habitat could also be factors contributing towards the effects seen.

100


-------


liver
function

reproductive
effects

whole
animal
effects

Raquette Middle- WHS
Fox Upstream- WHS
Fox Middle- WHS
Raquette Upstream- WHS
Fox Middle- 1MB
Fox Downstream- LMB
Fox Upstream- LMB
Fox Middle- SF
Fox Downstream- SF
Cuyahoga Downstream- SF
Cuyahoga Upstream- LMB
I Raquette Middle- LMB

Raquette Downstream- LMB
| Raquette Upstream- LMB
Cuyahoga Downstream- WHS
Cuyahoga Middle- WHS
Cuyahoga Upstream- SF
Fox Upstream- SF
Raquette Upstream- SF
Raquette Downstream- SF
Raquette Middle- SF





uuholo

«1 IWIC

animal

effects









1







¦



pathology

condition

Fox Downstream- LMB
Fox Middle- LMB
Raquette Downstream- LMB
Cuyahoga Upstream- LMB
Fox Middle- SF
Fox Upstream- SF
Raquette Downstream- SF
Raquette Upstream- SF
Cuyahoga Downstream- SF
Raquette Middle- SF
Cuyahoga Upstream- SF
Raquette Upstream- LMB
Raquette Middle- LMB
Fox Upstream- WHS
Raquette Upstream- WHS
Raquette Middle- WHS
Fox Middle- WHS
Cuyahoga Downstream- WHS
Cuyahoga Middle- WHS
Fox Downstream- WHS
Fox Upstream- LMB
Fox Downstream- SF

Figure D.4. Clustering of adverse effects in resident white suckers (WHS), largemouth bass
(LMB) and sunfish (SF). Darker red colors indicate stronger effects. Results from female
fish (left) and male fish (right).

D.3.4 Some CECs Occur at Concentrations Predicted to Cause Biological Effects

As a result of detecting CECs in in water and sediment, and finding that CECs may be a
contributing factor impact the biology and physiology of fish living in areas polluted by
CECs, screening values were developed as a tool for natural resource managers.

These screening values will help managers evaluate the water at their locations to
better understand if fish in those locations are at risk for impacts by CECs. Effect-
specific pairs of mean Screening Values (SV) were developed for 14 CECs in water for
which there was adequate peer reviewed literature available (Figure D.5). Screening
values are estimated threshold concentrations of chemicals that define our expectations
about adverse effects in target biota. A set of Comprehensive SVhigh and SVlow values
was developed for multiple effect categories using all adverse effects reported in peer
reviewed literature for a given CEC. A subset of adverse effects was used to derive a
set of Population-relevant SVhigh and SVlow values, to focus Ecological Hazard
Assessments (EHA) on the potential for population-level impacts. The SVlow is the
threshold concentration of a CEC at or below which it is unlikely to cause an effect to
freshwater fish, while an SVhigh is the threshold concentration above which it is likely
that an adverse effect has occurred. For values that fall between these two thresholds, it
is uncertain if an effect will occur.

101


-------
SVs for each of the 14 chemicals were developed for five population-relevant biological
effect categories (Gefell et al. 2019a):

1.	Mortality

2.	Reproductive

3.	Developmental

4.	Behavioral

5.	Growth.

Seven additional 'comprehensive' effect categories were evaluated for potential impacts
to fish health.

An EHA was then conducted for the 24 waterbodies sampled for CECs in 2010-2014
(Figure D.1), using the CEC concentrations detected in the water samples. The primary
purposes of the EHA were to determine a location's relative hazard from CECs, and
then rank CECs and sampled sites in terms of potential for direct impacts to fisheries
from water-borne CECs. It was evaluated whether ecological hazard to fish due to CEC
exposure could be discerned in the dataset, and if so, what is the nature and extent of
that hazard. For each exposure data point, a set of hazard scores was developed from
simple comparisons of CEC concentration to the corresponding pair of CEC-specific
SVs. A hazard score of 1 was assigned where the individual CEC concentration was
less than the SVlow, a score of 3 was assigned where the CEC concentration exceeded
the SVhigh, and CEC concentrations that fell between the SVlow and SVhigh received a
score of 2. At each project location an exceedance of SVlow was observed for at least
one of the CECs, in one of the effect categories (Gefell et al. 2019b), while an SVhigh
exceedance in at least one effect category was observed in 17/24 locations. These
results indicate environmental concentrations likely impact native fish health, but much
is unknown and more information is needed to fully evaluate the impacts of CECs on
native fish communities. Detailed information for each of the 14 chemicals and
associated screening values and EHA can be found in Gefell et al. (2019a and 2019b).

102


-------
Figure D.5. Comprehensive (ellipse) and Population-relevant (rectangle) Mean screening
values (high=solid) for fourteen Contaminants of Emerging Concern (CEC) commonly
detected in water samples in tributaries to the Laurentian Great Lakes. Measured detected
concentrations for each CEC in water samples (Choy et al. 2017) are indicated by the
horizontal colored columns. Red rectangles/ ellipses indicate that environmental
concentrations exceeded screening values. Percent detection for each CEC (out of 150
water samples) is indicated in the right hand side column. Black outlined hollow bars for
each chemical indicate range of concentrations measured by Bradley et al. 2017 in 38 US
streams.

D.4. Management Implications

Since CECs are ubiquitous throughout Great Lakes tributaries, and fish are exhibiting
effects associated with exposure to CECs, it is important for fisheries and natural
resource managers to be aware of CECs. Natural resource managers need to have a
good understanding of where and what CECs are found in their management areas,
and how those CECs could impact fish species and populations in order to make sound
management decisions. Understanding and accounting for CECs in the environment is
one more piece of the puzzle in natural resource management. Possible considerations
for natural resource management could include:

103


-------
•	Surveying for CECs before beginning restoration or mitigation projects to understand
how the CECs in the project area may impede the success of the project and
applying screening values to the samples taken to identify potential risk.

•	Prioritizing restoration, mitigation, or release sites with lower occurrence and
presence of CECs to reduce the risk of impacts and ensure success to those
projects.

•	Surveying/monitoring for CECs if incidences of concern arise, such as altered age
structure or decreases in populations, especially when all other environmental
factors are favorable.

•	Working together with state, federal, local, or tribal agencies, private land owners or
industries to spread awareness of the impacts CECs have on the environment.

•	Education campaigns on improving disposal and use of CECs to reduce presence in
aquatic systems.

•	Providing information to WWTPs or other industries to help them formulate
wastewater treatment processes to reduce loading of CECs into the environment.

•	Working with USFWS National Wildlife Refuges, Private Lands Division, Fish
Hatcheries, Fish and Wildlife Conservation Offices, Fish Health Centers, to name a
few, to help prioritize areas for projects and best use of restoration efforts.

Best management practices and strategies need to be adaptive and change as new
information arises. These suggestions are based on current information and may be
revised as new results and information become available. However, it is important for
natural resource managers to understand that CECs are present in most water and
sediments in Great Lakes tributaries, and have the potential to negatively impact natural
resources. As this team continues to investigate and assess those impacts and risks to
aquatic life, more information on how to adjust best management strategies will become
available.

D.5. Knowledge Gaps

•	Is the potential for severe CEC-related impacts at measured CEC exposure levels
masked in current fish populations and communities? That is, have the most
dramatic CEC-related impacts to resident fish communities already occurred in
waterbodies that have received long-term CEC inputs, leaving only the most CEC-
tolerant species and strains within species to investigate in our current field
assessments?

•	Can the composition of CEC mixtures be predicted based on land-use
characteristics?

104


-------
•	If fish are exposed to complex mixtures of CECs, what are the aggregate biological
effects of CEC mixtures and are mixture effects predictable from individual
compound exposures?

•	As CECs occur in Great Lakes tributaries year-round, what are the aggregate life-
time exposure effects?

•	As CECs co-occur with other environmental stressors (such as nutrient pollution
and changing water temperature), how do these multiple stressors interact to affect
aquatic biota?

•	Can aggregate effects of life-time and complex mixture exposures be extrapolated
to population level consequences for exposed populations of fishes?

•	Can the aggregate life-time complex CEC mixture exposure effects be extrapolated
to species that are difficult to study directly (i.e. listed species)?

D.6. Disclaimer

The findings and conclusions in this report are those of the author(s) and do not
necessarily represent the views of the U.S. Fish and Wildlife Service or the U.S.
Environmental Protection Agency

D.7. Acknowledgements

Funding was provided from the Great Lakes Restoration Initiative through the U.S. Fish
and Wildlife Service's Contaminants of Emerging Concern Team. We'd like to thank all
of those who assisted with field work, reviewing of this appendix and reviewing other
associated works. This assistance was instrumental in the successful completion of this
work. Several generations of graduate and undergraduate students in the St. Cloud
State University Aquatic Toxicology Laboratory assisted with field work logistics,
sampling, and analysis along with US Geological Survey technicians: Michael Menheer,
Jeffrey Zeigweid, and Molly McCool.

D.8. References

Blazer VS, Hoffman J, Walsh HL, Braham RP, Hahn C, Collins P, Jorgenson Z, Ledder
T. 2014a. Health of white sucker within the St. Louis River area of concern
associated with habitat usage as assessed using stable isotopes. Ecotoxicol 23:236-
251.

Blazer VS, Mazik PM, Iwanowicz LR, Braham RP, Hahn CM, Walsh HL, Sperry AJ.
2014b. Assessment of the fish tumor beneficial use impairment in brown bullhead
(Ameiurus nebulosus) at selected Great Lakes Areas of Concern: US Geological
Survey Open-File Report 2014-1105, 17 p. http://dx.doi.Org/10.3133/ofr20141105.

Blazer VS, Mazik PM, Iwanowicz LR, Braham R, Hahn C, Walsh HL, Sperry AJ. 2014c.
Monitoring of Wild Fish Health at Selected Sites in the Great Lakes Basin—Methods

105


-------
and Preliminary Results: US Geological Survey Open-File Report 2014-1027 p. 31,
http://dx. doi. org/10.3133/ofr20141027.

Choy SJ, Annis ML, Banda JA, Bowman SR, Brigham ME, Elliott SE, Gefell DJ,

Jankowski MD, Jorgenson ZG, Lee KE, Moore JN, Tucker WA. 2017. Contaminants
of emerging concern in the Great Lakes Basin: A report on sediment, water, and fish
tissue chemistry collected in 2010-2012. US Fish and Wildlife Service Biological
Technical Publication. 2017; BTP-R3001-2017:1-80.

Crago J, Bui C, Grewal S, Schlenk D. 2016. Age-dependent effects in fathead minnows
from the anti-diabetic drug metformin. Gen Comp Endocrinol 232:185-190.

Edwards QA, Sultana T, Kulikov SM, Garner-O'Neale LD, Yargeau V, Metcalfe CD.
2018. Contaminants of emerging concern in wastewater in Barbados, West Indies.
Bull Environ Cont Toxicol 101:1-6.

Elliott SM, Brigham ME, Kiesling RL, Schoenfuss HL, Jorgenson ZG. 2018.

Environmentally Relevant Chemical Mixtures of Concern in U.S. Tributaries to the
Great Lakes. Integrat Environ Assess Manag DOI: 10.1002/ieam.4041.

Elliott SM, Brigham ME, Lee KE, Banda JA, Choy SJ, Gefell DJ, MinarikTA, Moore JN,
Jorgenson ZG. 2017. Contaminants of emerging concern in tributaries to the
Laurentian Great Lakes: I. Patterns of occurrence. PLoS ONE 12(9): e0182868.
DOI: 10.1371 /journal, pone.0182868.

Elliott, S.M. Lee, K.E. Brigham, M.E. Langer, S.K. Trombley, M.M. Menheer, M.A.
Furlong, E.T. Foreman, W.T. Jorgenson, Z.G. Choy, S.J. Moore, J.N. Banda, J.B.
and Gefell, D.J. 2015, Chemicals of Emerging Concern in Water and Bottom
Sediment in Great Lakes Areas of Concern, 2013—Analytical Methods, Collection
Methods, Environmental Data, and Quality Assurance: U.S. Geological Survey Data
Release, http://dx.doi.Org/10.5066/F7DF6P9D.

Elliott, S.M. Lee, K.E. Brigham, M.E. Langer, S.K. McCool, M.M. Menheer, M.A.

Furlong, E.T. Foreman, W.T. Jorgenson, Z.G. Choy, S.J. Moore, J.N. Banda, J.A.
Gefell, D.J. and Schoenfuss, H.L. 2016, Chemicals of Emerging Concern in Water
and Bottom Sediment in the Great Lakes Basin, 2014 - Analytical Methods,

Collection Methods, Quality-Assurance Analyses, and Data: U.S. Geological Survey
data release, http://dx.doi.org/10.5066/F7TH8JS6.

Gefell, D.J., J. Banda, J.N. Moore, A.L. Secord, W.A. Tucker. 2019a. Ecological Hazard
Assessment of Contaminants of Emerging Concern in the U.S. Great Lakes Basin:
Part B - Supplemental Document: Technical Resources for Ecological Hazard
Assessments of Contaminants of Emerging Concern in Freshwater Fish. In Print.

Gefell, D.J., M. Annis, J. Banda, A. Bellamy, S. Choy, S. Hummel, Z. Jorgenson, J.N.
Moore, A.L. Secord, W.A. Tucker. 2019b. Ecological Hazard Assessment of
Contaminants of Emerging Concern in the U.S. Great Lakes Basin: Part A -

106


-------
Screening Assessment of Relative Hazard to Fish from Surface Water Exposures to
Fourteen Contaminants of Emerging Concern in the Great Lakes Basin. In Press.

Heberer T, Reddersen K, Mechlinski A. 2002. From municipal sewage to drinking water:
fate and removal of pharmaceutical residues in the aquatic environment in urban
areas. Water Sci Technol 46(3):81-88.

Hyland KC, Blain AC, Higgins CP. 2015. Accumulation of contaminants of emerging
concern in food crops - part 2: Plant Distribution. Environ Toxicol Chem 43:2222-
2230.

Jorgenson ZG, Thomas L, Elliott SM, Cavallin JE, Randolph EC, Choy SJ, Alvarez DA,
Banda JA, Gefell DJ, Lee KE, Furlong ET, Schoenfuss HL. 2018. Contaminants of
Emerging Concern Presence and Adverse Effects in Fish: A Case Study in the
Laurentian Great Lakes. Environ Poll 236:718-733. DOI:
10.1016/j.envpol.2018.01.070.

Lee, K.E. Langer, S.K. Menheer, M.A. Foreman, W.T. Furlong, E.T. and Smith, S.G.
2012, Chemicals of emerging concern in water and bottom sediment in Great Lakes
areas of concern, 2010 to 2011 —Collection methods, analyses methods, quality
assurance, and data: U.S. Geological Survey Data Series 723, 26 p.

Maruya KA, Dodder NG, Sengupta A, et al. 2016. Multimedia screening of contaminants
of emerging concern (CECS) in coastal urban watersheds in southern California
(USA). Environ Toxicol Chem 35(8): 1986-1994. doi:10.1002/etc.3348.

Maruya KA, Vidal-Dorsch DE, Bay SM, Kwon JW, Xia K, Armbrust KL. 2012. Organic
contaminants of emerging concern in sediment and flatfish collected near outfalls
discharging treated wastewater effluent to the southern California bight. Environ
Toxicol Chem 31(12):2683-2688.

Mazik PM, Braham RP, Hahn CM, Blazer VS. 2015. Assessment of general health of
fishes collected at selected sites within the Great Lakes Basin in 2012. Department
of Interior, Fish and Wildlife Service, Cooperator Science Series FWS/CSS-112-
2015, Washington, D.C.

Morales-Caselles C, Desforges J-PW, Dangerfield N, Ross PS. A risk-Based

characterization of sediment contamination by legacy and emergent contaminants of
concern in coastal British Columbia, Canada. Archives of environmental
contamination and toxicology. 2017;73(2):270-284. doi:10.1007/s00244-017-0403-z.

Munkittrick KR. 1992. A review and evaluation of study design considerations for site-
specifically assessing the health offish populations. J Aquat Ecosys Health 1:283-
293.

Nilsen E, Smalling KL, Ahrens L, Gros M, Miglioranza KSB, Pico Y, Schoenfuss HL. In
press. Critical review: grand challenges in assessing the adverse effects of
contaminants of emerging concern on aquatic food webs. Environ Toxicol Chem.

107


-------
Odendaal C, Seaman MT, Kemp G, Patterton HE, Patterton HG. 2015. An LC-MS/MS
based survey of contaminants of emerging concern in drinking water in South Africa.
S Afr J Sci 111(9/10); 1-6.

Pendall R, Poethig E, Treskon M, Blumenthal E. 2017. The future of the Great Lakes
Region. Urban Institute, Washington, DC, 51pp.

https://www.urban.org/sites/default/files/publication/89087/great lakes O.pdf

Pritz CMF, Schrlau JE, Simonich SLM, Blett TF. 2014. Contaminants of emerging
concern in fish from Western U.S. and Alaskan national parks - spatial distribution
and health thresholds. J Amer Water Res Assoc 50:309-323.

Sengupta A, Lyons JM, Smith DJ, Drewes SJ, Snyder SA, Heil A, Maruya KA. 2014.
The occurrence and fate of chemicals of emerging concern in coastal urban rivers
receiving discharge of treated municipal wastewater effluent. Environ Toxicol Chem
33(2):350-358. doi: 10.1002/etc.2457

Thomas LM, Jorgenson ZG, Brigham ME, Choy SJ, Moore JN, Banda JA, Gefell DJ,
MinarikTA, Schoenfuss HL. 2017. Contaminants of emerging concern in tributaries
to the Laurentian Great Lakes: II. Biological consequences of exposure. PLoS ONE
12(9): e0184725. DOI: 10.1371 /journal.pone.0184725

U.S. Fish and Wildlife Service, National Wildlife Refuge System, Midwest Region, 2017.
Retrieved from: https://www.fws.gOv/midwest/refuges/visit.htmlU.S. Fish and Wildlife
Service, Endangered Species, 2018a. Retrieved from:

https://www.fws.g0v/midwest/endangered/lists/e_th_pr.htmlU.S. Fish and Wildlife
Service, Fisheries, Midwest Region, 2018b. Retrieved from:
https://www.fws.g0v/midwest/fisheries/#YagerTJB, Furlong ET, Kolpin DW, Kinney
CA, Zaugg SD, Burkhardt MR. 2014. Dissipation of contaminants of emerging
concern in biosolids applied to non-irrigated farmland in Eastern Colorado. J Amer
Water Res Assoc 50(2):343-357. doi: 10.1111 /jawr. 12163.

108


-------
Appendix F

Transcriptional Effects-Based
Monitoring of Contaminants of Emerging
Concern in Great Lakes Tributaries:
Biological Effects and Chemical Specific
Impacts

Lead Organization: U.S. Army Corps of Engineers Research and Development Center,
Vicksburg, MS, USA

Contributing Authors: Ed Perkins, Natalia Garcia-Reyero, and Lyle Burgoon.

Contributing Investigators: Ed Perkins, Natalia Garcia-Reyero, Lyle Burgoon, J. Erik
Mylroi, and Mitch Wilbanks

Corresponding Author Contact: edward.j.perkins@erdc.dren.mil
F.1. Problem Statement and Scope

Increasing numbers of chemicals and other contaminants are detected in waters,
sediment and animals in Great Lakes watershed many of which have little to no
toxicological or water quality information is available. Given that these contaminants
appear as complex mixtures, available information for individual chemicals may under
estimate the effects of contaminant exposure in Great Lakes Tributaries. This makes it
difficult for managers and other decision makers to identify which contaminants
represent threats to the use of their resources, watersheds or waterways. Here we
discuss approaches using effects-based monitoring with transcriptomics, AOPs and
water quality assessments to identify the presence and deleterious effects of CECs so
that they can be appropriately managed. As part of the Great Lakes Restoration
Initiative (GLRI) Action Plan I, our goals were to develop methods and tools using
untargeted transcriptomics to (1) Detect biological activities associated with mixtures of
CECs in Great Lakes Tributaries, (2) Identify what contaminants could be potentially
driving adverse effects or impairments associated with exposure to Great Lakes
tributaries, and (3) Link transcriptional changes as a result of exposure in Great Lakes
tributaries to apical adverse effects and impairments.

F.2. Detecting Biological Activities Associated with Mixtures of CECs in Great
Lakes Tributaries using Transcriptomics

F.2.1. Strategy - In Situ Exposure of Model Organisms

141


-------
In this appendix, we are principally concerned with determining the potential of CECs in
Great Lakes tributaries to cause harmful effects in fish living in the water. However, it is
difficult to know what chemical or other stressors native fish might have experienced,
when they were exposed and where they were exposed. To overcome this challenge,
our strategy is to use laboratory raised fathead minnow placed in cages and then
deployed into water bodies for 2, 4 or 8 days as a surrogate for native fish (see
Appendix E Kahl et al. 2014). Caged fish were used to examine the effects of CECs at
14 different sites and three different tributaries (Figure F.1). Fathead minnows are easily
obtained small fish commonly used in ecotoxicology that are widely distributed in
freshwater lakes and streams across much of North America (Ankley and Villeneuve
2006). Using laboratory reared fish in cages has several advantages including control of
the exposure period and location, simplicity, and low cost of deployment that can be
used by managers. As discussed in Appendix E, this strategy can also be used with
other fish species.

F.2.2. Strategy - Laboratory Exposure of Model Organisms

The impact of many CECs on gene expression model organisms such as Fathead
minnow are often limited or unknown. We used the strategy of laboratory exposures of
fish to understand the potential contribution of individual CECs to the overall effects of
mixtures of CECs on fish. Laboratory exposures provide a high level of control over the
experimental conditions including known length of exposure, well characterized
chemistry, and minimization of other stressors that might confound results.

F.2.3. Endpoints and Analyses

The biological endpoints that we use are changes in gene expression measured using
transcriptomics. Our goal in using untargeted transcriptomics, or global gene expression
analysis, was to capture as many changes or adverse effects as possible in exposed
fish rather than focus on one or a few specific changes such as a change in vitellogenin
protein. Transcriptomics is the simultaneous measurement of messenger RNAs from
large numbers of different genes in a cell and is used to examine gene expression
changes (step 4 below). Since we are using tools to capture changes in a very broad
range of genes rather than a small number of specific genes, it is considered an
untargeted approach and can capture many different biological effects. To detect effects
of chemical exposure we looked at gene expression two different organs after fish are
exposed: Ovaries of females were tested because ovaries are reproductive organs and
the site of estrogen synthesis. Livers of males were tested as metabolism and toxicity of
CECs such as PAHS or pharmaceuticals can often be seen in livers. Male livers are
also interesting to look at since estrogenic effects are more easily observed in male
livers.

142


-------
Box F.l. The way chemicals can cause changes in gene
expression that lead to toxicity or adverse effects can
be summed up as:

Since expressed genes do
not always directly translate
to changes in proteins or,
ultimately, adverse effects,
we use the expression of a
collection of genes belonging
to different biological
pathways (e.g., a sequence
of enzymatic steps resulting
in production of a chemical or
physical change) to indicate
potential activities and
functions that have changed,
or been impaired, in exposed
fish. To do this, we examine
how many genes involved in
different biological functions
or pathways changed relative
to controls (gene enrichment
or enriched biological
pathways). We also used
chemical gene regulator
relationships for the same
genes in humans to
understand what chemicals
might be present based on
gene expression patterns
(Kramer et al., 2014). In addition to gene expression, we used physiological and
biochemical endpoints for the caged fish including plasma hormone (17(3-estradiol and
testosterone) and vitellogenin levels and gene expression levels of the PAH responsive
cyplal (see Appendix E for details). We compared gene expression in fish tissues to
types and levels of CECs detected in water samples obtained via grab sampling. Water
was tested for 134 organic compounds indicative of industrial, domestic, or agricultural
wastewaters, including a suite of 48 pharmaceuticals (Lee et al. 2012).

1.

2.

3.

4.

5.

6.

7.

A chemical enters biological system.

A chemical interacts with a receptor - typically some
type of protein.

This receptor becomes activated or inactivated by
chemical interaction activating or deactivating signals
within cells.

Activation or deactivation of signals leads to changes
in gene expression, which are changes in the amounts
of specific messenger RNAs within the cells.

Some of these changes in RNAs are translated into
changes in the amount and type of proteins that are
generated within the cell

Changes in the amount and type of proteins results in
new signals that cause changes to cells
These changes can include causing some cells to
divide and produce more cells (tumors develop when
cells divide out of control, making too many cells),
activate or inactivate the immune system, cause cells
to die (which can also activate the immune system),
cause cells to change how they use energy which
changes the amount of energy molecules available for
the organism to use, or impair reproduction.

143


-------
Figure F.l. Locations of case studies conducted as part of Action Plan I and corresponding study sites. Size of each marker
reflects the number of caged fish exposures conducted at each site. Map surface layer credits: ESRI, HERE, Garmin,
OpenStreetMap Contributors, and the GIS user community. Map created by Kendra Dean (10/24/18). This figure from
Appendix E. Final figure under development.

144


-------
F.2.4. Case Study: Effects-Based Monitoring of Waste Water Treatment Plant
Effluent at Increasing Distances from Point of Discharge

Male fathead minnows were deployed at varying distances from a WWTP effluent outfall
to determine if biological effects associated with the effluent diminished with distance
from the point of discharge (Garcia-Reyero, unpublished). Gene expression was
examined in livers of male exposed for 4d at each site. The impact of exposure site on
gene expression decreased with distance from the effluent discharge consistent with a
reduction in the number of chemicals detected in water samples farther from the
discharge site. Males exposed closest to the discharge site had enriched biological
pathways similar to those induced by 17(3-estradiol, while males at the distant site
(250m) had fewer pathways similar to 17(3-estradiol. The similarity of male fish deployed
near the outfall to males exposed to 17(3-estradiol was consistent with increased plasma
17(3-estradiol in males at that site. Enrichment of pathways in males deployed distant
from the discharge site indicated the presence of stressors different from those near the
discharge site. Analysis of gene expression indicated the potential gene regulator
effects of several dibenzofurans, polycyclic aromatic hydrocarbons (PAHs), and
polychlorinated biphenyls (PCBs). Overall the estrogenic effect of effluent reduced with
distance from the discharge site. At farther distances, other CECs, such as PAHs and
PCBs, likely drove the transcriptional effects seen in deployed fish.

F.2.5. Key Findings/Progress - Detecting Biological Activities with
Transcriptomics

•	Transcriptional-effects based monitoring can be used to survey tributaries for
chemical effects on wildlife including estrogenic and effects caused by other
CECs.

•	Transcriptomics can be used to identify potential chemicals causing effects even
if they were not looked for in analytical tests of the exposure water.

•	The impact of specific CECs in mixtures and surface water can be assessed by
comparing laboratory exposures of a target CEC to mixture effects.

F.3. Associating Detected Chemicals with Biological Effects Using
Transcriptomics and Evidence-Based Approaches

The goal of this section was to determine if changes in gene expression in exposed fish
can be related to the concentration of specific chemicals detected in the water they
were exposed to. Many different chemicals and CECs can be detected in surface
waters especially those into which waste water treatment plants discharge. It would be
of great use to identify, or reduce the number of, contaminants that could potentially
drive adverse effects so that managers and other decision makers can prioritize efforts
and best utilize limited resources.

F.3.1. Statistical Approaches for Associating Chemicals with Transcriptomics
and Biological Effects

145


-------
Here we used the knowledge of how chemicals may cause biological effects by
activation, or deactivation, of signals by chemicals can lead to changes in gene
expression. In this approach we counted the number of genes whose expression level
patterns were highly similar to (correlated with) the concentration levels of each
chemical detected in surrounding water (Perkins et al., 2017). We then looked at which
chemicals had the largest number of genes whose expression correlated with chemical
concentration. Exposure to increasing concentrations of chemical has generally been
found to result in increasing numbers of changed genes (e.g. Deng et al 2011; Webster
et al 2015) indicating that chemicals with high numbers of correlating genes are likely to
have had greater biological impact than those with low numbers of correlating genes.
While several factors can change this relationship (e.g. length of exposure, potency,
mechanism of action and thresholds of toxicity), the number of genes affected by a
chemical provides an easily measured indicator of the impact of a chemical that is
suitable for screening and prioritization.

F.3.2. Evidence-Based Approaches for Associating Chemicals with Biological
Effects

Evidence-based approaches utilize existing evidence, or data, that a chemical causes
biological effects to extrapolate what effect that chemical would have on wildlife
exposed to it in surface water. The evidence is generally based on published literature,
toxicological assays (e.g. ToxCast https://www.epa.gov/chemical-research/toxcast-
dashboard), chemical gene/protein relationships or other health hazard data. Here we
utilized two different evidence-based approaches, one that uses evidence of a
chemical's known impact on gene expression, and a second that uses known chemical
effects on a fish reproductive biomarker, vitellogenin and bioassay data related to tumor
formation.

F.3.2.1. Single Chemical Gene Expression Effects

Exposure of a fish to an individual CEC can establish what kind of genes and biological
pathways are changed in response to that chemical. This can also be used to establish
a gene signature characteristic of exposure to the chemical. Known signatures and
biological impacts of individual CECs have the potential to be used to determine if that
CEC is causing biological effects when present in a mixture (see 8.3.4).

F.3.2.2. Causal Networks Linking Gene Expression and Chemicals

Knowledge on what genes are expressed in response to a specific chemical exposure
can be useful in identifying chemical(s) that an animal may have been exposed to.
Causal networks linking drug and chemical interactions to gene expression and apical
effects in humans that enables prediction of potential upstream regulators of gene
expression can be useful in identifying conserved chemical gene interactions conserved
across species (Kramer et al., 2014). Vertebrates have many biological and regulatory
pathways in common, so causal networks such as the Ingenuity Knowledge Base
(QIAGEN Bioinformatics, Redwood City, CA) can be used to identify chemicals
potentially impacting exposed fathead minnows.

146


-------
F.3.2.3. Hazard Quotients

A rapid approach to estimate the potential of contaminated water to cause adverse
effects on wildlife is to divide the exposure amount (concentration in water) by a
reference concentration (a known concentration where the chemical can cause toxicity
or activity in an assay). One advantage of this approach is that results from cell-based
assays representing biological pathways or outcomes of interest (e.g. estrogen receptor
activation, cell proliferation) can be used to establish toxicity reference values based on
the point of departure from control assays. A hazard quotient approach was used as it is
a well establish approach that allows incorporation of uncertainty values for
extrapolating from human cell-based assays to fish. Risk assessors have long used
hazard quotients to described the potential of different media to cause toxic effects
(EPA 1989, EPA 2009, 2018). A hazard quotient (HQ) is derived by dividing the
exposure concentration by a toxicity reference value. When an HQ>1 is present,
enough chemical is present to potentially cause an effect. Since we are using a toxicity
reference value based on human in vitro data, the toxicity reference value was divided
by an uncertainty factor of 100 to account for extrapolation of species (human to fish)
and for cell assays to animal (in vitro to in vivo). A hazard Index is the sum of HQs for
chemicals having similar modes of action.

F.3.3. Case Study: Transcriptomics of Bisphenol A Laboratory Exposures to
Assess Bisphenol A Contribution to Estrogenicity in Waste Water Treatment
Plant Effluent

A major concern for CECs is the potential to cause reproductive effects by increasing
estrogenicity of surface waters. Chemical monitoring only provides a partial picture of
potential risk posed by complex mixtures of contaminants in aquatic systems. Here we
used transcriptomics to examine the impacts of effluent on fathead minnows in the
laboratory in comparison to a CEC (Bisphenol A) thought to cause estrogenic effects in
the effluent (Garcia-Reyero unpublished). Estrogenic effects have been observed in
male fish exposed to treated wastewater that discharges to the St Louis Bay. Bisphenol
A (BPA), a known estrogenic chemical, was among the contaminants detected. To
determine whether BPA was a major contributor to estrogenic effects, male and female
adult fathead minnow were exposed in the laboratory to effluent collected directly from
the discharge or BPA. Exposure to 50% effluent, but not BPA, had an estrogenic effect,
increased levels of vitellogenin, on male fish. Both BPA and effluent exposures caused
significant changes in liver transcriptomics. However, the gene signature of BPA derived
from laboratory exposed fish only partially overlapped with the effects on
transcriptomics caused by 50% effluent. The increase of vitellogenin by effluent, but not
by BPA, and the partial overlap of the gene signatures of BPA and effluent indicates
that BPA was not likey to be a major cause of estrogenicity in the effluent.

F.3.4. Case Study: Integrated Application of Transcriptomics and Evidence-Based
Approaches Assessing the Estrogenic Impact of CECs

We applied a combination of HQ and statistical approaches with caged female fish to
identify potential CECs causing estrogenic effects near Waste Water Treatment plants

147


-------
(WWTP) in the St louis Bay area (Perkins et al., 2017). Analysis of the water chemistry
near discharge sites identified 4 chemicals (naphthalene, 4-nonylphenol, BPA, estrone)
at concentrations known to cause any biological effect (HQ>1). Three of these
(naphthalene, 4-nonylphenol and BPA) were the most highly ranked/impactful
chemicals by covariation with gene expression in ovaries, but only BPA also had an
HQ>1 for vitellogenin production in male fathead minnows (Figure F.2). Causal network
analysis of transcriptomics of ovaries from exposed females indicated the effects of
several chemicals associated with hospital waste consistent with WWTP having
untreated hospital waste delivered to its system. Overall, the combination of HQ and
statistical association of detected chemicals with transcriptomics and causal network
analysis provides evidence that BPA is a significant contributor to estrogenic effects in
the St Louis Bay. However, it also indicates that other CECs may potentially be causing
biological effects such as PAHs and chemicals in hospital waste.

148


-------
Chemicals detected across sampling sites

"O "< ftj "D

3 °

il

o o

Q_ CL>

O 3

f - £

Iff

t) a h
3 SI j J TJ

n- O (t> m S 5

^ S: 9 r\ 3

V 1 W S- O -< „

o- 5

- o o 9 "?• -2 O c«
Oc*. » < a 0 S o <
3 a> qj o — ^

ST a?

o
x

£ *

5 -<"

2 g- o

O 3

ro g

C"Q.y,2dS5 ^X§ &> 3 Q-O 33

n 2. 2l * 2. n * 2. * 3 
3
3*

^ %
« 3
o 3
£ 3'
3 ro

n

o
a
-<_
"O

cr 3-
<0 ro

? no

O	i/> fD ~1"

t	?. 3 a
o

o	*

x
<

3 Q.

r«j —•

o 2.

"O 3-

 H*

s k

_ O

0> ix> 3 - Z1

< s - £*

9- 9 3 o 3"

" Is 13
|o« JS I-

SiJjlTD-oS! 01 D „
; " 5 t S ? ° o o A 3- - ~

o=o SL a ' < ^'S.jgojs.a.®
H=:ii05r»«!. SJJin U 5.

3Q.33™3?i3";SS3 3 3
rtot(t'D'(ioni~'3— n 

?¦ c: sr

oj <
"O ro

3" 3

n

3"

O

n —

O) T3
3 -C

O
"O

3 "S.

N	*—*

o, "D

3"

C

S .? s

OJ

o

3

fD —

-< C T5
2. rt> ri> g"
333"
*I> fD ft) ro

7* T>
3" 3
rt> O

3
> °

re Si nj



































1

l

l

l

l



1



2

2

2





3

3





4







ll



i



3







u

E



































X

1

1

l

J.

u

1

J

2

2

2

2

2

2



3

3

3

4







H













3

u

E



































l

i

l

l

l



1





2



2





3

3

3



4











3
3

u

u





































i









1

1





2













3



4







n















¦























l

l

l

l

l



1

1

2

2

2

2

2



3

3

3

3









i





H



161

13

u

u



































l











1





2











3

3













'i













u



1









¦















u





l











1





2













3





35

39

1

n











~

~



WLSSD-1 2d
WLSSD-1 4d
WLSSD-2 2d
WLSSD-2 4d
SMTP 2d
SMTP 4d
SMTP 8d

~ Analytical support [ Analytical and CTD support jH] Analytical and CLR support	¦ Analytical, CLR and CTD support

Greater than the chemical concentration u = Predicted upstream chemical regulator	E = Increased plasma estradiol levels in

needed to cause an in vitro estrogenic effect	caged male fish

n Greater than the chemical concentration	r"g~l Numberof genes associated

1—1 needed to cause a biological effect	1—1 with chemical by CLR

Figure F.2. Prioritization of chemicals detected near waste water treatment plants in St Louis bay. Chemicals with more
evidence supporting biological impacts are on the right and chemicals with the least supporting evidence are on the left.
Evidence of chemical impact include the number of differentially expressed genes that covary with a chemical (CLR), the
number of differentially expressed genes associated with a chemical in the Comparative Toxicogenomics Database (CTD),
presence detected by analytical chemistry, and hazard quotient for biological effect or estrogenic effect. Note that not all
chemicals had information available in CTD.

149


-------
F.3.5. Case Study: Determining CECs Associated with Transcriptomics Effects in
Great Lakes Tributaries

The Maumee and Detroit rivers have a number of different inputs in the rivers including
combined sewage overflow sites, agricultural sources, and urban sources such as
WWTP effluent. Both rivers have had tumor formation in fish identified as a beneficial
use impairment (Blazer et al. 2014; https://www.epa.gov/great-lakes-aocs/about-
maumee-river-aoc: https://www.epa.gov/great-lakes-aocs/about-detroit-river-aoc). We
used HQ and statistical approaches with transcriptomics and water chemistry to identify
chemicals detected in the mixture of CECs that represent potential threats to fish health
(Perkins unpublished). HQ were developed for estrogenicity and tumor formation based
on in vitro assay data. Analysis of water chemistry identified four chemicals at sufficient
levels to cause estrogenicity (HQ>1) with two at much higher HQ than the rest (17-beta-
Estradiol»> Estrone» BPA >4-Androstene-3,17-dione> 4-Nonylphenol). However,
estrogenic effects in males was found at only two sites and did not correlate with HI for
estrogenicity. Two chemicals were identified at sufficient levels to potentially cause
tumor formation (HQ fluoranthene >> tris (2-chloroethyl) phosphate). PAHs had the
most impact across all sites (agricultural and urban) when measured by transcriptomics
associated with specific chemicals, followed by pharmaceuticals, pesticides and
alkylphenol ethoxylates (Figure F.3). On an individual chemical level (Figure F.4),
phenanthrene, fluoranthene, and pyrene were associated with the highest number of
gene expression changes. These data indicate that PAHs have significant impacts on
fish, with fluoranthene a significant concern due to its high HQ and association with a
large number of gene expression changes.

150


-------
WWTP discharge

Alkylphenol ethoxylates
Antimicrobial
Aromatics
Flame retardant
Flavors, Fragrances
Herbicides
Hormones
Industrial chemicals
Sterols
PAHs
Pesticides
Pharmaceuticals
Plasticizer/Solvent

U	o	o	Q	—I

^	^	Q	^	^

<	*	m	<	0

5	5	o	^

00	U	I—	>	(J



6

0

18

21

8

22

1

0

5

10

3

6

6

7

6

8

6

8

4

5

7

7

2

4

1

0

6

11

7

8

6

12

12

3

4

3

2

2

4

16

6

15

4

2

12

12

7

12

4

2

8

9

3

7

49

46

45

77

45

75

8

7

28

28

16

26

14

7

66

45

21

46

2

2

10

3

1

6

Figure F.3. Ranking of chemicals in Maumee and Detroit rivers using the sum of covarying
differentially expressed genes in caged fish exposed at sites upstream of, and nearby,
WWTP discharge sites. The total number of differentially expressed genes covarying with a
chemical class are listed in each box with higher values shaded a darker red. Boxes with
zero represent chemicals with no significant covarying differentially expressed genes.

151


-------
A. Chem B. DEG:Chem
WWTP WWTP
SCTYGN SCTYG

Alkylphenol
ethoxylates

Antimicrobials |
Aromatics

Flame retardant
plasticizers

Flavor-
Fragrances

Herbicides]

Hormones

Industrial
UV inhibitor

>0.1
Mg'L |
0.0001

¦u. 1

I

PAHs

Pesticides

Pharmaceuticals

Plasticizers
Industrial solvents |

Sterol |

4-Nonylphenol

Tris(2-chloroethyl)phosphate

17-beta-Estradiol

4-Androstene-3,17-dione
Estrone

Fluoranthene

>15

Differentially
Expressed genes

Bis phenol A

Figure F.4. Prioritization of individual chemicals detected near waste water treatment
plants in the Maumee and Detroit Rivers. Heat maps of the average surface water
chemistry and gene expression correlated to average surface water chemistry across
deployment sites in the Maumee and Detroit Rivers. A. Heat map of average chemistry
concentration in surface water during 4 days exposure (Chem). Classes of chemicals are
listed are identified on the left. Blank boxes= concentrations below detection limits. Black
SCT=sites on Maumee river, Blue YGN=sites on Detroit river. WWTP=Waste Water
Treatment Plant outfalls. S=Swan Creek, C— Clark Oil site, T=Toledo WWTP outfall,
Y=Wyandotte WWTP outfall, G=Grosse He WWTP outfall, N=Toledo WWTP outfall. B.
The number of differentially expressed genes that were correlated with average chemistry
at deployment sites (DEG:Chem). Blank boxes indicate no significant association between
genes and chemistry. Chemicals on the right were identified as being present at
concentrations sufficient to cause estrogenicity (normal type) or tumors (bold type).

152


-------
F.3.6 Key Findings/Progress - Associating Detected Chemicals with Biological
Effects Using Transcriptomics and Evidence-Based Approaches

•	Statistical association of water chemistry with gene expression from exposed
caged fish can be used to develop a biological impact ranking based upon the
number of genes that covary with chemicals in surface water. This can be used
to prioritize management actions to reduce costs and lower the number of CECs
that must be managed.

•	The combination of evidence-based approaches with transcriptomics analysis
provided more weight of evidence that specific chemicals in CEC mixtures cause
biological and adverse effects.

•	PAHs appear to be present at many sites at high concentrations and have
significant biological impacts on exposed fish. This indicates that PAHs are
potential threats to wildlife in the Great Lakes.

•	Gene expression changes can be used to identify potential exposure to chemical
not tested for such as hospital waste chemicals.

F.4. Linking Biological Activities to Adverse Apical Effects in Wildlife

Effects-based monitoring can detect changes in fish as a result of being exposed to
CEC mixtures. However, these effects need to be related to adverse apical effects in
wildlife to determine if the tributary is adversely impacted and to identify what CECs
might potentially cause these effects. Transcriptomic changes can be linked to apical
effects because gene expression changes can lead to changes in cells, tissues and
health of an animal (Box F.1). The AOP provides a framework that describes how a
molecular initiating event can lead to activation of key events that cause and adverse
outcome (Ankley et al 2010). Using the AOP framework we can develop pathways in
which gene expression changes can be linked to apical effects.

F.4.1. The Adverse Outcome Pathway (AOP) Framework with Transcriptomics
Effects-Based Monitoring

Since an AOP is a series of events with one leading to another, key events leading an
outcome can be measured to predict the likelihood of an adverse event occurring
(Perkins et al., 2019a). This is especially useful when using untargeted monitoring
approaches such as transcriptomics as an AOP serves as a framework to organize
information and test hypotheses as to whether evidence exist to support the activation
on an AOP. Here we have developed an AOP network describing the events that lead
from sustained liver cypla activation and/or steatosis activation causing liver damage
that results in proliferative regeneration and subsequent tumor formation (Figure F.5).
The AOP network uses gene expression as indicators of molecular events in the
network which lets one use transcriptomic analysis of exposed animals to assess how
much of the pathway might be activated. This permits linkage of transcriptomic impacts

153


-------
to tumor formation, a common biological use impairment (BUI) in Great Lakes
Tributaries.

F.4.2. Coupling Organ Damage Models with Transcriptomics Data

While most attention has focused on adverse outcomes that impact reproduction, liver
damage in fish can result in tumor formation can result in tumor formation, metabolic
impairments or other diseases. Tumor formation can lead to impacts on resources and
sportfishing and therefore is considered a BUI. We developed a network of AOPs that
describe how chemicals can cause liver toxicity by impairing fatty acid metabolism or
steatosis (Burgoon et al., 2019 unpublished, Perkins et al., 2019b). We combined the
steatosis AOP network with gene expression to create a steatosis AOP Bayesian
Network model (AOPBN) to quantitatively model the transcriptomic effects. The
steatosis AOPBN is available in the US Army's Bayesian Inference for Substance and
Chemical Toxicity risk assessment software tool (Burgoon et al 2019 in review;
https://github.com/DataSciBurgoon/bisct). The steatosis model calculates the probability
of a chemical causing steatosis given the activation or inactivation of genes in the
network. The model describes signaling, fatty acid metabolism and transport pathways
important in steatosis using the activity of key proteins and measurements of fatty acid
influx, efflux, lipogenesis, and fatty acid beta oxidation. Here, gene expression was used
as an indicator of increased or decreased levels of protein and activity. Since sustained
steatosis over long periods results in repeated cellular injuries, it can lead to
regenerative proliferation and eventually tumor formation.

154


-------
macrophage

frizzled

disheveled

Hypoxia

hiflan

Mipemi

IncTCF?

•ccndl

foxml

regenerative cell
proliferation

Gl/S transition

cydin-cdk
complex

cdk4

Sustained
activation of
Cyp1a1

Oxidative 	k

Steatosis	stress 1/

Steatosis AOP network and
Bayesian network model

Repeated H\ Regenerative
injury	proliferation i^Tumor

	I

Regenerative proliferation AOP
network

Figure F.5. AOP network representing sustained cytochrome p450 activity or steatosis creates oxidative stress resulting in
repeated liver tissue injury, leading to regenerative proliferation, tumor formation and carcinogenesis.

155


-------
F.4.3. Case Study: Transcriptional Effect-Based Monitoring of Potential for CEC
Mixtures to Cause Tumor Development

The Maumee and Detroit Rivers have input from agricultural, industrial and urban
sources. Both Maumee and Detroit rivers have experienced degradation offish and
wildlife populations and fish tumors or other deformities (https://www.epa.gov/great-
lakes-aocs/about-maumee-river-aoc; https://www.epa.gov/great-lakes-aocs/about-
detroit-river-aoc). Surveys of native fish have indicated tumor formation as an adverse
outcome in these rivers (Blazer et al. 2014). We examined how CECs and
transcriptomics can be linked to adverse outcomes with effects-based monitoring of
CECs present in different areas near the outlets of the rivers (Perkins unpublished). To
do this we examined the hypothesis that deployment of fathead minnows near WWTP
discharge sites in the Maumee and Detroit Rivers would activate genes in the AOP
network leading to tumor formation and carcinogenesis via steroidogenesis and
regenerative proliferation (Figure F.5). A high number of genes related to tumor
formation were activated in livers of exposed fish (Figure F.6). Analysis of gene
expression results supporting the AOP indicated sustained activation of cyplal and
oxidative stress, steatosis activation at all sites, and activation of key genes controlling
regenerative proliferation. This is consistent with the association of PAHs as the most
impactful chemicals on gene expression, the high fluoranthene HQ for tumor formation,
activation of tumor pathway related genes, and elevated levels of cypla and oxidative
stress related gene expression consistent with exposure to PAHs in native fish in the
Detroit river (Braham et al., 2017).

F.4.3.1 Key Findings/Progress - Development of Translational Tools and
Frameworks for Linking transcriptional Measurements to Adverse Effects in
Wildlife

•	A combination of AOPs, HQ, and association approaches can be used to
support the assessment of the contribution of individual chemicals to adverse
effects seen due to exposure to CEC mixtures.

•	AOP networks and models combined with transcriptomics can be used to
assess activation of key events leading to adverse effects seen in wildlife
such as tumor formation as a result of CEC exposure.

156


-------
A

U§§9=!h
< ^ j < O w
Toxicity Function ^ u h ^ 5 h

Liver cancer
Liver tumor
Liver carcinoma
Hepatic steatosis
Hepatocellular carcinoma
Hepatomegaly
Hepatic injury
Liver cholangiocarcinoma
Liver metastasis by bladder
Hepatocellular adenoma
Hepatocellular damage
Hepatic adenocarcinoma

• • • •

•	• a a ¦

•log(p-value)

• ¦

B

wo
C

o

y
c
3

IL.

X

o

2	u	§	o	=!	H

q;	Z	ui	Z	i/i	2

<	<	-j	2 comparison across sites

r

1.3

>8.8

Figure F.6. Heatmap comparing liver toxicity functions enriched in differentially expressed genes from caged fathead
minnows exposed along the Maumee and Detroit rivers. Dots denote values below significance. A. Significance of enrichment
for genes associated with toxicological functions. More significant functions are represented by darker purple. B. Activation of
liver toxicity functions across deployment sites. Orange is activated, blue is inhibited.

157


-------
F.5 Management Implications and Future Directions

In this section we described how gene expression data can be used by risk managers to
identify potential adverse effects in waters of the Great Lakes tributaries. We described
two different types of gene expression platforms: caged fish and laboratory exposed
fish. We have demonstrated that transcriptomics combined with these systems can be
used to identify CECs that are likely to cause harmful effects so that managers can
focus efforts on smaller list of CECs. The benefits of using these platforms are that
predictions of potential chronic toxicity can be made based on relatively short exposure
times. This is especially helpful when determining if management actions have been
successful for mitigating irreversible toxicities, such as tumor formation or structural
abnormalities. Due to the nature of irreversible toxicity, it would be possible to find fish
and other organisms with tumors and structural abnormalities even though the water
quality has improved. The use of caged fish or laboratory exposures enables the use of
newly hatched and reared organisms would not develop the irreversible toxicity if no
chemical is present.

Transcriptomics effect-based monitoring is most effectively used to link CECs to
adverse outcomes when used in combination with evidence-based approaches such as
Hazard Quotients, water chemistry, and AOPs representing adverse outcomes of
concern.

Currently, our models perform well for tumor formation and liver toxicity, including fatty
liver (steatosis). We are extending our models and validating our approaches for
additional toxicities that may be of interest to risk managers. Currently, we recommend
that our models may have the greatest impact on those welfare effects that are
associated with tumors in fish. These welfare effects include many of the ecosystem
services associated with sport fishing, recreation, and cultural uses of water bodies, and
the related economic impacts.

F.6 Action Plan I Products

Perkins EJ, Habib T, Escalon BL, Cavallin JE, Thomas L, Weberg M, Hughes MN,
Jensen KM, Kahl MD, Villeneuve DL, Ankley GT, Garcia-Reyero N. Prioritization of
Contaminants of Emerging Concern in Wastewater Treatment Plant Discharges
Using Chemical:Gene Interactions in Caged Fish. Environ Sci Technol. 2017
Aug1 ;51 (15):8701 -8712. doi: 10.1021/acs.est.7b01567.

Perkins EJ, Gayen K, Shoemaker JE, Antczak P, Burgoon L, Falciani F, Gutsell S,
Hodges G, Kienzler A, Knapen D, McBride M, Willett C, Doyle FJ, Garcia-Reyero N.
Chemical hazard prediction and hypothesis testing using quantitative adverse
outcome pathways. ALTEX. 2019a; 36(1 ):91 -102.

Perkins EJ, Ashauer R, Burgoon L, Conolly R, Landesmann B, Mackay C, Murphy CA,
Pollesch N, Wheeler JR, ZupanicA, ScholzS. Building and Applying Quantitative

158


-------
Adverse Outcome Pathway Models for Chemical Hazard and Risk Assessment.
Environ Toxicol Chem. 2019b; 38(9): 1850-1865. doi: 10.1002/etc.4505.

Burgoon 2019. A Steatosis Adverse Outcome Pathway Bayesian Network model and
interface (https://github.com/DataSciBurgoon/bisct)

F.7 Other References

Ankley GT, Bennett RS, Erickson RJ, Hoff DJ, Hornung MW, Johnson RD, Mount DR,
Nichols JW, Russom CL, Schmieder PK, Serrrano JA, Tietge JE, Villeneuve DL.
Adverse outcome pathways: a conceptual framework to support ecotoxicology
research and risk assessment. Environ Toxicol Chem. 2010 Mar;29(3):730-41. doi:
10.1002/etc.34.

Ankley, GT. The fathead minnow in aquatic toxicology: Past, present, and future. Aquat.
Toxicol. 2006. 78, 91-102.

Blazer VS, Mazik PM, Iwanowicz LR, Braham RP, Hahn CM, Walsh HL, Sperry AJ..
Assessment of the fish tumor beneficial use impairment in brown bullhead (Ameiurus
nebulosus) at selected Great Lakes Areas of Concern. 2014. US Geological Survey
Open-File Report 2014-1105, 17 p., http://dx.doi.o.	3/ofr20141105.

Braham RP, Blazer VS, Shaw CH, Mazik PM. Micronuclei and other erythrocyte nuclear
abnormalities in fishes from the Great Lakes Basin, USA. Environ Mol Mutagen.
2017 Oct;58(8):570-581. doi: 10.1002/em.22123. Epub 2017 Sep 4.

Deng Y, Meyer SA, Guan X, Escalon BL, Ai J, Wilbanks MS, Welti R, Garcia-Reyero N,
Perkins EJ. Analysis of common and specific mechanisms of liver function affected
by nitrotoluene compounds. PLoS One. 2011 Feb 8;6(2):e14662.

Kramer A, Green J, Pollard J, Tugendreich S.. Causal analysis approaches in Ingenuity
Pathway Analysis. Bioinformatics, 2014. 30 (4), 523-530.

Lee KE, Langer SK, Menheer MA, Foreman WT: USGS Data Series 723: Chemicals of
Emerging Concern in Water and Bottom Sediment in Great Lakes Areas of Concern,
2010 to 2011— Collection Methods, Analyses Methods, Quality Assurance, and
Data. 2012. Data Series 723 US Geological Survey, Reston, VA.

U.S. Environmental Protection Agency (USEPA). EPA's 2014 National Air Toxics
Assessment Technical Support Document. 2018. Office of Air Quality Planning and
Standards, Research Triangle Park North Carolina 27711.

https://www.epa.gov/sites/production/files/2018-
09/documents/2014 nata technical support document pdf

USEPA. Risk Assessment Guidance for Superfund Volume I Human Health Evaluation
Manual (Part A). EPA/540/1-89/002.1989. Office of Emergency and Remedial
Response, U.S. Environmental Protection Agency Washington, D.C. 20450

159


-------
USEPA. Risk Assessment Guidance for Superfund Volume I: Human Health Evaluation
Manual (Part F, Supplemental Guidance for Inhalation Risk Assessment). EPA-540-
R-070-002 OSWER 9285.7-82. 2009. Office of Superfund Remediation and
Technology Innovation, U.S. Environmental Protection Agency Washington, D.C.
20450

Webster AF, Chepelev N, Gagne R, Kuo B, Recio L, Williams A, Yauk CL. Impact of
Genomics Platform and Statistical Filtering on Transcriptional Benchmark Doses
(BMD) and Multiple Approaches for Selection of Chemical Point of Departure (PoD).
PLoS One. 2015 Aug 27; 10(8):e0136764.

160


-------
Contaminants of Emerging Concern n the Great Lakes:
GLRI Integrated Phase II Group Progress Report


-------
Contaminants of Emerging Concern in the Great Lakes:
GLRI Integrated Phase li Group Progress Report

Contributing Authors:

Steph Hummel1, Gerald Ankley2, Lyle Burgoon3, Steve Corsi4, Christine Custer5, Kimani Kimbrough6,

Heiko Schoenfuss7, Sarah A. Zack8, and Elizabeth Murphy9

1	U.S. Fish and Wildlife Service, Ecological Services, Bloomington, MN, USA

2	U.S. EPA,Office of Research and Development, Great Lakes Toxicology and Ecology Division, Duluth, MN, USA

3	U.S. Army Corps of Engineers, Engineer Research and Development Center, Vicksburg, MS, USA

4	U.S. Geological Survey, Upper Midwest Water Science Center, Middleton, Wl, USA

5	U.S. Geological Survey, Upper Midwest Environmental Sciences Center, La Crosse, Wl, USA

6	National Oceanic and Atmospheric Administration, National Centers for Coastal Ocean Science, Silver
Spring, MD, USA

7	St. Cloud State University St. Cloud, MN, USA

8	Illinois-Indiana Sea Grant College Program, University of Illinois Extension, Woodstock, IL, USA.

9	U.S. EPA, Great Lakes National Program Office, Chicago, IL, USA.

Project Officer: Derek Ager
Great Lakes National Program Office
77 West Jackson Blvd, Chicago, IL, USA 60604-3590

This report and its contents are UNCLASSIFIED.

DISTRIBUTION A: Approved for public release: distribution unlimited.

The findings and conclusions in this article are those of the author(s) and do not necessarily represent the
views of the agencies involved.

Cover photo: Detroit River, Mlcigari, Photo Credit: Kimani Kimbrough

This page: Superior Falls at the mouth of the Montreal River on the border of
Wisconsin and Michigan. Photo credit: Austin Baldwin.

I!


-------
Table of Contents

1.	Executive Summary					IV

2.	Key Findings	V

3.	Introduction			1

Results

4.	Surveilance Program 			2

5.	Integrated Assessment Case Studies (IACS)			6

6.	Priority Contaminant Mixtures (PCM)	10

7.	Products Associated with Key Findings	11

Sampling the Manitowoc River in Wisconsin.
Photo credit: Austin Baldwin.

Ill


-------
Executive Summary

Under Action Plan II (2015-2019) of the Great Lakes Restoration Initiative (GLRI), Federal and Academic
partners continued an investigation that began under Action Plan I (2010-2014) into the presence and
distribution of contaminants of emerging concern (CECs) in the Great Lakes, and their potential impacts
on fish and wildlife. The term CECs is applied to a broad range of chemicals that are currently present in
the environment due to their historical or ongoing use. Data are currently lacking to determine whether fish,
wildlife, or humans are being exposed to CECs or if negative health or environmental effects are expected
if exposure to CECs occurs. Pharmaceuticals, personal care products, flame retardants, some current-use
pesticides, polycyclic aromatic hydrocarbons, and poly- and perfluorinated chemicals are often identified as
CECs, but currently there is no definitive or comprehensive list.

This progress report seeks to provide an update on the results of the activities undertaken in the 2016 field
season (and other completed work to-date), which built on the results of the Phase I work conducted under
GLRI Action Plan I. Specifically, Phase I was focused on the development of methods and gathering of CEC
data to guide the design and help achieve the objectives of Phase II. The activities presented in this report
contributed towards fulfilling the GLRI Action Plan II Objective 1.2.2: Identify emerging contaminants and
assess impacts on Great Lakes fish and wildlife. Three specific goals were further developed from Objective
1.2.2:

1.	Characterize and evaluate the extent to which CECs might threaten fish and wildlife populations
relative to other chemical stressors present in the Great Lakes.

2.	Pilot and develop state-of-the-art surveillance techniques for biological effects from CECs in the
Great Lakes basin.

3.	Develop information and tools for resource managers to better manage and address potential CEC
threats to fish and wildlife populations.

A three-pronged approach was used to implement these goals which included:

1.	Implementing a surveillance program to determine which CECs occurred with the greatest
frequency and abundance across the Great Lakes basin to identify seasonal and spatial patterns of
occurrence.

2.	Conducting Integrated Assessment Case Studies (IACS) focusing on a particular waterbody and/
or suite of CECs and conducting an analysis of the effects of CECs on organisms. The 2016 IACS
focused on pesticides in the Maumee River, Ohio.

3.	Evaluating the toxicity of priority contaminant mixtures (determined by the work conducted in
Phase I) on fish and native mussels in a laboratory setting.

Manitowoc River in Wisconsin.
Photo credit: Austin Baldwin.

IV


-------
Key Findings

•	Pesticides were found in water at all sampled sites, with the greatest concentrations occurring in
May-July for herbicides. Complex chemical mixtures were common, and an average of 18 pesticide
compounds, including pesticides and pesticide degradation products, were detected in each sample.

o Many pesticides were detected in the Maumee River water samples. The watershed includes
inputs associated with agricultural practices and urban influences such as runoff from residential,
commercial, and park lands. The Maumee River had the most pesticide and pesticide degradates
detected of the 16 Great Lakes watersheds sampled. Seventy compounds were detected of the 231
analyzed throughout the study with a maximum of 57 compounds detected in one individual sample.

•	Polycyclic aromatic hydrocarbons (PAHs) were a class of contaminants detected throughout
the basin in water and occasionally biota dreissenid mussels and aquatic insects). A multiple lines-
of-evidence approach indicated that coal-tar-sealed pavement was the most likely source of PAHs in
sediment at 80% of the 71 locations sampled, and vehicle emissions were the primary source at 10% of
sites.

•	Observed concentrations of CECs would not be expected to produce lethal effects in fish and
wildlife species, but some CECs were detected in aquatic life. These results indicated potential
impacts to the species offish (e.g., lake sturgeon, sunfish, and fathead minnows) and mussels
(pocketbook, spike, fatmucket, zebra, and quagga mussels) examined as part of this effort. These
possible effects include anatomical changes (e.g., feminization of males), impacts on the nervous
system, declining offspring production, and developmental delays. The possibility of impacts on these
types of endpoints warrants further investigation.

o Results from caged fish (sunfish, fathead minnow) studies indicated some molecular and biochemical
changes in exposed animals, but it is unclear the degree to which this was due to CECs nor what
the consequences of these changes might be in terms of impacts on overall fish populations.
Inconclusive results were found surrounding developmental delays in fish, which were seen in some
years and at some sites but not consistently.

o Initial metabolomics results for dressinid mussels found patterns which showed differences

between sites. Work is still in progress to understand the relationship between CECs and dressinid
metabolomics data.

o Native mussel bio-effects assessments did not find relationships between environmental CEC

exposure and glycogen and gamete counts in wild caught native mussels, but CECs were present in
tissues indicating other assessments may be warranted to determine if CECs impact native mussel
biology.

Milwaukee River, Wisconsin. Photo credit: Kimani Kimbrough.

V


-------
o Laboratory exposures with formulated CEC mixtures designed to mimic the field with fish and native
mussels did not impact survival but yielded decreased number of offspring produced in fish. Native
mussel glycogen, fatty acid, sperm production, and behavior were not affected by exposure to CEC
mixtures in the laboratory, but offspring viability decreased.

• Pesticides and herbicides were rarely detected in tree swallow tissues, with two exceptions.

The insect repellent ingredient DEET, or diethyltoluamide, was detected in nearly all liver tissue samples,
and desethylatrazine, a byproduct of the herbicide atrazine, was detected in virtually all tree swallow
tissues including in swallows nesting at remote Wl lakes. These compounds were not expected to be
found in tree swallows because of their chemical properties and modes of action, so follow-up studies
are warranted to learn more about the metabolic and ecosystem pathways leading to this exposure.

Manitowoc River in Wisconsin.
Photo credit: Austin Baldwin.

VI


-------
Introduction

Contaminants of emerging concern (CECs) are chemicals identified that may pose threats to the Great
Lakes ecosystem, currently in use and for which we lack a clear understanding of whether fish, wildlife, or
humans are being impacted by exposure. The term CECs has been applied to a broad range of commercial
products, including pharmaceuticals, personal care products, flame retardants (e.g., polybrominated
diphenyl ethers; PBDEs), some current-use pesticides, polycyclic aromatic hydrocarbons, poly- and
perfluorinated chemicals, and household chemicals, as well as a variety of industrial chemicals. In general,
there is a lack of information about the toxicity of these chemicals and their occurrence, movement, and
fate in the environment. In order to reduce and prevent adverse ecological effects from CECs to fish and
wildlife, and associated negative impacts on related economic and recreational activities in the Great Lakes,
the presence and deleterious effects of CECs must be better understood and appropriately managed.

This collaborative research effort is a continuation and expansion of the work previously conducted
under the Great Lakes Restoration Initiative (GLRI) Action Plan I GLRI Action Plan I included work which
characterized and determined the presence of CECs across the Great Lakes basin and the development
of methods designed to address the different Phase II objectives. . GLRI Action Plan II built on the work of
Action Plan I by continuing surveillance of CECs and investigating the biological effects CECs have on fish,
mussels, and wildlife residing in the Great Lakes basin. This work will contribute towards better resource
management efforts throughout the Great Lakes.

Sheboygan River, Wisconsin. Photo credit: Kimani Kimbrough.

A three-pronged approach was used to implement
research goals for GLRI Action Plan II which
included:

1.	Implementing a surveillance program to
determine which CECs occurred with the greatest
frequency and abundance across the Great
Lakes basin, to identify seasonal changes in
patterns of occurrence.

2.	Conducting Integrated Assessment Case
Studies (IACS) focusing on a particular
waterbody and/or suite of CECs and conducting
an analysis of the effects of CECs on organisms.
The 2016 IACS, which is summarized in this
report, focused on the Maumee River/watershed,
Ohio, and evaluated several chemicals, with an
emphasis on pesticides.

3.	Evaluating the toxicity of priority contaminant
mixtures (determined by the work conducted in
Phase I) on fish and native and invasive mussels
in a laboratory setting.

This research effort was comprised of individual and
collaborative projects from multiple federal agencies
and academic institutions, and was overseen by the
U.S. Environmental Protection Agency (EPA), Great
Lakes National Program Office. Partners include
the U.S. Geological Survey, National Oceanic and
Atmospheric Administration, U.S. Army Corps of
Engineers, U.S. Fish and Wildlife Service, Saint
Cloud State University, and the U.S. EPA Office of
Research and Development.

This progress report details results of the research
completed during the 2016 field season and
completed works to date.

GLRI CEC 2016

1

Phase 2 Progress Report


-------
Surveillance Program (SP)

The surveillance program focused on achieving the specific goal "characterize and evaluate the extent to
which CECs occur throughout the Great Lakes Basin". Projects identified and determined which CECs
occurred in water and biota (fish, mussels, and tree swallows) with the greatest frequency and abundance
across the Great Lakes basin and identified seasonal changes in patterns of occurrence. Surveillance
studies discussed below include efforts from water years (Oct. - Sept.) 2016 and 2017. Specific objectives
included:

•	Expand the information on pesticide occurrence, seasonality, and potential biological effects across a
land use gradient in Great Lakes tributaries.

•	Expand information on PAH occurrence, magnitude, potential biological effects, and sources across a
land use gradient from very little to full urban influence.

•	Test for potential biological effects using molecular techniques, including Attagene (a commercially-
available set of high-throughput in vitro assays), metabolomics, and transcriptomics.

•	Determine CEC prevalence in dreissenid mussels throughout the Great Lakes tributaries.

•	Determine relations of CEC presence with watershed attributes including the amount of urban land
cover, wastewater treatment effluent influence, agricultural crops and pasture land cover.

•	Develop methods to evaluate potential influence to fish and wildlife in the Great Lakes basin using
CEC concentration data in comparison to concentrations of potential biological concern using effects
measured for several thousand chemicals through the USEPAToxCast program.

•	Characterize CEC concentrations in lake sturgeon tissues.

•	Prioritize chemicals and sites by potential for adverse biological effects.

Niagara River, New York. Photo credit: Kimani Kimbrough.
GLRI CEC 2016

2

Phase 2 Progress Report


-------
Key Findings:

• Monthly water samples were collected during
variable flow conditions from 16 Great Lakes
tributaries during water year 2016 (October 2015
and September 2016; Figure 1). Agricultural land
use varied from 0.1-79% of the drainage area
for these watersheds. Pesticides were detected
at all sites. At least one pesticide or degradation
product was detected in 190 of the 198 pesticide
samples collected, and 104 compounds were
detected in at least one sample out of the 231
parent compounds and degradation products
analyzed.4'

o Pesticides were detected in each month of
the year with the greatest concentrations
occurring in May-July for herbicides.

o The presence of complex chemical mixtures
was common; samples had an average of 18
pesticide-related compounds (i.e., parents or
metabolites) detected per sample and 30 or
more compounds were detected in 15% of
samples.

o Metolachlor, 2,4-D, diuron, and atrazine
were the chemicals of greatest concern
given their widespread occurrence (detected
at >11 sites) and potential for adverse
biological effects (exceedance of bioeffect
thresholds at >12 sites and in >35% of
samples). Concentrations were compared to
bioeffect thresholds including EPA Aquatic life
benchmarks and endpoints within the ToxCast
high-throughput screening database. Among
the 16 tributaries studied, the Maumee River
basin had frequent detection of pesticides;
exceedances of bioeffect thresholds occurred
for all 18 Maumee River samples, including
exceedances for 15 compounds. The
Maumee River also had the most chemicals
detected across the duration of the study (70)
and in a given sample (57).

'Superscript numbering following each Key Findings
bullet corresponds to the principal investigator
conducting the study and generating the key finding.
Numbers identifying associated principal investigator
can be found below the author list.

—	Watersheds samptaci

		 Great Lakes watershed

boundary

—	Water

—	National boundary

—	State/province boundary
A Sampling site

Figure 1; 167 Tributaries surveyed for the presence of pesticides from Oct. 2015 - September 2016. The Maumee
River basin had a high detection of pesticides and the most chemicals detected.

GLRI CEC 2016

3

Phase 2 Progress Report


-------
Pharmaceutical and personal care products
(PPCPs) were detected at all dreissenid mussel
sites and were highest at urban sites near
municipal wastewater treatment plant (VWVTP)
discharges. The most common compounds
detected were sertraline (an antidepressant),
nonylphenol and nonylphenol ethoxylates
(surfactants).6

Polycyclic aromatic hydrocarbons (PAHs) were
the dominant contaminant class of CECs in
dreissenid mussel tissues and found throughout
the basin. Levels were highest in river harbors
and bays with high percentage of impervious
surfaces but also found in nearshore and offshore
lakes zones (Kimbrough et al., 2019; Figure 2).6

Statistical models were developed to predict
the probability of CEC occurrence in water and
sediment across the Great Lakes Basin. Water
and sediment data, along with 21 watershed
characteristics, from 24 US Great Lakes
tributaries were used in the models. Watershed
characteristics included land use, number of
permitted point sources, and distance to point
sources all of which were important predictors
of CEC occurrence. Developed land use and
distance to point sources were most often
important predictors of CEC occurrence. Of
the 24 sites sampled, sites with the greatest
predicted occurrence of CECs included the Fox
River, Milwaukee River, North Shore Channel,
Little Calumet River, Cuyahoga River, and
Tinker's Creek (Keisling et al 2019).Results from
the predictive models can be used to assess
vulnerability of Great Lakes tributaries to CEC
occurrence to guide future research and/or
resource management decisions.1

Figure 2: Great Lakes PAH dreissenid mussel locations
occur throughout the basin. Some of the locations have
multiple sites.

GLRI CEC 2016

4

Phase 2 Progress Report


-------
• Streambed sediments were sampled arid
analyzed for PAHs at 71 locations across 26
Great Lakes tributaries in 2017 (Figure 3). The
results were compared to Threshold Effect
Concentrations (TEC; 1,610 pg/kg for IPAH16)
and Probable Effect Concentration (PEC; 22,800
pg/kg for IPAH16) defined as consensus-
based freshwater sediment quality guidelines
(Ingersoll et al., 2001). TEC and PEC are used
as indicators to measure the potential for adverse
biological effects. The TEC was exceeded at
62% of sampling locations and the PEC was
exceeded at 41% of sampling locations. Based
on source apportionment conducted using
multiple lines-of-evidence, coal-tar-sealed
pavement was the most likely source of PAHs
in sediment at 80% of the 71 locations sampled,
vehicle emissions were the primary source at
10% of sites, and a primary source could not be
determined for 10% of the sites.4

• PPCPs and PBDEs were detected in lake
sturgeon serum and gamete samples from
populations at four sites located in the lower
Great Lakes (Detroit River, Lower Niagara River,
Upper Niagara River, and St. Lawrence River).
Four PPCPs (bentropine, DEET, hydrocortisone,
and amitriptyline) were found in at least 25% of
adult lake sturgeon serum samples across all four
study sites, and three PPCPs (sertraline, DEET,
and 10-hyrdoxy-amitriptyline) were found in at
least 25% of lake sturgeon gamete samples from
the St. Lawrence River study site. Twenty out
of 40, polybrominated diphenyl ethers (PBDEs)
were found in every serum sample at every site.
A total of 14 PBDEs were found in all gamete
samples (Banda et al. 2020, in review).1

Watersheds sampled,
with number of sites
per watershed

Great Lakes watershed
boundary

Water

National boundary
State/province boundary

300 Kilomete

Figure 3: Map of the Great Lakes Basin and watersheds sampled for PAHs in sediment for during 2017. Numbers indicate the
number of sampling sites within each watershed

GLRI CEC 2016

5

Phase 2 Progress Report


-------
• The software package toxEval was developed
to facilitate analysis and visualization of
potential biological effects for complex chemical
concentration data sets that represent multiple
chemicals, sampling locations, and samples per
location (DeCicco et al. 2018). These techniques
were used to evaluate a data set from Phase
I of the GLRI that included 709 water samples
from 57 sites that included analysis of 67 organic
compounds.24

o Comparison using ToxCast high-throughput
screening results (https://www.epa.gov/
chemical-research/exploring-toxcast-
data-downloadable-data) and established

water quality benchmarks identified 12
chemicals with the greatest potential for
biological effects at multiple sites including
4-nonylphenol, bisphenolA, metolachlor,
atrazine, DEET, caffeine, tris(2-butoxyethyl)
phosphate, tributyl phosphate, triphenyl
phosphate, benzo(a)pyrene, fluoranthene,
and benzophenone (Corsi et al. 2019).

o Evaluation of chemical mixtures through
examination of adverse outcome pathway
networks commonly predicted effects on
reproduction and mitochondrial function.

Maumee River, Ohio Photo credit: W. Edward Johnson

Phase 2 Progress Report


-------
Integrated Assessment Case Studies (IACS)

The Maumee River in Ohio was chosen as the location for the IACS in 2016. The Maumee River is one of
the most agriculturally influenced rivers in the Great Lakes Basin, which allowed the IACS to focus on CECs
classified as pesticides, and their potential effects on biota including native fish, deployed (caged) fish,
dreissenid mussels, native fresh water mussels, and tree swallows. In addition to pesticides, select other
CECs (e.g., a suite of PPCPs) were measured. Sampling was conducted along a gradient from agricultural
land use (upstream) to urban/industrial land use (downstream) and timed to coincide with summer pesticide
application with an initial sampling occurring before (April/May), and a second after pesticide application
(June/July). Projects under IACS looked to connect observed exposure effects with the goals of identifying
novel biomarkers of CEC exposure, and determination of relevant thresholds of response for targeted
biological endpoints.

Western Lake Erie. Photo credit: W. Edward Johnson.

Key Findings:

• Chemistry and biology results for the Maumee
River along the gradient of sites indicate a
system potentially heavily influenced by land use
(both agricultural and urban land use). These
inputs reflect various degrees of non-point source
contamination and input from specific point-
source discharges, predominantly WWTPs.2

o There is a distinct seasonality relative to
potential agricultural impacts on the system,
with marked changes in concentrations of
some nutrients and herbicides associated
with crop production cycles. In some
instances (e.g., metalochlor and atrazine)
observed seasonal elevations in surface

• Both resident and caged sunfish from seven sites
in 2016 exhibited biological stress responses that
may be attributable to CECs present in Maumee
River water and sediment.17

o Greater responses were seen at the
downstream (urban) sites, and greater
changes in resident fish compared to the
caged fish. CECs were found in every
sunfish tissue sample analyzed in this study.
Chronic exposure of agricultural and urban
contaminants may alter sunfish anatomy and
physiology which could lead to population
declines and altered ecosystem functioning
(Cipoletti et al. 2020, in review).

water concentrations in June/July versus
April/May which were sufficient, based on
existing effects data for possible impacts
on biological components of the system,
such as aquatic plants. Conversely, non-
pesticidal contaminants (i.e., PPCPs) at sites
predominantly impacted by WWTPs tended
not to exhibit large seasonal variations.

GLRI CEC 2016

7

Phase 2 Progress Report


-------
•	21 -day exposures of fathead minnows at
embryonic, larval and adult life stages, in 2016
and 2017 to in-situ water samples from the same
seven Maumee River sites where sunfish were
assessed, resulted in biological effects that
differed by life stage and by year.17

o Fecundity was the most sensitive

variable measured in adults. Embryonic
morphological development delays were
seen in 2016 but not 2017. Contaminants
were detected in every tissue sample, with
six pesticides and eight pharmaceuticals
detected in at least half of tissue samples.
No clear upstream (agricultural) to
downstream (urban) gradient was seen
because many of the CECs measured
were ubiquitous, occurring at multiple sites
(Cipoletti et al. 2019; Figure 4).

•	Dreissenid mussels exposed to Maumee River,

2016

Q 2 4 6 S 1ft il 14 Ifi 16 20
Dxy of «xpourv

accumulated measurable concentrations of
current use pesticides common to corn, soybean
and wheat agriculture including herbicides
(metolachlor and atrazine, trifluralin, and atrazine
transformation product desethylatrazine), and the
insecticide chlorpyrifos, and its transformation
product (chlorpyrifos oxon) (Kimbrough et al.
2018).6

• Multiple lines of evidence indicate generalized
physiological responses in fish to chemical
stressors at sites throughout the Maumee
watershed.2

o For example, enzymes related to the
metexample, enzymes related to the
metabolism of chemicals like pesticides and
pharmaceuticals were elevated in fathead
minnows caged for 4-days at several of the
study sites. The nature of these responses
could indicate adaptive/compensatory
responses to the chemical exposures and,
potentially, adverse impacts on population-
relevant endpoints such as survival and
growth.

o Overall, there was little evidence in the caged
fish responses consistent with endocrine-
disrupting chemicals like estrogens.

a 2 4 * « 10 12 14 1« It 29
Day of tKfWs-uf*

Figure 4: IACS fathead minnow exposure fecundity results.
Each line corresponds to the site where water collections took
place. Superscript letters following site code indicate significant
differences in total fecundity if the letters are different and non-
significant differences if the letters are the same. In both years
some sites had significantly higher fecundity than the controls.

GLRI CEC 2016

8

Phase 2 Progress Report


-------
• PPCPs (n = 140) were rarely detected in tree
swallow tissue or sediment samples from 6 sites
along the Maumee River in 2016 and 2017.5

o Tissues analyzed included egg, liver, and
diet samples all of which were pooled (n =
5 individuals per pool) by site. The same
tissues were also sampled at one or two
remote lakes in north central Wisconsin for
comparative purposes. Of the 34 PPCPs
detected in at least one sample, 26 were
detected in at least two samples. The most
frequently detected PPCP was DEET which
was detected in nearly all liver tissue and
sediment samples, but in only ~30% of
egg samples, lopamidol, a frequently used
contrast agent by medical professionals, was
detected in diet and liver samples, but not
in egg or sediment samples. Interestingly,
both DEET and lopamidol were detected
in samples from the remote lakes, as well
as, from the more urbanized and populated
Maumee River.

• Preliminary results for targeted metabolomics in
dreissenid mussels exposed to the Maumee and
Detroit Rivers, indicate differences in metabolite
profiles associated with the specific location of
sites relative to non-point source contamination
and proximity to WWTPs.6

o Because responses to environmental
variables may be dependent on life stage,
sex, size or age, longitudinal studies require
a deeper understanding of dreissenid mussel
physiology over a season. Methodological
(caging) influences may be important to
acknowledge in the validation of dreissenid
mussels as bioindicators in priority areas.

• Non-organochlorine pesticides (n = 34), which
included triazine herbicides, fungicides, and
organophosphate and carbamate insecticides,
were rarely detected in tree swallow samples
except for desethylatrazine, a metabolite of
atrazine.5

o Desethylatrazine was detected in nearly
all swallow egg, nestling carcass, and
diet samples. It was only detected in the
sediments at one of six sites. The parent
compound, atrazine, was detected in diet
samples at all Maumee River sites, but
only at the most upstream sites in swallow
tissues and sediment. Ametryn, another
triazine herbicide, was detected, but in only
one composite egg and one composite
nestling sample over the 2-year study period.
Neither the PPCPs nor these pesticides were
expected to be detected in avian tissues
because of their chemical composition and
mode of action.

GLRI CEC 2016

9

Phase 2 Progress Report


-------
•	Preliminary anticholinesterase (AChE) biomarker
response results in dreissenid mussels caged

in the Maumee River showed severe inhibition,
indicating the potential for significant neurotoxicity
associated with pesticide exposures. Additionally,
perturbation of the biomarker responses
related to oxidative stress (glutathione and lipid
peroxidation) was observed at some sites.6

o Multiyear studies with dreissenid mussels
facilitated the identification of "normal
ranges" of biomarkers based on an extensive
database, providing greater potential for
diagnostic approaches that do not necessarily
require a specific reference site as part of
the monitoring strategy. Therefore more
biomarker studies suggest the potential
for wide-spread cellular and physiological
impacts throughout the watershed that may
be related to pesticide exposures.

•	Twenty-one day exposure of native freshwater
adult mussels in the Maumee River watershed
(2017: Sp. plain pocketbook) and Milwaukee
River watershed (2017 and 2018: Spp. plain
pocketbook, fat mucket, and spike) to in-situ
water did not result in significant differences

in biological responses along the exposure
gradients. However, temporal differences were
seen in sperm production and CEC presence in
Milwaukee. Limited sample numbers and high
variability did not allow trends to be elucidated for
the endpoints (behaviors, fatty acids, glycogen,
teste development, and sperm density) and
species assessed (Rzodkiewicz et al. 2019, in
review).1

•	Native freshwater mussel health assessment
was completed in the Maumee River in 2016.
Preliminary results revealed although mussel
CEC tissue concentrations significantly differed
by site, glycogen concentrations and egg and
sperm counts showed no relationship to CEC
concentrations. This indicates glycogen and
gamete counts are not useful indicators when
assessing CEC exposure. Native freshwater
mussels bio-accumulate some CECs which
could impact health and lead to population
impairments.1

•	Preliminary analyses suggest that Great
Lakes tributaries with similar overall chemistry
profiles did not have the same changes in
gene expression related to hepatocyte toxicity
in cell lines (including liver cancer pathways),
suggesting overall chemistry profiles are not
predictive of potential liver toxicity.3

•	Preliminary analyses suggest that no
developmental abnormalities were observed in
zebrafish embryos exposed to surface waters
from the Genesee River, Saginaw River, Indiana
Harbor and the Oswego River.3

•	Preliminary analyses suggest that concentrated
pollutants at levels greater than those seen in
the environment at sites within the Kinnickinnic
River and Menomonee River, may have impacts
on zebrafish embryo survival and zebrafish
embryos with developmental abnormalities may
be observed.3

Some of the species found during the native freshwater
mussel health assessment (fatmucket and rayed bean).
Photo credit: Daelyn Woolnough

GLRI CEC 2016

10

Phase 2 Progress Report


-------
Priority Contaminant Mixtures (PCM)

Much of the known toxicity data on CECs is for single chemical analyses, which is not representative of
environmental conditions. Complex mixtures of CECs are present in the environment from a variety of
sources such as wastewater treatment plant effluent, combined sewer overflows, and runoff. Projects
first worked towards defining chemicals and their relevant environmental concentrations from the
surveillance program data. Phase I surface water data were assessed to determine which chemicals and
at what concentrations are most common across the Great Lakes Basin. Two diverse mixtures emerged;
chemicals predominantly found in agricultural landscapes (atrazine, DEET, TBEP, bromacil, estrone, BPA,
nonylphenol), chemicals predominantly found in urban landscapes (sulfamethoxazole, fexofenadine,
desvenlafaxine, metformin, DEET, TBEP, estrone, HHCB, Methyl-1 H-benzotriazole, BPA, nonylphenol). The
results of this work were used to make environmentally relevant chemical mixtures that could be used in a
laboratory setting to conduct exposure toxicity assessments (Elliott et al. 2018).

Fathead minnows (a model small fish species) were exposed to those defined chemical mixtures over
multiple generations to characterize biological endpoints and to help better understand if CECs are the
source of observed effects in the IACS. Native freshwater mussels were exposed to the same chemical
mixtures, during peak female reproductive periods and monitored through larval attachment on host fish,
concluding when juveniles dropped detached from fish for a total of 100 day study

Milwaukee River, Wisconsin. Photo credit: W. Edward Johnson.

Key Findings:

• When fathead minnows were exposed in 2016
across three generations to a statistically
derived mixture of CECs commonly occurring in
Great Lakes tributaries associated with urban
iand-use (Elliott et ai. 2018), physiological
and reproductive changes were noted. In the
first generation exposed to the CEC mixtures,
an egg-yolk protein was present in males at
concentrations usually seen only in female
fish. However, this effect was suppressed in
the subsequent exposure generation. Female
fathead minnows exposed in the second
generation to higher concentrations of urban
CECs, experienced declining fertility. (Lina Wang,
2017. MS Thesis, St. Cloud State University).17

• Adult native mussel health assessments of
glycogen, fatty acids, and behavior were not
found to be sensitive to the evaluated CEC
mixes and concentrations. Additionally, CEC
life cycle exposures were performed on native
plain pocketbook early mussel life stages (i.e.,
embryos in vivo and post spawn-parasitic
glochidia on exposed largemouth bass host fish)
to assess population relevant endpoints. Trials in
2017 and 2018 suggest sensitivity of early native
mussel life stages to both agricultural and urban
exposures with delay of transformation from
glochidia to free living juveniles. (Gill, Rappold,
and Rzodkiewicz, 2019. MS Theses, Central
Michigan University).1

GLRI CEC 2016

11

Phase 2 Progress Report


-------
Products Associated With Key Findings In This

Progress Report

Milwaukee River, Wisconsin. Photo credit: Kimani Kimbrough.

Surveillance Program

Completed Reports and Publications

All data collected as part of the surveillance program
are available:

USGS Water Data for the Nation. 1 October 2019.
https://waterdata.usgs.gov/nwis.

Banda, J.A., D.J. Gefell, V. An, A. Bellamy, Z.
Biesinger, J. Boase, J. Chiotti, D. Gorsky, T.
Robinson, S. Schlueter, J. Withers, S. L. Hummel.
2020. Contaminants of emerging concern: Body
burden characterization in lake sturgeon. Manuscript
in review.

Baldwin, A.K., Corsi, S.R., Oliver, S.K., Lenaker,
PL., Nott, M.A., Mills, M.A., Norris, G.A., Paatero,
P., 2020. Primary sources of PAHs to Great
Lakes tributaries using multiple lines-of-evidence,
Environmental Toxicology and Chemistry, in press.

Corsi, S.R., L.A. De Cicco, D.L. Villeneuve, B.R.
Blackwell, K A. Fay, G.T. Ankley, A.K. Baldwin. 2019.
Prioritizing chemicals of ecological concern in Great
Lakes tributaries using high-throughput screening

data and adverse outcome pathways. Science of
The Total Environment 686, 995-1009. https://doi.
org/10.1016/j.scitotenv.2019.05.457

De Cicco, L . S.R. Corsi, D.L. Villeneuve, B.R.
Blackwell, G.T. Ankley. 2018. toxEval: Evaluation
of measured concentration data using the ToxCast
high-throughput screening database or a user-
defined set of concentration benchmarks. R package
version 1.0.0., https://code.usgs.gov/water/toxEval,


Keisling, R. L., S. M. Elliott, L. E. Kammel, S.J. Choy,
S. L. Hummel. 2019. Predicting the occurrence of
chemicals of emerging concern in surface water and
sediment across the U.S. portion of the Great Lakes
Basin. Science of the Total Environment. 651:838-
850. https://doi.Org/10.1016/j.scitotenv.2018.09.201.

Kimbrough, K, W. E. Johnson, A. Jacob, M.

Edwards, E. Davenport. 2018. Great Lakes Mussel
Watch: Assessment of Contaminants of Emerging
Concern. Silver Spring, MD. NOAA Technical
Memorandum NOS NCCOS 249, 66 pp. https://
repository, library, noaa. gov/view/noaa/19484.

GLRI CEC 2016

12

Phase 2 Progress Report


-------
Product Topics in Development

Basiri wide assessment of pharmaceuticals in
dreissenid mussels based on habitat type (river,
nearshore, offshore).

Basin wide assessment of pesticides in dreissenid
mussels based on habitat type (river, nearshore,
offshore).

Using caged and in situ dreissenid mussels to
characterize chemical contamination in agricultural
and urban watersheds.

Legacy and contaminants of emerging concern
(CECs) in tree swallows along an agricultural to
industrial gradient: Maumee River, OH

An assessment and classification of pharmaceuticals
and personal care products (PPCPs) along the Great
Lakes Basin Coastal Zone: Relationship to mixed
land-use and point sources.

A multi-matrix assessment of pesticides
environmental occurrence and their link to pollution
gradients in the Maumee and Ottawa River
watersheds and riverine systems.

A predictive analysis of chemicals of emerging
concern (CECs) occurrence and distribution in the
Great Lakes Basin coastal riverine systems and
adjacent mixed land-use watersheds.

The distribution, seasonality, and potential biological
effects of pesticides in Great Lakes tributaries.

GLRI CEC 2016

13

Phase 2 Progress Report


-------
Integrated Assessment Case Studies

Product Topics in Development

Completed Reports and Publications

Cipoletti, N , Z. G. Jorgenson, J. A. Banda, S. L.
Hummel, S. Kohno, H. L. Schoenfuss. 2019. Land
use contributions to adverse biological effects in
a complex agricultural and urban watershed: A
case study of the Maumee River. Environmental
Toxicology and Chemistry 38:1035-1051. https://doi.
org/10.1002/etc.4409.

Cipoletti, N , Z. G. Jorgenson, J. A. Banda, S.
L Hummel, H. L Schoenfuss. 2020 Impacts of
Agricultural and Urban Land Use in the Maumee
River Watershed on the Anatomy and Physiology
of Caged and Resident Sunfish (Lepomis spp.)
Manuscript in review.

Kimbrough, K., A. Jacob, E Davenport, W.E.
Johnson, M. Edwards. 2019. Characterization of
Polycyclic Aromatic Hydrocarbon in the Great Lakes
Basin using Driessenid Mussels. Manuscript in
review.

Using dreissenid mussels for targeted and
untargeted metabolomics in place-based
environmental health assessment of rivers with
agricultural and urban influence.

Basin wide assessment of targeted and untargeted
metabolomics in established populations of
dreissenid mussels based on habitat type (river,
nearshore, offshore).

Using cellular biomarkers in dreissenid mussels for
place-based environmental health assessment of
rivers with agricultural and urban influence.

Impacts of Agricultural and Urban Land Use in the
Maumee River Watershed on the Anatomy and
Physiology of Caged and Resident Sunfish
(Lepomis spp.).

A case study of native freshwater mussel health in
the Maumee River.

side the mobile exposure laboratory for the 21
situ water exposure along the Maumee River
16. Photo Credit: USFWS.

Rzodkiewicz, L.D., M. Annis, S. L. Hummel, D.
A. Woolnough. 2019. Contaminants of emerging
concern may pose prezygotic barriers to freshwater
mussel recruitment. Manuscript in review.

Assessments of CECs, including PPCP and non-
organochlorine pesticides, in tree swallows nesting
along the Maumee River, 2016 and 2017.

Assessment of targeted metabolomics in tree
swallows based on habitat type and exposure to a
range of CECs.

Novel pathway-based approaches for assessing
biological hazards of complex mixtures of
contaminants: Application to an integrated
assessment of the Maumee River.

GLRICEC2016	14	Phase 2 Progress Report


-------
Priority Contaminant Mixtures

Completed Reports and Publications

Cipoletti, N. 2018. Complex agricultural mixtures:
assessing effects on aquatic species (Pimphales
promelas and Leomis spp.) through short-term
field and multi-generational laboratory exposures.
M.Sc. Thesis. St Cloud State University, St Cloud,
Minnesota.

Elliott, S. M., M E. Brigham, R. L Kiesling, H
l. Schoenfuss. 2018. Environmentally relevant
chemical mixtures of concern in waters of the United
States tributaries to the Great Lakes. Integrated
Environmental Assessment and Management.
9999:1 -10. https://doi.Org/10.1371 /journal,
pone.0182868.

Gill, S. 2019. Effects of a mixture of contaminants of
emerging concern found in agricultural waterways on
the freshwater mussel Lampsilis cardium and host
fish Micropterus salmoides. M.Sc. Thesis. Central
Michigan University, Mount Pleasant, Michigan.

Rappold, J.C. 2019. Effects of a mixture of urban
contaminants of emerging concern on Lampsilis
cardium in a laboratory setting and waste water
treatment plant discharge influence on field deployed
Amblema plicata from the Great Lakes Region.
M.Sc. Thesis. Central Michigan University, Mount
Pleasant, Michigan.

Rzodkiewicz, L.D. 2019. Contaminants of emerging
concern exposure may alter unionid reproductive
success. M.Sc. Thesis. Central Michigan University
Mount Pleasant, Michigan.

Wang, L. 2017. Three generational exposure of
Pimphales promelas to an urban contaminants of
emerging concern mixture. M.Sc. Thesis. St Cloud
State University, St Cloud, Minnesota.

Product Topics in Development

Laboratory effects assessment of Agricultural and
Urban CEC mixture exposure to fathead minnows
over multiple generations .

Laboratory effects assessment of Agricultural and
Urban CEC mixture exposures to native freshwater
mussels from glochidia to larval stages and their host
fish.

Native freshwater mussel microbiome changes with
exposure to CEC mixtures.

Native freshwater mussel transformation rate
modeling after exposure to CEC mixtures.

Understanding the Ecological Consequences of
Ubiquitous Contaminants of Emerging Concern in
the Laurentian Great Lakes Watershed: A Continuum
of Evidence from the Laboratory to the Environment.

Maumee River 2016 Photo ci

GLRI CEC 2016

15

Phase 2 Progress Report


-------

-------