Using whole-genome sequencing and a One Health approach to understand the epidemiology of zoonotic pathogens and antimicrobial resistance at the
human, wildlife, environmental, and livestock interface in southern Ontario
by
Nadine Anna Vogt
A Thesis
presented to
The University of Guelph
In partial fulfilment of requirements for the degree of
Doctor of Philosophy
in
Population Medicine
Guelph, Ontario, Canada
© Nadine Anna Vogt, June 2021
ABSTRACT
USING WHOLE-GENOME SEQUENCING AND A ONE HEALTH APPROACH TO UNDERSTAND THE EPIDEMIOLOGY OF ZOONOTIC PATHOGENS AND
ANTIMICROBIAL RESISTANCE AT THE HUMAN, WILDLIFE, ENVIRONMENTAL, AND LIVESTOCK INTERFACE IN SOUTHERN ONTARIO
Dr. Nadine Anna Vogt
University of Guelph, 2021
Advisor(s):
Dr. David L. Pearl
Dr. Claire M. Jardine
This thesis examines the role of racoons in the ecology of Campylobacter, Salmonella, E. coli
and associated antimicrobial resistance, using wildlife and environmental samples from a
previous repeated cross-sectional study of wildlife on swine farms and conservation areas in
southern Ontario, along with isolates collected through public health surveillance for Salmonella
and E. coli. Using whole-genome sequencing data from Salmonella and E. coli isolates,
microbial population structure was assessed using core-genome multi-locus sequence types
(cgMLST; 3002- and 2513-loci Enterobase schemes, respectively), and antimicrobial resistance
(AMR) genes and plasmid replicons were identified using in silico tools. Outputs from
bioinformatics analyses were used in logistic regression analyses to assess associations with
source type, location type, and farm location. Potential transmission of Salmonella and E. coli
was assessed based on similarities between cgMLST types, as well as examination of extensively
multi-drug resistant isolates. Based on results from mixed multivariable logistic regression,
Campylobacter carriage among raccoons was associated with climatic variables such as season,
temperature, and rainfall, but not with demographic factors (age, sex), or location type (swine
farm vs. conservation area). Assessment of Salmonella from human, livestock, wildlife, soil, and
iii
water samples revealed that Salmonella obtained from raccoons and soil on farm sites were more
similar, and, on average, demonstrated fewer allelic differences than water, human, and livestock
isolates obtained from the broader geographic region. Apart from one Salmonella Heidelberg
isolate that differed at 4 cgMLST loci, no raccoon Salmonella isolates clustered together with
human Salmonella isolates. Comparisons of Salmonella and E. coli isolates from raccoons, soil
and swine manure pits suggest that on- and between-farm transmission may be occurring.
Results from statistical analyses revealed associations between certain genes and plasmid
replicons with source; in many cases, livestock and water samples were significantly more likely
to harbour resistance genes and plasmid types compared to wildlife. Overall, our findings
underscore the complexity of AMR and pathogen transmission in the ecosystem, and suggest that
anthropogenic sources such as conservation areas, swine farms, and contaminated water
represent potential sources of these organisms for wildlife in this region.
iv
DEDICATION
“Creativity and ego cannot go together. If you free yourself from the comparing and jealous
mind your creativity opens up endlessly. Just as water springs from a fountain, creativity springs
from every moment. You must not be your own obstacle. You must not be owned by the
environment you are in. You must own the environment, the phenomenal world around you. You
must be able to freely move in and out of your mind. This is being free. There is no way you can’t
open up your creativity. There is no ego to speak of. That is my belief.”
-Zen Buddhist nun, Jeong Kwan
To Petey, who taught me that one of the most difficult truths is regret.
v
ACKNOWLEDGEMENTS
I am so fortunate to have had the opportunity to pursue a PhD alongside a team of
talented, hard-working, and passionate colleagues that has allowed me to grow both as a
researcher and as a person. Thank you to my advisory committee, David Pearl, Claire Jardine,
Nicole Ricker, and Mike Mulvey for their guidance, support, and expertise. Mike, thank you for
your critical eye, helpful feedback, and collaboration from afar. Nicole, thank you for jumping
into this project head-first, and for your feedback and support. David, thank you for
accommodating the many impromptu meetings in your busy schedule, allowing me to barrage
you with modeling questions. Claire, I am grateful to you for your encouragement, and your
support in my acquiring training and exposures through workshops and conferences, to help
prepare me to work on this project. I would also like to extend a special thank-you to David and
Claire for their emotional support and compassion during a particularly challenging time in my
life and for providing much-needed words of encouragement and advice.
This project would undoubtedly not have been possible without the expertise and support
of my good friend and colleague Benjamin Hetman. When I decided to tackle a project spanning
two different research fields, I had little grasp of what that would entail, and perhaps naively
assumed it wouldn’t be “that bad”. How wrong I was! Ben, I’m not sure you knew what you
were taking on, but I definitely couldn’t have been more “new” to coding and command line.
Throughout the many ups and downs, Ben, you permitted me the necessary independence that I
needed, and allowed me to experience failures first-hand, only to be there to pick me up when I
needed clarification and understanding. Your unwavering support and vision for this project is
what kept me going when I felt like quitting (often!). I am deeply indebted to you and look
forward to future collaborations together. I only hope that one day I can repay your kindness and
support.
Sam, my good friend, and colleague¾who could have guessed that when we met in first
year veterinary school that we would both be here doing PhDs together? How serendipitous that
I asked to spend a day “in the field” back in 2012, while you were trapping raccoons for this
project for another PhD student¾little did I know that I would find myself behind a computer
vi
screen with data from those very animals for my own PhD project many years later. Thank you
for sharing your resilience, your passion, and your commitment to the health and well-being of
wildlife, and animals in general.
I would not be the same person I am today without the support and mentorship of Jan
Sargeant along the way. I don’t know what I would have done if not for the opportunities to
engage in side projects (a.k.a. “passion projects”) that have helped me to find my future direction
as a researcher. Jan was always encouraging, kind, and willing to lend an ear to my sympathies,
and I will always cherish our conversations about humans, animals, epidemiology, and (most
importantly!) outdoor adventures. You are the first to congratulate me on a job well done, but
also there to offer me honest feedback when needed. Jan, you have taught me many important
life lessons, and I look forward to continuing our collaborations together!
I would like to thank my epi peers and friends, Melissa, Lee, Angie, Emily and others, for
their support and for being a source of inspiration and friendship during what seemed like an
unending project. I also don’t know where I would be without the support of my friends and
family. Heidi, Ru, Andy, Asha, Ivanca, Kirstie, Kathryn, Rae, Scott, Emma, Tara, Randy and
Dor, it’s because of all of you that I remembered to slow down, laugh, and enjoy the sweeter
parts of life during a grueling three and a half years. Heidi, thanks for reminding me that life
can’t be all about words on a page, and for joining me on countless adventures in the wilderness,
near and far.
Adam, my brother, my best friend, I never imagined that we would be here together,
working towards a collective goal of bettering the world, one line of coding at a time (thank
goodness for Remote Desktop!). I am so grateful to have shared this project with you, and for the
endless support you provided in the countless iterations of bioinformatics troubleshooting, data
cleaning and extraction, as well as in devising conceptual approaches for this project. You helped
me to find strength I never knew I had, both in technical computer skills, and in other areas of
my life. You have led by example, and I have learned monumental life lessons because of you.
This thesis would not have been possible without the steadfast support of my family, my
in-laws, and extended family. With so many bumps in the road, my family was always there to
vii
believe in me and to allow me to make mistakes and grow. A special thanks to my parents for
graciously allowing us the use of their cottage during the pandemic, and thanks to both my
parents and my in-laws for helping care for our pets without hesitation. I’m lucky to have a sister
who is always game to meet me somewhere in the world after a conference or workshop and take
on an adventure, in the form of “type-II” fun! I am also so fortunate to have a furry and feathered
family in Mick, Smokey, Rosie, Petey (may he rest in peace) and Sakura. I couldn’t imagine a
“worse” time to get a dog, and yet it turned out to be the best time! Saku, thank you for being the
best dog a girl could ask for, and for unexpectedly walking into my life at a time when I needed
you the most.
Finally, the person I am most thankful for and indebted to is my husband Christian. I
don’t have words to express my eternal gratitude for your constant encouragement, and your
unwavering belief in me. You unselfishly provided advice, acted as a sounding board, and
impressively juggled both your own PhD work and our collaborations together. I can only hope
to emulate your chess-like approach to critical thinking, and your incredible ability to take a
complicated issue and distill it down into a simple problem. You have taught me important
research skills, but more importantly, you have taught me to find and stand by truth in my life. I
couldn’t ask for a better partner in crime.
viii
STATEMENT OF WORK
Dr. Nadine A. Vogt performed all data cleaning, statistical analyses, conceptualization of
thesis chapters, and the writing of this thesis in its entirety, with edits and input from Benjamin
Hetman, and Drs. David Pearl, Claire Jardine, Nicole Ricker, and Michael Mulvey. Isolation of
Campylobacter was performed by colleagues at the Canadian Campylobacter laboratory in
Lethbridge, Alberta. Isolation of Salmonella and E. coli for the wildlife study was performed by
colleagues at the Canadian Research Institute in Food Safety (CRIFS). Selection of Salmonella
and E. coli isolates from the wildlife study for sequencing was performed by Drs. Jane Parmley
and Nadine Vogt. Coordination of DNA extraction was performed by Drs. Nadine Vogt and
Claire Jardine, with assistance from Andrea Desruisseau, Gabhan Chalmers, and Amrita Bharat.
DNA extraction was performed by Gabhan Chalmers at the University of Guelph, and by
colleagues at the National Microbiology Laboratory in Winnipeg. Sequencing was performed by
colleagues at the National Microbiology Laboratories in Guelph, and in Winnipeg. Acquisition
of FoodNet livestock data from the Public Health Agency of Canada and human data from
Public Health Ontario was performed by Dr. Nadine Vogt, with assistance from Dr. Claire
Jardine, Dr. Richard Reid-Smith, James Robertson, Kim Ziebell, Stephanie Kadykalo, Andrea
Nesbitt, Manon Fleury, Laura Mataseje, and Amrita Bharat. The bioinformatics pipeline was
developed by Benjamin Hetman, and all raw data were processed by Dr. Nadine Vogt with
technical assistance from Benjamin Hetman and Adam Vogt. Code used for cleaning
bioinformatics data was developed by Adam Vogt and overseen by Dr. Nadine Vogt. The
generation of in silico antimicrobial resistance data for Chapter 5 was performed by Dr. Nadine
Vogt and Adam Vogt and overseen by Dr. Nicole Ricker. Generation of core genome sequence
typing data for Chapters 3-5 was performed by Benjamin Hetman.
STUDY DESIGN and FUNDING: Funding for sequencing of Salmonella and E. coli
samples previously obtained through a longitudinal study of wildlife was provided by a grant
from the Ontario Ministry of Agriculture, Food and Rural Affairs (OMAFRA), and study design
for this grant was developed by Drs. Richard Reid-Smith, Jane Parmley, David Pearl, Claire
Jardine, and Michael Mulvey. Funding for the original longitudinal wildlife study was provided
by the National Sciences and Engineering Research Council (NSERC) and the United States
ix
Department of Agriculture (USDA). Dr. Nadine Vogt was supported by an OVC Scholarship, a
Graduate Tuition Scholarship from the University of Guelph, an Ontario Graduate Scholarship
(2019), and a National Sciences and Engineering Research Council-PGS-Doctoral Scholarship
(2020-21).
DATA COLLECTION: Collection of samples for the wildlife study was performed by Dr.
Samantha Allen, Dr. Kristin Bondo, Erin Harkness, Samantha Kagan and Mary Thompson.
Bryan Bloomfield engaged the participation of landowners, and Tami Harvey and Barbara
Jefferson submitted laboratory samples. Collection of livestock and water samples for Chapters 4
& 5 were performed as part of the FoodNet surveillance program developed by the Public Health
Agency of Canada. Collection of human data in Chapter 5 was performed by Public Health
Ontario.
RESULTS SHARING: Chapter 2 has been published in Zoonoses and Public Health (2020;
doi: 10.1111/zph.12786). Findings from Chapter 2 were presented at the International
Conference on Diseases in Nature Communicable to Man in 2020, and at the Conference of
Research Workers in Animal Diseases in 2020. Results from Chapter 5 were presented at the
Graduate Student Research Symposium at the Ontario Veterinary College in 2020.
x
TABLE OF CONTENTS
Abstract ........................................................................................................................................... ii
Dedication ...................................................................................................................................... iv
Acknowledgements ......................................................................................................................... v
Statement of work ........................................................................................................................ viii
Table of Contents ............................................................................................................................ x
List of Tables ................................................................................................................................ xv
List of Figures .............................................................................................................................. xix
List of Abbreviations .................................................................................................................... xx
1 Introduction, literature review, and study rationale ......................................................... 1
1.1 Introduction ..................................................................................................................... 1
1.2 Literature review ............................................................................................................. 5
1.2.1 Epidemiology of foodborne pathogens in humans ................................................. 5
1.2.2 Mechanisms and transmission of antimicrobial resistance ................................... 14
1.2.3 Molecular epidemiology ....................................................................................... 20
1.3 Study rationale and objectives ...................................................................................... 29
1.4 References ..................................................................................................................... 33
2 Epidemiology of Campylobacter jejuni in raccoons (Procyon lotor) on swine farms and
in conservation areas in southern Ontario ............................................................................... 50
2.1 Abstract ......................................................................................................................... 50
2.2 Introduction ................................................................................................................... 51
2.3 Methods......................................................................................................................... 53
2.3.1 Sample collection .................................................................................................. 53
2.3.2 Campylobacter isolation and speciation ............................................................... 54
xi
2.3.3 Statistical methods ................................................................................................ 55
2.3.4 Assessing model fit ............................................................................................... 57
2.3.5 Spatial scan statistics ............................................................................................. 57
2.4 Results ........................................................................................................................... 58
2.4.1 Samples collected and isolates .............................................................................. 58
2.4.2 Univariable and multivariable analyses ................................................................ 58
2.4.3 Spatial scan statistics ............................................................................................. 60
2.5 Discussion ..................................................................................................................... 60
2.6 References ..................................................................................................................... 66
2.7 Tables ............................................................................................................................ 70
3 Genomic epidemiology of Salmonella and Escherichia coli and associated
antimicrobial resistance in raccoons, swine manure pits, and soil samples on swine farms in
southern Ontario: A risk factor analysis .................................................................................. 79
3.1 Abstract ......................................................................................................................... 79
3.2 Introduction ................................................................................................................... 81
3.3 Methods......................................................................................................................... 83
3.3.1 Sample collection .................................................................................................. 83
3.3.2 Culture and susceptibility testing .......................................................................... 84
3.3.3 DNA extraction, whole-genome sequencing and genome assembly .................... 85
3.3.4 Analysis of whole-genome assemblies ................................................................. 85
3.3.5 Statistical analyses ................................................................................................ 87
3.4 Results ........................................................................................................................... 88
3.4.1 Dataset ................................................................................................................... 88
3.4.2 Distribution of serovars and MLST types ............................................................. 89
3.4.3 Population structure based on cgMLST ................................................................ 90
xii
3.4.4 In silico determination of acquired AMR genes and plasmid replicons ............... 91
3.4.5 Comparison of Salmonella and E. coli ................................................................. 93
3.4.6 Sensitivity and specificity of in silico AMR prediction ........................................ 93
3.4.7 Statistical results ................................................................................................... 94
3.5 Discussion ..................................................................................................................... 95
3.6 Conclusions ................................................................................................................. 102
3.7 References ................................................................................................................... 103
3.8 Tables .......................................................................................................................... 108
3.9 Figures......................................................................................................................... 123
4 Genomic epidemiology of antimicrobial resistance in Escherichia coli isolates from
wildlife, livestock, and environmental sources in southern Ontario, and the potential for
transmission: A risk factor analysis ........................................................................................ 129
4.1 Abstract ....................................................................................................................... 129
4.2 Introduction ................................................................................................................. 130
4.3 Methods....................................................................................................................... 133
4.3.1 Dataset ................................................................................................................. 133
4.3.2 Culture and susceptibility testing ........................................................................ 136
4.3.3 DNA extraction, whole-genome sequencing, and genome assembly ................. 137
4.3.4 Analysis of whole-genome assemblies ............................................................... 137
4.3.5 Statistical analyses .............................................................................................. 139
4.4 Results ......................................................................................................................... 140
4.4.1 Description of dataset ......................................................................................... 140
4.4.2 Distribution of serovars and MLST types ........................................................... 141
4.4.3 Population structure based on cgMLST .............................................................. 142
4.4.4 In silico determination of acquired AMR genes and plasmid Inc types ............. 142
xiii
4.4.5 Sensitivity and specificity of in silico AMR prediction ...................................... 143
4.4.6 Associations between source type, location type, and the carriage of resistance genes and plasmid replicons ............................................................................................... 143
4.5 Discussion ................................................................................................................... 144
4.6 Conclusions ................................................................................................................. 150
4.7 References ................................................................................................................... 152
4.8 Tables .......................................................................................................................... 157
4.9 Figures......................................................................................................................... 170
5 Genomic epidemiology of non-typhoidal Salmonella enterica and associated
antimicrobial resistance in raccoons (Procyon lotor), humans, livestock, and environmental
sources in southern Ontario, and the potential for zoonotic transmission: A risk factor
analysis ....................................................................................................................................... 180
5.1 Abstract ....................................................................................................................... 180
5.2 Introduction ................................................................................................................. 181
5.3 Methods....................................................................................................................... 184
5.3.1 Dataset ................................................................................................................. 184
5.3.2 Salmonella culture and phenotypic susceptibility testing ................................... 187
5.3.3 DNA extraction, whole-genome sequencing and genome assembly .................. 188
5.3.4 Analysis of whole-genome assemblies ............................................................... 189
5.3.5 Statistical analyses .............................................................................................. 191
5.4 Results ......................................................................................................................... 192
5.4.1 Description of dataset ......................................................................................... 192
5.4.2 Serovar distribution ............................................................................................. 193
5.4.3 MLST results ...................................................................................................... 194
5.4.4 Population structure based on cgMLST .............................................................. 195
5.4.5 In silico determination of acquired AMR genes and plasmid replicons ............. 196
xiv
5.4.6 Sensitivity and specificity of in silico AMR identification ................................ 197
5.4.7 Associations between source type and carriage of genes and plasmid replicons 197
5.5 Discussion ................................................................................................................... 198
5.6 Conclusions ................................................................................................................. 206
5.7 References ................................................................................................................... 208
5.8 Tables .......................................................................................................................... 214
5.9 Figures......................................................................................................................... 227
6 Summary discussion and conclusions ............................................................................. 234
6.1 Summary discussion ................................................................................................... 234
6.1.1 Major findings ..................................................................................................... 236
6.1.2 Limitations and future work ................................................................................ 239
6.2 Conclusions ................................................................................................................. 241
6.3 References ................................................................................................................... 243
xv
LIST OF TABLES
Table 2.1: Proportion (%) of raccoon fecal swabs testing positive for Campylobacter jejuni overall and by age, sex, location type, season, and year in southern Ontario from May-November 2011-2013 (n=1086 samples) ........................................................................... 70
Table 2.2: Results from univariable multi-level logistic regression models showing associations between the occurrence of Campylobacter jejuni in raccoon fecal swabs and raccoon sex and age, year, season, location type, mean air temperature, and total rainfall in Ontario, Canada (n=1086) ........................................................................................................................................ 71
Table 2.3: Results from multi-level multivariable logistic regression models showing associations between the occurrence of Campylobacter jejuni in raccoon fecal swabs and year, season, air temperature, total rainfall and interaction effects in Ontario, Canada (n=1086) ........ 73
Table 2.4: Contrasts derived from the multi-level logistic regression model for the occurrence of Campylobacter jejuni in raccoon fecal swabs (Table 2.3) in Ontario, Canada to interpret interaction effects between season and mean air temperature in the 14 days prior to sampling (n=1086) ........................................................................................................................................ 75
Table 2.5: Contrasts derived from the multi-level logistic regression model for the occurrence of Campylobacter jejuni in raccoon fecal swabs (Table 2.3) in Ontario, Canada to interpret interaction effects between year and total rainfall in the 14 days prior to sampling (n=1086) .... 76
Table 2.6: Contrasts derived from the multi-level logistic regression model for the occurrence of Campylobacter jejuni in raccoon fecal swabs (Table 2.3) in Ontario, Canada to interpret interaction effects between year and mean air temperature in the 14 days prior to sampling (n=1086) ........................................................................................................................................ 77
Table 2.7: Statistically significant spatial, temporal and space-time clusters of Campylobacter jejuni positive raccoon fecal swabs in Ontario, Canada using spatial scan statistics with Bernoulli models (n=1086) ........................................................................................................................... 78
Table 3.1: Resistance genes and plasmid replicons identified in silico in antimicrobial resistant Salmonella enterica from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013 (n=159)............................................................................108
Table 3.2: Frequency of Salmonella enterica legacy multi-locus sequence types of isolates obtained from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013 (n=159) ........................................................................................................ 109
Table 3.3: Multi-locus sequence types identified using whole-genome sequencing data from Escherichia coli isolates obtained from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013 (n=96) ............................................................... 110
xvi
Table 3.4: Serovars identified using whole-genome sequencing data from Escherichia coli isolates obtained from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013 (n=96) ............................................................................................ 111
Table 3.5: Frequencies of plasmid replicon types identified using whole-genome sequencing data from Salmonella enterica and Escherichia coli isolates obtained from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013 ........................ 112
Table 3.6: Frequencies of acquired antimicrobial resistance genes identified using whole-genome sequencing data from Escherichia coli isolates obtained from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013 (n=96) ........................... 113
Table 3.7: Test sensitivity and specificity for in silico identification of acquired antimicrobial resistance genes in Salmonella enterica and Escherichia coli isolates obtained from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013 ......................................................................................................................................................114
Table 3.8: Univariable logistic regression models assessing the association between source type, farm location, and year of sampling and the occurrence of select antimicrobial resistance genes and plasmid replicons in Salmonella enterica isolates from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013 (n=159) ................................ 115
Table 3.9: Contrasts from univariable logistic regression models assessing the association between source type, farm location, and year of sampling and the occurrence of select antimicrobial resistance genes and plasmid replicons in Salmonella enterica isolates from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013 (n=159) ............................................................................................................................... 118
Table 3.10: Univariable logistic regression models assessing the association between source type, farm location, and year of sampling, and the occurrence of select antimicrobial resistance genes and plasmid replicons in Escherichia coli isolates obtained from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013 (n=96) ............ 119
Table 4.1: Distribution of sequence types of international importance identified in antimicrobial resistant Escherichia coli isolates collected from wildlife, and environmental sources in southern Ontario, Canada 2011-2013 (n=200) ..........................................................................................157
Table 4.2: Multi-locus sequence types identified using whole-genome sequencing data by source type for antimicrobial resistant Escherichia coli isolates from wildlife, livestock, and environmental sources in southern Ontario, Canada 2011-2013 (n=200) .................................. 158
Table 4.3: Serovars identified using whole-genome sequencing data for antimicrobial resistant Escherichia coli isolates from wildlife, livestock, and environmental sources in southern Ontario, Canada 2011-2013 (n=200) ........................................................................................................ 159
xvii
Table 4.4: Frequencies of antimicrobial resistance genes identified from whole-genome sequencing data of antimicrobial resistant Escherichia coli isolates from wildlife, livestock, and environmental sources in southern Ontario, Canada 2011-2013 (n=200) .................................. 160
Table 4.5: Distribution of plasmid incompatibility (Inc) types identified using whole-genome sequencing data by source type for antimicrobial resistant Escherichia coli isolates from wildlife, livestock, and environmental sources in southern Ontario, Canada 2011-2013 (n=200) ........... 162
Table 4.6: Test sensitivity and specificity for in silico identification of antimicrobial resistance genes in antimicrobial resistant Escherichia coli isolates from wildlife, livestock, and environmental sources in southern Ontario, Canada 2011-2013 (n=200) .................................. 163
Table 4.7: Logistic regression models assessing the association between source type and the occurrence of select plasmid incompatibility types and antimicrobial resistance genes in antimicrobial resistant Escherichia coli isolates collected from wildlife, livestock and environmental sources in southern Ontario, 2011-2013 (n=200, dataset A) .............................. 164
Table 4.8: Contrasts from logistic regression models assessing the association between source type and the occurrence of select antimicrobial resistance genes in antimicrobial resistant Escherichia coli isolates collected from wildlife, livestock and environmental sources in southern Ontario, 2011-2013 (n=200, dataset A) ...................................................................................... 167
Table 4.9: Logistic regression models assessing the association between source type, location type, and the occurrence of select plasmid incompatibility types and antimicrobial resistance genes in antimicrobial resistant Escherichia coli isolates from wildlife and soil samples collected on swine farms and conservation areas in southern Ontario, 2011-2013 (n=146, dataset B) ..... 168
Table 5.1: Distribution of Salmonella enterica serovars from raccoons, humans, livestock, and environmental sources in southern Ontario, 2011-2013 (n=608)............................................... 214
Table 5.2: Counts of most frequently identified sequence types of Salmonella enterica isolates from raccoons, livestock, and environmental sources in southern Ontario, 2011-2013 (n=608) ..... ..................................................................................................................................................... 215
Table 5.3: Distribution of Salmonella enterica legacy multi-locus sequence types for isolates obtained from raccoons, humans, livestock, and environmental sources in southern Ontario, 2011-2013 (n=608) ..................................................................................................................... 216
Table 5.4: Frequencies of antimicrobial resistance genes identified from whole-genome sequencing data of Salmonella enterica isolates from raccoons, humans, livestock, and environmental sources in southern Ontario, Canada, 2011-2013 (n=608) ................................. 218
Table 5.5: Distribution of plasmid replicon types identified using whole-genome sequencing data by source type for Salmonella enterica isolates from raccoons, humans, livestock, and environmental sources in southern Ontario, Canada, 2011-2013 (n=608) ................................. 221
xviii
Table 5.6: Test sensitivity and specificity for in silico identification of antimicrobial resistance genes in Salmonella enterica isolates from raccoons, humans, livestock, and environmental sources in southern Ontario, 2011-2013 (n=608) ....................................................................... 222
Table 5.7: Logistic regression models assessing the association between source type and the occurrence of select antimicrobial resistance genes and plasmid replicons in Salmonella enterica isolates from raccoons, humans, livestock, and environmental sources in southern Ontario, 2011-2013 (n=608) ............................................................................................................................... 223
Table 5.8: Contrasts from logistic regression models assessing the association between source type and the occurrence of select antimicrobial resistance genes and plasmid replicons in Salmonella enterica isolates collected from raccoons, humans, livestock, and environmental sources in southern Ontario, 2011-2013 (n=608) ....................................................................... 225
xix
LIST OF FIGURES
Figure 3.1. Population structure of 159 Salmonella enterica isolates from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario based on 3002-loci cgMLST scheme from Enterobase. ............................................................................................................ 124
Figure 3.2. Population structure of 64 isolates of Salmonella enterica isolates (serovars Agona, Infantis, Typhimurium, Poona) from raccoons, swine manure pits, and soil on swine farms in southern Ontario based on 3002-loci cgMLST scheme from Enterobase. ................................. 126
Figure 3.3. Population structure of 96 Escherichia coli isolates from wildlife, soil, and swine manure on swine farms in southern Ontario based on 2513-loci cgMLST scheme from Enterobase. .................................................................................................................................. 128
Figure 4.1. Classification of source types for overall dataset (A), and subset of data (B), with sample sizes and independent variables analyzed...................................................................... 170
Figure 4.2. Causal diagram for analytical modeling. .................................................................. 171
Figure 4.3. Population structure of 200 Escherichia coli isolates from wildlife, livestock, and environmental sources in southern Ontario based on 2513-loci cgMLST scheme from Enterobase. .................................................................................................................................. 175
Figure 4.4. Population structure of 146 Escherichia coli isolates from wildlife and soil on swine farms and conservation areas in southern Ontario based on 2513-loci cgMLST scheme from Enterobase. .................................................................................................................................. 179
Figure 5.1. Causal diagram for analytical modeling................................................................... 227
Figure 5.2. Population structure of 608 Salmonella enterica isolates from raccoons, livestock, humans, and environmental sources in southern Ontario based on 3002-loci cgMLST scheme from Enterobase .......................................................................................................................... 230
Figure 5.3. Heatmap of all serovars identified in humans and their proportional occurrence in Salmonella enterica isolates from raccoons, livestock, and environmental sources in southern Ontario, 2011-2013 (n=608) ....................................................................................................... 232
Figure 5.4. Population structure of 127 Salmonella enterica isolates (serovars Newport, Typhimurium, Infantis) from raccoons, livestock, humans, and environmental sources in southern Ontario based on 3002-loci cgMLST scheme from Enterobase .................................. 233
xx
LIST OF ABBREVIATIONS
AIC ¾ Akaike’s Information Criterion AMC ¾ Amoxicillin-clavulanic acid AMP ¾ Ampicillin AZM ¾ Azithromycin AMR ¾ Antimicrobial resistance BIC ¾ Bayesian Information Criterion BLUPs ¾ Best linear unbiased predictions CHL ¾ Chloramphenicol CI ¾ Confidence interval CIP ¾ Ciprofloxacin cgMLST ¾ Core genome multi-locus sequence typing CIPARS ¾ Canadian Integrated Program for Antimicrobial Resistance Surveillance DT ¾ Definitive type ESBL ¾ Extended spectrum beta-lactamase FOX ¾ Cefoxitin GEN ¾ Gentamicin KAN ¾ Kanamycin MLST ¾ Multi-locus sequence typing NAL ¾ Nalidixic acid NARMS ¾ National Antimicrobial Resistance Monitoring Systems OR ¾ Odds ratio PFGE ¾ Pulsed-field gel electrophoresis SOX ¾ Sulfisoxazole ST ¾ Sequence type STR ¾ Streptomycin SXT ¾ Trimethoprim-sulfamethoxazole TIO ¾ Ceftiofur TCY ¾ Tetracycline VPC ¾ Variance partition coefficient WGS ¾ Whole genome sequencing
1
CHAPTER 1
1 Introduction, literature review, and study rationale
1.1 Introduction
The global burden of foodborne illness caused by bacterial agents was estimated by the World
Health Organization to be 18 million disability adjusted life years in 2010 (World Health
Organization, 2015). Typhoidal Salmonella, non-typhoidal Salmonella, pork tapeworm, and
enteropathogenic Escherichia coli (EPEC) are the predominant agents responsible for this global
burden (World Health Organization, 2015). Since the 1980s, the incidence of foodborne illness
has increased by over 3300-fold, and over 600-fold in the United States and Canada, respectively
(Todd, 1990; CDC, 2020; Government of Canada, 2019). Dramatic changes in and centralization
of food production, processing, and distribution are believed to contribute to this increasing
incidence of foodborne illness in industrialized countries (Institute of Medicine, 1998). As a
result of centralization and specialization, livestock farms are increasingly large-scale operations
focused on a single species, resulting in fewer and larger herds (Institute of Medicine, 1998).
Although some agents have long been recognized as causes of foodborne illness (e.g., non-
typhoidal Salmonella), other pathogens, such as Campylobacter spp. and E. coli 0157 have only
recently been identified as important causative agents of foodborne illness (Riley et al., 1983;
Mintz, 2003). The majority of foodborne infections result in self-limiting gastroenteritis,
however, the medical treatment of severe cases of foodborne illness may be complicated by
2
antimicrobial resistance (AMR; Michael & Schwarz, 2016). The incidence and prevalence of
antimicrobial resistant infections has been rapidly increasing worldwide ever since the initial
discovery of this phenomenon in the 1930s (Davies & Davies, 2010). Prior speculation that
humans are entering a post-antibiotic era has become a widely recognized reality, with resistance
mechanisms identified for every clinically available antimicrobial (McDermott et al., 2002).
Antimicrobials with potentially lethal side effects (e.g., colistin) are increasingly being used to
treat resistant bacterial infections that were, historically, treatable with antimicrobials that cause
few to no side effects (Ryan et al., 1969; Singh, 2013; World Health Organization, 2015). Aside
from increased mortality, antimicrobial resistant infections can negatively impact recovery times,
increase the risk of complications, and can result in permanent, long-term health effects (Tanwar
et al., 2014). Together, foodborne illness and AMR represent a substantial burden to public
health, with negative social and economic implications (World Health Organization, 2015).
In recent decades, produce (i.e., fruits and vegetables) is increasingly playing a role as a
source of foodborne pathogens (CSPI, 2006). In the United States between 1990 and 2004,
31,496 cases of illness were attributed to produce, followed by poultry (16,280 cases), beef
(13,220 cases), eggs (11,027 cases), and seafood (9,969 cases; CSPI, 2006). Public education
concerning safe food handling practices are likely to have played a role in the reduction of the
burden of foodborne illness from animal products (Kendall et al., 2017), but the increasing
incidence of cases related to produce is puzzling and concerning (Carstens et al., 2019). The
contamination of produce with organisms such as Listeria, Salmonella, and E. coli 0157, as well
as protozoan parasites, is particularly concerning since vegetables and fruits are usually
consumed raw and rinsing of contaminated produce prior to ingestion cannot remove harmful
microbes which are embedded within the plant itself (Carstens et al., 2019; Whipps et al., 2008).
3
With the increasing popularity of One Health research approaches that consider humans and
animals within a shared environment, wildlife are increasingly coming under scrutiny for their
possible role in the maintenance and transmission of causative agents of foodborne illness in
humans (Fredriksson-Ahomaa, 2018; Singh & Gajadhar, 2014; Silva et al., 2014). To date, there
is considerable evidence demonstrating that avian and mammalian wildlife are capable of
carrying bacteria that are responsible for causing foodborne illness in humans (Hald et al., 2016;
Arrausi-Subiza et al., 2016; Facciola et al., 2017; Greig et al., 2015). Further, a number of
epidemiological investigations have identified the presence of wildlife living near agricultural
settings as a risk factor for the isolation of bacteria and/or antimicrobial resistant organisms from
produce fields and livestock animals (Strawn et al., 2013; Cernicchiaro et al., 2012; Hassel et al.,
2019). These approaches, however, have typically been limited to prevalence surveys and risk
factor analyses that provide evidence of putative associations, as opposed to hypothesis-testing
studies that are focused on determining whether there is evidence of transfer between wildlife
and humans on the basis of molecular typing data (Greig et al., 2015; Ramey & Ahlstrom, 2020).
With the decreased cost of DNA sequencing, increased throughput, and the increasing
availability of open-source bioinformatics tools, genomic data are quickly becoming the new
standard for epidemiologic research addressing questions of source attribution (Besser, 2018).
Thus, there is a need to employ these highly discriminatory genomic data in epidemiological
investigations comparing wildlife and human bacterial isolates, in order to bridge the fields of
genomics and epidemiology (Croucher et al., 2013).
The aim of this thesis is to examine the role of raccoons as potential sources of
Campylobacter, Salmonella, and AMR determinants. Raccoons represent a prevalent wildlife
species that often occupy environments that are shared by humans, livestock, and other domestic
4
animals such as pets (Rosatte et al., 2010). As opportunistic foragers, raccoons commonly
inhabit both urban and rural landscapes, and have been shown to reach population densities
upwards of 70 raccoons/km2 in urban locations (Prange et al., 2011). Rural raccoons may occupy
vast territories, particularly in prairie environments where food is sparse; territories as large as
50km2 have been documented in Canadian prairie habitats (Rosatte, 2008). Additionally,
raccoons are considered a relatively social species, particularly in urban environments where
high raccoon densities are common (Prange et al., 2011). With this combination of feeding and
behavioural attributes, high prevalence and widespread distribution, raccoons as a species merit
further study in the context of human foodborne illness and AMR. Indeed, a range of bacterial
pathogens have been isolated from raccoons, including Salmonella, Campylobacter, Clostridium
difficile, Listeria monocytogenes, and Yersinia enterocolitica (Bondo et al., 2015; Bondo et al.,
2016; Mutschall et al., 2020; Nowakiewicz et al., 2016). Notably, several of these studies include
data obtained from a southern Ontario population of raccoons-data which will be further
examined within this thesis to address current research gaps regarding the role of raccoons in
foodborne illness and related AMR. Understanding the distribution, transmission, and movement
of AMR determinants and pathogenic bacteria is an important step in the prevention of
foodborne illness, especially for food products such as produce which is typically consumed raw.
Results from this thesis will help to inform future public health policy, farm biosecurity
practices, pathogen reduction programs, and antimicrobial stewardship programs. Using a
combination of statistical modeling and genomic epidemiology to identify risk factors and
examine the population structure of bacteria from raccoons, humans, livestock, and
environmental sources, we hope to gain insights about the movement and transmission of
bacteria and associated AMR within a southern Ontario ecosystem. Ultimately, our aim is to
5
contribute positively to the future health of humans, the environment, and animals, both domestic
and wild. With these issues in mind, the following literature review examines:
1.2.1 Epidemiology of foodborne pathogens in humans
1.2.2 Mechanisms and transmission of antimicrobial resistance
1.2.3 Molecular epidemiology
1.2.4 The role of wildlife in the epidemiology of foodborne pathogens and associated
antimicrobial resistance
1.2 Literature review
1.2.1 Epidemiology of foodborne pathogens in humans
Foodborne illness is caused by the consumption of contaminated foodstuff or beverages,
resulting in various symptoms that may include nausea, vomiting, diarrhea, and/or abdominal
cramps. Worldwide, it has been estimated that 600 million cases of foodborne disease occur
every year, resulting in approximately 420,000 deaths (World Health Organization, 2020).
Although parasites and viruses cause a certain proportion of foodborne illness, bacteria are most
frequently implicated, most commonly: Staphylococcus aureus, Salmonella, Campylobacter,
Listeria monocytogenes, and pathogenic E. coli (Abebe et al., 2020). Most cases of foodborne
illness are self-limiting; however, some patients require hospitalization, supportive care, and/or
antimicrobial treatment (Abebe et al., 2020). Certain bacteria are particularly virulent and result
in a high proportion of hospitalization among infected individuals. For instance, infection with
6
Listeria results in hospitalization of upwards of 90% in some populations (Jemmi & Stephan,
2006). In some instances, serious long-term complications have been linked to foodborne
infections, most notably, Guillain-Barré syndrome, an ascending paralytic condition that is
associated with previous Campylobacter infection (Acheson & Allos, 2001). A recent scoping
review has revealed that although there is a considerable body of literature examining long-term
sequelae with pathogens such as Salmonella and Campylobacter, there is comparatively little
work on pathogenic E. coli, Shigella, Listeria, and Yersinia in this field (Pogreba-Brown et al.,
2020).
There is variability in the incidence of foodborne illness among different segments of the
population, depending on demographic factors such as age and sex (Strassle et al., 2019). A
study of outbreaks from 1998-2015 in the United States revealed that foodborne illness
associated with ingestion of contaminated dairy products occurred most frequently in children
less than 5 years of age (Strassle et al., 2019). Meanwhile, in this same study, men were most
likely to be afflicted with foodborne illness associated with the consumption of beef, pork, and
wild game, and women were most likely to experience foodborne illness in association with the
ingestion of contaminated vegetables, grains, beans and sprouts (Strassle et al., 2019).
Worldwide, the most common sources of foodborne illnesses are thought to be animal products
such as meat, eggs, and dairy products (Abebe et al., 2020). With increasing global awareness of
healthy eating habits that include the regular consumption of fruits and vegetables, the incidence
of foodborne infections associated with produce has markedly increased in recent years (Balali et
al., 2020; Krishnasamy et al., 2020). In the United States, contaminated produce was responsible
for a large proportion of cases of foodborne illness between 1990 and 2004 (n=31,496 cases;
CSPI, 2006), and, in Canada, over 1360 cases of foodborne illness between 1991 and 2000 were
7
linked to produce (Sewell & Farber, 2001). Since vegetables and produce are often consumed
raw or processed minimally, the risk of foodborne illness may be increased as compared with
foods that are cooked to internal temperatures which destroy harmful microorganisms (Carstens
et al., 2019). Furthermore, washing and rinsing of produce prior to consumption is ineffective for
the removal of pathogens that have been internalized within plant tissues, or that are embedded
within biofilms on plant surfaces (Whipps et al., 2008). Given the challenges in reduction or
destruction of microbial contaminants in fresh produce, the identification of potential sources of
contamination is paramount to the prevention of outbreaks associated with produce (Iwu &
Okoh, 2019). Several routes and sources of contamination for produce are well-recognized,
including farm animal reservoirs which may contribute to the contamination of surface waters
used for irrigation, or through the spreading of raw or inadequately composted manure on
produce fields (Jung et al., 2014). The prevalence of pathogenic E. coli and Salmonella enterica
serovars shed in the feces of asymptomatic farm animals such as cattle, sheep, pigs, and poultry
has been shown to be as high as 70% of individuals in a herd (Bailey et al., 2001; Dargatz et al.,
2003; Hussein and Bollinger, 2005; Keen et al., 2006; Doane et al., 2007). Other sources of pre-
harvest contamination of produce may include dust, insects, and fertilizers (Carstens et al.,
2019). Finally, wild animals are also believed to contribute to the transmission of pathogens into
the food chain via the contamination of both irrigation and recreational water, and direct
transmission of pathogens to livestock or produce, but their role is not well understood (Park et
al., 2015; Alegbeleye et al., 2018).
Campylobacter
8
Campylobacter is a genus of Gram-negative flagellar bacteria which prefer microaerophilic
growth conditions (Lamas et al., 2018). Campylobacter jejuni and Campylobacter coli are
responsible for the majority (>90%) of reported clinical cases in humans (Bojanic et al., 2017;
Butzler, 2004). Campylobacteriosis is the leading bacterial cause of foodborne illness in both
Canada and in Europe (Facciola et al., 2017; Pintar et al., 2017), and the second most frequent
bacterial cause of foodborne illness in the United States (Gebreyes et al., 2020). Infection with
Campylobacter typically results in a mild gastroenteritis; the majority of cases do not require
medical intervention. In a small proportion of individuals (<0.1%), previous infection with
Campylobacter is thought to trigger a serious–and potentially life-threatening–autoimmune
condition, Guillain-Barré syndrome (Acheson & Allos, 2001). Other serious medical conditions
have been documented in close temporal association with Campylobacter infection, including
esophageal cancer, colorectal cancer, irritable bowel syndrome, and Barrett’s esophagus (Lamas
et al., 2018), as well as non-gastrointestinal ailments such as lung infections, meningitis, reactive
arthritis, brain abscesses, and bacteremia (Facciola et al., 2017, Kaakoush et al., 2015). Since the
majority of Campylobacter infections cause mild illness, campylobacteriosis is believed to be
vastly underreported, and some estimates suggest that for every case reported, there are an
additional 30 cases within the community (Thomas et al., 2013). Most cases (99%) of
campylobacteriosis occur sporadically, rather than as outbreaks (EFSA, 2016). A considerable
proportion of Campylobacter infections are believed to result from cross-contamination or
indirect exposures (e.g., contaminated water), rather than exposure to primary contaminated
sources, thus source attribution is challenging (Facciola et al., 2017).
In addition to foodborne exposure, Campylobacter infections may also be related to
waterborne exposure or contact with animals. Mammals and birds are natural reservoirs for
9
Campylobacter and may be colonized with this bacterium without showing any overt signs of
clinical illness (Chlebicz & Slizewska, 2018). Campylobacter has a particular predilection for
both wild and domestic avian species due to their high body temperature (Hansson et al., 2018).
Domestic poultry are widely recognized as a major source of human foodborne Campylobacter
infections (Facciola et al., 2017; Pintar et al., 2015). Close contact with farm animals, household
pets, and animals in petting zoos have all been linked to Campylobacter infections (Pintar et al.,
2015). The results from one study even suggest that exposure of young children to playgrounds
contaminated with wild bird feces may be a potential source of infection for that particular
demographic (French et al., 2009). A comparative exposure assessment of Campylobacter in
Ontario revealed that the most common sources of exposure, in order from highest to lowest,
were: household pets, chicken, living on a farm, raw milk, visiting a farm, recreational water,
beef, drinking water, pork, vegetables, seafood, petting zoos, and fruits (Pintar et al., 2017). At
present, there is limited information about whether feeding of raw meat to companion animals
places pet owners at a higher risk of infection (Procter et al., 2014); however, a recent study of
retail raw meat pet food in New Zealand discovered a 28% prevalence of contamination with
Campylobacter (Bojanic et al., 2017). The persistence of Campylobacter in the environment for
long periods of time, and on foods stored at low temperature is a contributing factor to foodborne
infections (Chlebicz & Slizewska, 2018), and one study demonstrated that Campylobacter spp.
can persist on the surfaces of vegetables and fruits for up to eight days (Newell et al., 2016).
Over time, selection pressures have resulted in a growing prevalence of Campylobacter
that exhibit AMR to fluoroquinolones and tetracyclines (Iwu & Okoh, 2019). The widespread
use of fluoroquinolones by the poultry industry is believed to have been a major contributor to
the emergence and persistence of fluoroquinolone-resistant Campylobacter (Moser et al., 2020).
10
Following a ban on the use of this class of antimicrobials in livestock/poultry, this resistance
does not appear to be decreasing (Shiaffino et al., 2019), demonstrating the importance of
antimicrobial resistance prevention. In 2017, the World Health Organization recognized
fluoroquinolone-resistant Campylobacter spp. as the fourth most important human bacterial
pathogen (Shiaffino et al., 2019). Unfortunately, a shift to macrolides as a first-line treatment in
place of fluoroquinolones for Campylobacter infections necessitating antimicrobial therapy is
probably a contributing factor to the recent increasing prevalence of macrolide-resistant
Campylobacter (Shiaffino et al., 2019).
Non-typhoidal Salmonella
Salmonella are a genus of Gram-negative rod-shaped bacteria that inhabit the intestinal
tract of both warm and cold-blooded animals (Michael & Schwarz, 2016). Salmonella consists of
two species, Salmonella enterica and Salmonella bongori, the latter of which is considered non-
pathogenic to endotherms (Forshell & Wierup, 2006). Salmonella enterica is comprised of six
subspecies that are widely recognized as pathogenic in humans and warm-blooded animals.
Salmonella enterica subsp. enterica is responsible for the vast majority of Salmonella infections
in humans (>99%; Desai et al., 2013). To date, over 2400 serovars of Salmonella enterica have
been identified, based on the presence of surface antigens (Forshell & Wierup, 2006). Different
Salmonella serovars exhibit important differences in the host species they are able to infect, as
well as the clinical disease they induce in those species. Salmonella enterica subsp. enterica
serovar Typhi and Salmonella enterica subsp. enterica serovar Paratyphi are adapted specifically
to humans, and are the causative agents for typhoid fever, whereas they do not cause clinical
illness in animals (Chen et al., 2013). Other serovars represented within the subspecies enterica
11
which are not specifically adapted to humans are collectively referred to as non-typhoidal
Salmonella. Many non-typhoidal serovars are specifically adapted to certain animal hosts, and
infected individuals may or may not show clinical illness: Salmonella Abortus Ovis in sheep,
Salmonella Choleraesuis in pigs, Salmonella Gallinarum in poultry, Salmonella Abortus Equi in
horses, and Salmonella Dublin in cattle (Forshell & Wierup, 2006). Although these host-adapted
serovars rarely cause illness in humans, they often result in serious, invasive infections in their
animal hosts (Cheng et al., 2019). Other serovars such as Typhimurium, Enteritidis, Hadar, and
Infantis, among others, are not specifically adapted to certain hosts, and may cause clinical
illness in humans and in a variety of animal species (Forshell & Wierup, 2006). Based on a
global survey conducted between 1990 and 1995, the most common serovars responsible for
causing salmonellosis in humans were Enteritidis and Typhimurium (Herikstad et al., 2002).
These two serovars were also identified among the top 10 most common serovars responsible for
human illnesses in Canada between 2009 and 2014, in addition to the following: Heidelberg,
Thompson, Newport, Infantis, Typhi, Javiana, Saintpaul, and subsp. I:4,[5],12:i:– (Furukawa et
al., 2018). All serovars of Salmonella enterica subsp. enterica are capable of causing illness in
humans, but not all infected humans become clinically ill, and still others become carriers, or are
sub-clinically ill, and can actively shed Salmonella in their feces (Scallan et al., 2011). One
source estimates that between 7 and 66% of humans infected with non-typhoidal Salmonella do
not develop overt clinical illness (Anon, 1999).
The illness caused by non-typhoidal Salmonella varies from mild to serious
gastroenteritis that usually does not require antimicrobial therapy (Chen et al., 2013). More
serious infections can develop if Salmonella invades the bloodstream. These invasive infections
are more likely to occur in vulnerable populations such as the young, elderly and immuno-
12
compromised, and this life-threatening septicemia can result in serious infections elsewhere in
the body, including the bones and nervous system (Chen et al., 2013). Salmonella infections may
also result in long-term sequelae, including reactive arthritis, and neurological illness (Forshell &
Wierup, 2006). Overall, human fatalities due to infection with non-typhoidal Salmonella are rare,
but the increasing prevalence of multi-drug resistant Salmonella presents a major challenge for
the effective treatment of invasive bloodstream infections (Chen et al., 2013).
Non-typhoidal Salmonella is the leading cause of foodborne illness worldwide, causing
illness in both developing and industrialized countries; the global burden of disease is estimated
to be 93.8 million cases annually, with 155,000 deaths (Gebreyes et al., 2020; Majowicz et al.,
2010). In the United States, an estimated 1.4 million infections occur every year (Mead et al.,
1999), and non-typhoidal salmonellosis is the third most common cause of foodborne illness in
Canada (Thomas et al., 2013). The most common route of infection is via foodborne exposure; in
the United States, over 90% of reported non-typhoidal Salmonella cases are foodborne (Scallan
et al., 2011). Although salmonellosis is typically associated with contaminated animal-based
products, outbreaks related to contaminated produce are increasingly prevalent (Jain et al., 2019).
Foods which are commonly contaminated with Salmonella include: meat, eggs, milk, cheese,
shellfish, fresh fruit, fresh juice, spices, chocolate, and vegetables (Rabsch et al., 2013).
Livestock are considered primary reservoirs for human infections, but the prevalence of infected
animals varies markedly by herd, species, and serovar, with estimates ranging between 0 and
90% of animals in a herd (Christensen et al., 2002; Cook et al., 2005; EFSA, 2006; Thomas et
al., 2013). The trade of infected livestock, as well as contaminated animal feed between countries
has been shown to contribute to the introduction of Salmonella into previously uninfected herds
(Blood & Radostits, 1989; Crump et al., 2002). Coordinated efforts to dramatically reduce and/or
13
to eradicate Salmonella altogether from certain livestock species have been remarkably
successful in a number of European countries, most notably, Denmark, Sweden, Finland, and
Norway (Wegener et al., 2003; EFSA, 2006). Other routes/sources of infection for humans are
via travel and exposure to pets, particularly exotic species such as reptiles and amphibians
(Hernandez et al., 2013; Anderson et al., 2016; Besser, 2018; Hale et al., 2012). With the rising
popularity of raw feeding practices for companion animals, pets are increasingly being examined
as a potential source of zoonotic Salmonella infections in humans (Leonard et al., 2015; Thomas
& Feng, 2020); a recent study linked exposure to raw diets with several cases of severe/fatal
clinical salmonellosis in domestic cats and dogs based on genomic data (Jones et al., 2019). The
ability of Salmonella to persist in the environment, and to withstand harsh environmental
conditions is an important factor in the persistence and transmission of Salmonella between
different animal, human and environmental sources; Salmonella has been shown to persist in soil
for at least a year (Davies & Wray, 1996).
Since the 1990s, the prevalence of AMR among non-typhoidal Salmonella has been
increasing worldwide (Chen et al., 2013). Resistant Salmonella infections result in more frequent
hospitalizations, prolonged illness, and carry a higher risk of complications and treatment failure
(Mąka & Popowska, 2016). Widespread use of antimicrobials in livestock at subtherapeutic or
therapeutic doses is believed to be a major driving force in the emergence and persistence of
multi-drug resistant Salmonella organisms (Mąka & Popowska, 2016; Forshell & Wierup, 2006).
In poultry, antimicrobials used at subtherapeutic levels for growth promotion have been shown to
disrupt the gut flora and may in fact increase the susceptibility of birds to subsequent infection
with Salmonella (Smith & Tucker, 1978). Some serovars, such as Salmonella Enteritidis, rarely
display AMR and their widespread emergence globally is likely related to virulence properties
14
(Michael & Schwarz, 2006). However, the rapid emergence of an epidemic strain of Salmonella
Typhimurium (phage type DT 104) that was detected worldwide in various sources (cattle, pigs,
poultry, sheep, humans), was likely related to its ability to resist five different families/classes of
antimicrobials, namely, penicillins, chloramphenicol, aminoglycosides, sulfonamides, and
tetracyclines (Michael & Schwarz, 2006; Helms et al., 2005). Concerningly, certain strains of
Salmonella Typhimurium DT 104 with penta-resistance also displayed resistance to
ciprofloxacin (Threlfall, 2002). Fortunately, recent work demonstrates that the prevalence of this
particular multi-drug resistant strain of Salmonella has decreased (Leekitcharoenphon et al.,
2016). Increasing resistance among Salmonella to carbapenems, (fluoro)quinolones, and third-
generation cephalosporins has also recently been observed (Chen et al., 2013). Resistance to
some of these antimicrobials among non-typhoidal Salmonella has been shown to be mediated
by plasmid-mediated transfer of the following genes: blaCMY-2, qnr, blaTEM, blaCTX-M (Mąka &
Popowska, 2016; Chen. et al., 2013).
1.2.2 Mechanisms and transmission of antimicrobial resistance
The discovery of antimicrobials in the late 1920s revolutionized medicine and healthcare,
providing a valuable and necessary treatment for infections caused by bacteria, fungi, and viruses
(Aminov, 2010). In recent decades, the overuse of antimicrobials in human medicine and
livestock production, as well as the contamination of the environment with antimicrobial by-
products, are thought to have contributed to the current major global predicament of the "post-
antibiotic era", where few effective antimicrobial options are available (McDermott et al., 2002).
The presence of AMR in foodborne infections has direct negative implications for treatment
success, and can result in prolonged hospitalization, and increased risk of complications among
15
hospitalized patients (Tanwar et al., 2014). Although AMR is influenced by anthropogenic use of
these substances, natural AMR has been documented in bacteria found in remote locations
untouched by humans, demonstrating that bacteria and other microorganisms possess these
adaptive survival mechanisms regardless of human influence (Van Goethem et al., 2018).
Two broad types of AMR exist: intrinsic and acquired. Intrinsic resistance is defined as
the ability of an organism to resist damage or destruction by an antimicrobial due to functional or
structural attributes of that organism (Arzanlou et al., 2017). For example, Pseudomonas spp. are
naturally/intrinsically resistant to penicillin and other antimicrobials as a result of their
impermeable outer membrane (Suginaka et al., 1975). In contrast, acquired resistance is, as the
term suggests, a mechanism that is acquired, and is therefore not considered a standard attribute
for all members of that species (Arzanlou et al., 2017). Since acquired resistance is the major
issue driving the post-antibiotic era, this type of resistance will be the focus of the following
discussion.
There are four main mechanisms of action of antimicrobials: inhibition of cell wall
synthesis, damage to cell membrane function, inhibition of nucleic acid synthesis/function, and
inhibition of protein synthesis (McManus, 1997). To counter the action of antimicrobials,
bacteria commonly use the following mechanisms: reduced penetration of antimicrobial agent
into the bacterial cell, efflux pumps, modification/degradation of the antimicrobial agent, or
antimicrobial target modification (McManus, 1997). Some of these resistance mechanisms
confer resistance to multiple antimicrobials at once (e.g., efflux pumps), and there are usually
several co-occurring mechanisms that are responsible for AMR observed in an organism
(Schindler & Kaatz, 2016).
16
Acquired resistance to antimicrobials occurs in two major ways, via mutations in existing
genes, or via horizontal gene transfer (i.e., the transfer of new genes from other bacteria). There
are three main mechanisms by which AMR genes can move between different bacteria:
transformation (uptake of naked DNA from environment), transduction (transfer by
bacteriophages), and conjugation (transfer by plasmids or other conjugative elements; Boerlin &
Reid-Smith, 2008). Antimicrobial resistance genes can move within a bacterial cell, as well as
between chromosomes, plasmids, and bacteriophages (Boerlin & Reid-Smith, 2008). The major
genetic elements responsible for moving AMR genes are transposons, insertion sequences,
ISCRs (atypical insertion sequence elements with common regions), plasmids, integrons,
bacteriophages, genomic islands, and ICEs (integrative conjugative elements; Boerlin & Reid-
Smith, 2008). Intracellular movement of AMR genes is facilitated mainly by integrons and
transposons, but insertion sequences and ISCRs also play a role (Toleman & Walsh, 2011).
Insertion sequences and transposons, also known as "jumping genes", can mobilize AMR genes
directly as well as indirectly, due to their ability to cause insertions, deletions, and translocations
of DNA (Bennett, 2008). Insertion sequences are typically very short DNA sequences that code
only for transposition proteins that allow them to mobilize themselves (Toleman & Walsh,
2011). Transposons may contain accessory genes in addition to transposition protein-encoding
genes, which allows for other genetic elements and/or AMR genes to be transposed as well
(Bennett, 2008). Composite transposons which consist of two separate transposons, can move as
one unit, thereby mobilizing the DNA between them (Bennett, 2008). Integrons are an ancient
genetic element present in bacteria that confer a selective advantage since they store large
numbers of AMR genes, and can re-shuffle them when the cell is under stress (Gillings, 2014).
Integrons contain gene cassettes, and possess three main features: intI, a gene encoding an
17
integrase which incorporates new gene cassettes, attI a recombination site for new gene
cassettes, and Pc, a promoter (Gillings, 2014). There are several advantages conferred by
integrons: new genes can be acquired without disturbing existing gene cassettes, and are
expressed by the promoter, and thus immediately subjected to selection pressure (Gillings,
2014). In addition, cassettes can also be excised intact to form circular DNA elements, without
disrupting the remainder of the integron (Gillings, 2014). Although integrons are not themselves
mobile, they can be captured by other mobile genetic elements (Gillings, 2014). Integrative
conjugative elements (ICEs) are a diverse group of genetic elements that have a common ability
to excise from and integrate into chromosomes, and they also possess the ability to transfer
between bacterial cells in a fashion similar to plasmids (Toleman & Walsh, 2011). ISCRs contain
insertion sequence elements and conserved recombinase sequences, and transpose using rolling
circle replication–a process which frequently captures extraneous DNA and accessory genes
since termination frequently occurs beyond the termination site (Toleman & Walsh, 2011).
Found ubiquitously, bacteriophages are considered one of the most abundant
microorganisms in the environment (Colavecchio et al., 2017). These viruses can accidentally
capture bacterial DNA during replication and can transfer this DNA to another bacterial cell
during subsequent transduction events. Phages are considered a reservoir of AMR genes; the
extent to which bacteriophages contribute to the movement of AMR genes is unknown, although
there is evidence to suggest they contribute significantly to horizontal gene transfer in foodborne
pathogens in the Enterobacterales order (Colavecchio et al., 2017; Hassan et al., 2021).
Plasmids are the most widely recognized, and well-known genetic element responsible
for the transmission of AMR genes. Defined as a double-stranded molecular DNA that is distinct
18
from chromosomal DNA, plasmids have the ability to replicate autonomously (Rozwandowicz et
al., 2018). Due to their ability to replicate in a wide variety of bacterial hosts, and to readily
acquire and transfer AMR genes via genetic elements such as transposons and insertion
sequences, plasmids are considered a major driver in the worldwide transmission of AMR
determinants (Rozwandowicz et al., 2018). Some plasmids (conjugative plasmids) are able to
initiate transfer to another bacterial cell via conjugation, while others (mobilizable plasmids) rely
on conjugative plasmids in order to transfer to another cell (Smillie et al., 2010). Mobilizable
plasmids contain only some of the genes required for conjugation but are able to parasitize
conjugative plasmids to permit high rates of transfer to other bacterial cells (Smillie et al., 2010).
Plasmids are typed according to their incompatibility with one another, or Inc types, named this
way since the cell cannot maintain two plasmids of the same incompatibility group at the same
time since the same replication machinery is used; therefore, one or both types will be lost from
the cell, or will not successfully be integrated into the cell in the first place (Rowandowicz et al.,
2018). Not all Inc types are incompatible with one another, thus a single bacterial cell may
contain multiple plasmids Inc types so long as they are compatible with one another
(Rowandowicz et al., 2018). Two parent plasmids may also fuse to create a hybrid plasmid, a
phenomenon which complicates the study of long-term evolution of AMR and plasmids
(Shintani et al., 2015). This fusion event may occur due to the presence of genetic elements such
as transposons, and insertion sequences. The fusion of two plasmids can result in inversions and
rapid changes in the position of accessory genes, including AMR genes (Shintani et al., 2015).
Complex interactions between pathogenic bacteria and commensal organisms such as
generic E. coli may facilitate the transfer of resistance genes from commensal organisms to
pathogens and vice versa, via horizontal gene transfer (Skurnik et al., 2016). This process is
19
thought to occur frequently in the gastrointestinal tract of humans and animals, thus the study of
AMR in non-pathogenic E. coli isolated from the feces of humans and animals can provide
valuable insights about AMR, as these commensal organisms may act both as sentinels and as
reservoirs of AMR genes (Van Den Bogaard & Stobberingh, 2000).
The study of bacterial population structures can offer important insights as to the primary
mode of transmission of AMR determinants. Bacterial populations diversify and evolve mainly
through de novo mutations and recombination events following horizontal gene transfer, and
distinct population structures emerge depending on the relative proportion of each that is
occurring (Spratt & Maiden, 1999). On one end of the spectrum, if little to no recombination
occurs, the population is described as being "clonal" (Spratt & Maiden, 1999). Under this clonal
model, new genetic material is acquired only by random mutation, and is passed vertically from
parent cells to daughter cells through asexual reproduction (i.e., binary fission; Spratt & Maiden,
1999). In a clonal population, the relative proportion of cells with a mutation conferring a fitness
advantage will increase over time due to selection pressures. If recombination plays a major role
in generating genetic changes in a population, the structure of the population is described as
"non-clonal" (Spratt & Maiden, 1999). "Panmictic" populations represent populations where a
high frequency of recombination events randomize alleles in the population such that stable
clones are not consistently maintained (Spratt & Maiden, 1999). In non-clonal populations,
genetic diversity is generated through horizontal gene transfer. The majority of bacterial
population structures likely fall somewhere in between these two extremes and may be regarded
as being weakly-clonal (Spratt & Maiden, 1999). In weakly-clonal populations, recombination
plays an important role in the generation of genetic diversity, but it does not prevent the
emergence of stable clones (Spratt & Maiden, 1999). The complexity of population structures
20
should not be underestimated: populations can also shift between different structures (i.e.,
weakly-clonal to non-clonal, panmictic to apparently clonal) depending on environmental
conditions (Spratt & Maiden, 1999). The population structure of Salmonella enterica subsp.
enterica was previously believed to be strictly clonal of the basis of multi-locus enzyme
electrophoresis typing data, however, it has since been determined that Salmonella exhibits a
high frequency of recombination within this subspecies, but not between different Salmonella
subspecies (Lan et al., 2009; Didelot et al., 2011; Octavia et al., 2006). The presence of high
levels of recombination can distort signals about true population structure and the primary mode
of transmission of AMR genes (i.e., horizontal gene transfer vs. clonal expansion), thus the
potential confounding effect of recombination should be accounted for in phylogenetic analyses,
especially for organisms displaying high rates of recombination (Croucher et al., 2013).
1.2.3 Molecular epidemiology
Molecular epidemiological approaches were first applied within the context of veterinary
medicine for the control of livestock diseases and zoonoses in the 1970s (Gebreyes et al., 2020).
Molecular epidemiology represents an extension of classic epidemiology rather than a field in
and of itself; it combines classic epidemiology with data obtained from molecular biology to
achieve a detailed understanding of the distribution, determinants, and prevention of infectious
diseases (Duggal et al., 2019). Molecular data that provide a finer resolution for typing of
bacteria beyond the simple positive identification of a pathogen may be either phenotypic or
genotypic in nature (Maslow et al., 1993). The former refers to strain characteristics that are
typically determined via benchtop experiments that identify externally expressed characteristics
of the microorganism. In contrast, genotypic data are determined through the direct analysis of
21
genetic data (Maslow et al., 1993). Molecular epidemiology can address some of the limitations
of classical epidemiology; by utilizing strain-level data, it offers the higher resolution approach
needed for source attribution studies, outbreak investigations, as well as surveillance of
pathogens and AMR (Gebreyes et al., 2020). Compared with phenotypic methods, genotyping,
or "DNA fingerprinting", offers higher discriminatory power, higher throughput, and–in theory–
higher reproducibility between different laboratories (Gebreyes et al., 2020). Molecular typing of
plasmids and bacterial strains together can offer valuable insights about whether the transmission
of AMR genes is driven predominantly by epidemic plasmids in different hosts, or the clonal
spread of organisms containing these plasmids (Rozwandowicz et al., 2018). The use of
genotypic data in the absence of epidemiological methods should not be equated with molecular
epidemiology; genotypic data are commonly used in other fields such as evolutionary biology
(Gebreyes et al., 2020). An investigation must fulfill the classical epidemiological approach by
identifying the frequency, distribution, and determinants of disease at the population level in
order to be considered under the umbrella of 'molecular epidemiology' (Gebreyes et al., 2020). A
diversity of approaches is available for the generation of phenotypic and genotypic data for the
purposes of bacterial strain typing; the following discussion will be limited only to the most
commonly used approaches.
Phenotypic typing methods
Biotyping represents a phenotypic method that characterizes bacteria on the basis of a number of
attributes, including colony morphology, chemical susceptibility and resistance, environmental
tolerances, and biochemical reactions (Eberle & Kiess, 2012). Biotyping may be combined with
other typing methods (e.g., serotyping, phagetyping) to provide higher discriminatory power
22
(Barker & Old, 1989). With the recent advances of newer genotypic methods and whole-genome
sequencing data, biotyping is less frequently used. Other phenotypic methods that rely on
immunological reactions (e.g., serotyping) or susceptibility to bacteriophages continue to play an
important role in Salmonella source attribution studies (Ferrari et al., 2017). Serotyping permits
the classification of strains on the basis of antisera reactions to surface antigens (flagellar-H,
somatic-O; Ferrari et al., 2017). In fact, serotyping of enteric bacteria such as Salmonella and E.
coli continues to be the standard for classification and nomenclature of these organisms in spite
of recent advances with available genomic data. Unfortunately, serotyping is a time-consuming
process that requires expensive reagents, and highly trained laboratory staff, thus, in silico
serotyping methods based on genomic data are increasingly being developed and adopted
(Yoshida et al., 2016; Joensen et al., 2015).
Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-
TOF MS) has recently become the gold standard method for microbial identification (Schubert &
Kostrzewa, 2017). This method is rapid, cost-effective, and can be used directly on bacterial
cultures to confirm species and also to determine strain types (Schubert & Kostrzewa, 2017). In
the future, MALDI-TOF mass spectrometry will also likely play a role in AMR testing (Schubert
& Kostrzewa, 2017). Unfortunately, typing using MALDI-TOF MS may be restricted to certain
laboratories due to the prohibitive cost of the instrument itself (>$100,000; Vranakis et al.,
2012).
Genotypic molecular methods
Older genotypic methods such as ribotyping have been widely applied to the study of foodborne
pathogens but have recently fallen out of favour since manual ribotyping is both time- and labour
23
intensive (Schumann & Pukall, 2013). Restriction fragment length polymorphism (RFLP) is a
technique that exploits variation in homologous DNA sequences, and utilizes restriction enzymes
to cleave a target gene, thereby producing sequences of varying lengths (Grover & Sharma,
2016). RFLP has similarly become obsolete with the introduction of faster, more efficient
technologies; RFLP can take several weeks to produce results (Grover & Sharma, 2016).
Although RFLP is no longer used today, the founding principles gleaned from this technique
have provided valuable insights for modern-day genomics (Chambers et al., 2014).
Until recently, pulsed-field gel electrophoresis (PFGE) was considered the gold-standard
for molecular typing of bacterial foodborne pathogens. PFGE utilizes alternating electric currents
to separate large DNA fragments in an agarose gel, following digestion by restriction enzymes
(Neoh et al., 2019). PFGE is reproducible between laboratories and provides considerable
discriminatory power–although in some cases it may not provide adequate discrimination
(Champion et al., 2002). Unfortunately, PFGE is time-intensive and requires the use of
specialized equipment (Goering, 2010). Furthermore, PFGE results may lack discriminatory
power, and may be prone to confounding (i.e., by recombination events) if epidemiologically
related isolates have undergone genetic recombination (Wassenaar et al., 1998).
DNA sequence-based methods such as MLVA (multi-locus variable number of tandem
repeats) and MLST (multi-locus sequence typing) have also been widely used in molecular
studies. MLST is an historical method used to type bacterial strains based on genetic sequences
of internal fragments of housekeeping genes (usually 7). MLST is a cost-effective and rapid
method that provides several advantages; the results are reproducible, easily interpretable, easily
shared, and universal (Urwin & Maiden, 2003). Unfortunately, in certain cases (e.g., a bottleneck
24
event), MLST may not provide adequate discriminatory power (Urwin & Maiden, 2003). MLST
has been used extensively for molecular epidemiological investigations of foodborne diseases
and will likely continue to play an important role in the years to come; in spite of advances in the
acquisition of vast amounts of genomic data for more detailed analyses, in silico MLST will
likely be a valuable tool in allowing researchers to place new findings into an historical context
(Pérez-Losada et al., 2013).
Genomics
With the use of data generated from next-generation sequencing (NGS) technologies, genomic
epidemiology brings together the fields of epidemiology, genetics, and statistics. Advances in
computation and throughput, along with the decreased costs of sequencing, have placed NGS at
the forefront of modern epidemiological investigations, and this technology promises to
revolutionize our ability to detect and respond to outbreaks, conduct surveillance, and perform
large-scale, long-term source attribution studies (Croucher et al., 2013; Duggal et al., 2019).
Attempts to quantify the economic benefit of whole-genome sequencing (WGS) data have
suggested large monetary benefits for healthcare, as well as federal costs; one estimate from a
Canadian research group estimates that sequencing will provide a net economic benefit
somewhere between 5 and 90 million dollars every year as a result of improved detection and
mitigation of Salmonella outbreaks (Jain et al., 2019). With the use of a single technique, the
data generated from sequencing technology allows researchers to identify a wide array of genetic
attributes including strain identification, and the detection of AMR genes, plasmids, and
virulence factors, among others (Shen et al., 2019).
25
However, with genomic epidemiology still in its infancy, there are many challenges to
overcome with regards to comparability of data from different sequencing platforms and
bioinformatics pipelines, as well as separating the "noise" from the "signal" in this type of data
(Croucher et al., 2013). One particular challenge lies with the heterogeneity of different methods
used for strain-typing of microorganisms (Schürch et al., 2018). In an ideal world, WGS analyses
would provide closed genomes available for highly discriminatory, robust analyses based on
differences in single-nucleotide polymorphisms (SNPs; Croucher et al., 2013). Unfortunately,
resolving the location and identity of highly repetitive sequences based on short-read sequence
data represents a major challenge, thus the majority of analyzed genomic data are in the form of
draft genomes. In addition, "noise" from sequencing and processing errors, as well as
spontaneous rearrangements in homologous and non-homologous genes (i.e., recombination) can
distort genetic relationships, and the latter may act as a potential confounder in epidemiologic
investigations, if unaccounted for (Croucher et al., 2013). Whole genome sequencing data,
however, are continuous in nature; depending on the needs of the analysis or investigation,
different types of information can be extracted. For example, in the case of an outbreak, high
specificity, and high discriminatory power is most needed, therefore SNP-based approaches are
considered most useful (Besser, 2018). In contrast, for long-term, large-scale investigations of
distantly related isolates that may display substantial heterogeneity with regard to source and
strain-type, gene-by-gene approaches (e.g., whole genome/core-genome MLST) may be most
appropriate, to minimize the "noise" generated from SNP-based approaches (Croucher et al.,
2013). Statistical methods have recently been developed to account for recombination in SNP-
based analyses; since a single recombination event can introduce several polymorphisms at once,
it can be difficult to determine which substitutions have arisen from mutations vs. recombination
26
events (Didelot & Wilson, 2015). Although these statistical methods are able to adjust for the
effect of recombination in SNP-based phylogenetic analyses (Didelot & Wilson, 2015), SNP-
based phylogenies should generally be interpreted as a useful framework, rather than providing
evidence of definitive population structure and true epidemiological associations (Croucher et
al., 2013). Long-term, large-scale epidemiological analyses of genomic data from enteric
bacteria frequently employ gene-by-gene approaches (e.g., core genome MLST; cgMLST);
although these methods have lower discriminatory power than SNP-based approaches, they offer
valuable comparability for different epidemiological investigations across different regions and
time periods (Schürch et al., 2018). Gene-by-gene approaches may suffer from lower
discriminatory power as compared with SNP-based approaches, however, they likely provide
adequate discrimination for distantly related isolates, and are widely used for public health
surveillance (Nadon et al., 2017). For gene-by-gene approaches, recombination is accounted for
through the assignment of alleles, with smaller numbers of allelic changes for regions of
recombination versus mutations (Schürch et al., 2018).
To date, the analysis of genomic data for epidemiological investigations has been hyper-
focused on the core genome, defined as the set of genes that is common to a group of isolates
(Schürch et al., 2018). The accessory genome, or the set of genes that is external to the core
genome and also contained within the pan-genome–the set of all genes from a group of isolates–
likely represents a crucial link to our understanding of the evolution and structure of bacterial
populations, as well as the movement of genetic material between bacteria (Schürch et al., 2018).
It is anticipated that combined analyses with accessory, core, and pangenomes will play a major
role in future investigations in this field (Schürch et al., 2018).
27
1.2.4 The role of wildlife in the epidemiology of foodborne pathogens and associated
antimicrobial resistance
Due to challenges in identifying sources of outbreaks and causes of the increasing prevalence of
AMR globally, researchers are increasingly looking to ecosystems and wildlife as biological and
mechanical vectors of foodborne pathogens, multi-drug resistance bacteria, and AMR
determinants (Bonnedahl & Jährhult, 2014). Habitat encroachment, urbanization and increasing
interactions with wild animals are thought to contribute to the transmission of problematic
microorganisms between wildlife, humans, domestic animals, and their environments (Lejeune
& Pearl, 2014). In spite of the fact that wild animals are not typically directly exposed to
antimicrobials and pathogenic bacteria, there are many postulated routes of indirect exposure for
wild animals through the pollution and contamination of a shared environment. Anthropogenic
sources of bacteria and AMR by-products may enter into waterways that are used by wildlife for
foraging and drinking. Sewage, wastewater effluent from hospitals, and rainwater run-off from
crop fields fertilized with contaminated manure have proven to be concentrated sources of AMR
bacteria, and foodborne pathogens (Karkman et al., 2018; Cahill et al., 2019; Zhu et al., 2019;
Bicudo & Goyal, 2003). Wildlife may also be exposed to these microorganisms and resistant
bacteria on land, when foraging on farms, in cities, and in landfills (Dolejska & Papagiannitsis,
2018; Ahlstrom et al., 2018). Exposure of wildlife to heavy metals, and biocides in the
environment is also thought to promote the selection and maintenance of multi-drug resistance
(Wales & Davies, 2015; Romero et al., 2016; Van Breda et al., 2017).
Exploratory work (primarily in the form of cross-sectional surveys) has demonstrated the
presence of many different pathogenic bacteria (e.g., Salmonella, Campylobacter, E. coli 0157)
28
in the feces of a wide variety of wild avian and mammalian species (Chlebicz & Slizewska,
2018). Furthermore, bacterial clones of international importance (e.g., E. coli ST131), ESBLs,
and resistance to last-resort antimicrobials (e.g., colistin, carbapenems, vancomycin) mediated by
mobile resistance genes have all been isolated from and documented in wildlife (Tausova et al.,
2012; Jamborova et al., 2018; Veldman et al., 2013; Poirel et al., 2012; Oracová et al., 2018;
Báez et al., 2015; Fischer et al., 2013; Liakopoulos et al., 2016). Moreover, these wild animals
are generally considered to be healthy, showing no overt clinical signs of illness or infection.
Although much information has been gained over the past several decades with respect to the
role of wildlife in the increasing incidence of foodborne illness and antimicrobial resistant
infections, several research gaps still exist (Wang et al., 2017). It is unknown whether wildlife
are merely sentinels of environmental contamination/pollution, or whether they contribute as
sources and/or reservoirs of pathogenic/resistant bacteria (Wang et al., 2017; Bonnedahl &
Jährhult, 2014). There are concerns that the widespread movement and migration of wild animals
may contribute to the global dissemination of antimicrobial resistant bacteria (Bonnedahl &
Jährhult, 2014). Several studies have shown that wildlife living and foraging in close proximity
to anthropogenic sources are more likely to harbour bacteria with AMR (Cole et al., 2005, Allen
et al., 2011) and animals living in remote locations appear to be less contaminated (Rolland et
al., 1985; Thallet et al., 2010; Wheeler et al., 2012). However, wildlife living in these remote
locations have also tested positive for AMR bacteria (Van Breda et al., 2017; Santos et al.,
2013). The dynamics between environmental contamination and the carriage of bacteria and
AMR by wildlife appear to be quite complex and may be complicated by considerable
heterogeneity at the level of animal species and demographics, geographic location, season of
study, and study methods used (Benskin et al., 2009; Vogt et al., 2020).
29
1.3 Study rationale and objectives
The increasing adoption of One Health approaches, which recognize that human, animal and
environmental health are inextricably linked, are likely to shed important insights concerning the
epidemiology of foodborne illness and AMR. As complex phenomena, these issues require a
perspective that considers and weighs all components of the ecosystem–including wild animals–
in the relevant problem (McEwen & Collignon, 2018). Over the past several decades, much of
the research on these issues of AMR and foodborne illness has been focused on wild bird
species, since birds are capable of widespread movement through long distance migrations, and
there are obvious concerns about the introduction of AMR determinants into new locations
(Wang et al., 2017). There is comparatively little work on medium-sized terrestrial species such
as raccoons, a species which arguably merits further study due to their unique biology and
behaviour. Raccoons are easily recognizable by their physical attributes and behaviours, with a
distinct facial mask, and ringed tail, as well as their dexterous paws, and opportunistic foraging
behaviour. Raccoons represent a species of wildlife that is well-adapted to a vast number of
ecological niches in both urban and rural landscapes (French et al., 2019). In urban
environments, raccoons are known to interact with domestic animals and humans, as they forage
trash in urban environments and compete for territory; raccoons can reach remarkable population
densities in cities (up to 94/km2 has been recorded; Broadfoot et al., 2001). Not a social species
in the traditional sense, raccoons are understood to have a fusion-fission social structure, where
short-term social interactions are common, but long-term associations are rare (Prange et al.,
2011). As opportunistic foragers, raccoons may also occupy vast territories in rural landscapes if
they need to travel to find food; ranges up to 50km2 have been recorded in the Canadian prairies
30
(Rosatte, 2008). Raccoons are a remarkably adaptable species that have been introduced into and
gained a foothold in countries outside of their native range, such as Germany and Japan, where
they are now considered an invasive, pest species that–in some cases–can be hunted year-round
to control their expanding populations (Fischer et al., 2015; Osaki et al., 2019). Previous work in
raccoons has confirmed that they carry bacteria considered pathogenic to humans including:
Clostridium difficile, Yersinia, Salmonella, Campylobacter, Leptospirosis, and Tularemia
(Bondo et al., 2015; Lee et al., 2011; Bondo et al., 2016a; Duncan et al., 2012). Comparatively
little research is available concerning AMR in raccoons (Bondo et al. 2016a; Bondo et al.,
2016b; Baldi et al., 2019).
A previous repeated cross-sectional study examining a population of raccoons in southern
Ontario over a three-year period in both urban and rural locations will form the basis of the work
presented in this thesis (Bondo, 2015). In this thesis, we will use a combination of statistical
modeling and WGS data for bacterial isolates obtained from this previous study in order to
address current gaps in our understanding of the epidemiology of Salmonella, Campylobacter,
generic E. coli, and associated AMR in raccoons. In addition, we will employ a One Health
approach by using, where possible, a variety of sampling sources in addition to wildlife samples,
including human, livestock, and environmental samples. By employing a One Health approach
together with genomic data in our epidemiological investigations, we hope to contribute to a
deeper understanding of the epidemiology of foodborne illness and the movement and
transmission of AMR determinants between different components of the ecosystem. Overall, this
thesis represents an effort to better understand the role of raccoons in the epidemiology of
Campylobacter, Salmonella, E. coli and AMR in the context of public health. In order to address
these issues, this thesis has the following objectives:
31
1. Assess the impact of demographic, temporal, seasonal, and annual factors
on the occurrence of Campylobacter in raccoons on swine farms and
conservation areas (Chapter 2);
2. Assess for potential transmission of Salmonella and antimicrobial resistant
E. coli between raccoons, swine manure pits, and soil on swine farms
based on similarities between core-genome multi-locus sequence types
determined using whole-genome sequencing (WGS) data (Chapter 3);
3. Determine whether source types and farm locations predict the occurrence
of certain plasmids and AMR genes (determined in silico) in Salmonella
and resistant E. coli isolates from raccoons, swine manure pits, and soil
samples obtained on swine farms (Chapter 3);
4. Assess for potential transmission of antimicrobial resistant E. coli between
wildlife, livestock, and environmental sources at a broad scale, and at a
local scale (on swine farms and conservation areas) based on similarities
between core-genome multi-locus sequence types determined using WGS
data (Chapter 4);
5. Determine whether source types and location type predict the occurrence
of certain plasmids and AMR genes (determined in silico) in resistant E.
coli isolates from wildlife, livestock, and environmental sources (Chapter
4);
6. Assess for potential transmission of Salmonella between raccoons,
livestock, environmental samples, and human clinical cases based on
32
similarities between core-genome multi-locus sequence types determined
using WGS data (Chapter 5);
7. Determine whether source type predicts the occurrence of certain plasmids
and AMR genes (determined in silico) in Salmonella isolates from
humans, raccoons, livestock, and environmental sources (Chapter 5).
33
1.4 References
Abebe, E., Gugsa, G., & Ahmed, M. (2020). Review on major food-borne zoonotic bacterial pathogens. Journal of Tropical Medicine, 2020, 1–19.
Acheson, D., & Allos, M. (2001). Campylobacter jejuni infections: update on emerging issues and trends. Clinical Infectious Diseases, 32, 1201–1206.
Ahlstrom, C. A., Bonnedahl, J., Woksepp, H., Hernandez, J., Olsen, B., & Ramey, A. M. (2018). Acquisition and dissemination of cephalosporin-resistant E. coli in migratory birds sampled at an Alaska landfill as inferred through genomic analysis. Scientific reports, 8(1), 7361.
Alegbeleye, O. O., Singleton, I., & Sant’Ana, A. S. (2018). Sources and contamination routes of microbial pathogens to fresh produce during field cultivation: A review. Food Microbiology, 73, 177–208.
Allen, S. E., Boerlin, P., Janecko, H., Lumsden, J. S., Barker, I. K., Pearl, D. L., Reid-Smith, R., & Jardine, C. (2011). Antimicrobial resistance in generic Escherichia coli isolates from wild small mammals living in swine farm, residential, landfill, and natural environments in southern Ontario, Canada. Applied and Environmental Microbiology, 77, 882-888.
Aminov R. I. (2010). A brief history of the antibiotic era: Lessons learned and challenges for the future. Frontiers in Microbiology, 1, 134.
Anderson, T. C., Marsden-Haug, N., Morris, J. F., Culpepper, W., Bessette, N., Adams, J. K., Bidol, S., Meyer, S., Schmitz, J., Erdman, M. M., Gomez, T. M., & Barton Behravesh, C. (2017). Multistate outbreak of human Salmonella Typhimurium infections linked to pet hedgehogs - United States, 2011-2013. Zoonoses and Public Health, 64(4), 290–298.
Anon. (1999). Salmonella, strategidokument [in Swedish]. Socialstyrelsen, Stockholm.
Arrausi-Subiza, M., Gerrikagoitia, X., Alvarez, V., Ibabe, J. C., & Barral, M. (2016). Prevalence of Yersinia enterocolitica and Yersinia pseudotuberculosis in wild boars in the Basque Country, northern Spain. Acta Veterinaria Scandinavica, 58, 4.
Arzanlou, M., Chai, W. C., & Venter, H. (2017). Intrinsic, adaptive and acquired antimicrobial resistance in Gram-negative bacteria. Essays in Biochemistry, 61(1), 49–59.
Báez, J., Hernández-García, M., Guamparito, C., Díaz, S., Olave, A., Guerrero, K., Cantón, R., Baquero, F., Gahona, J., Valenzuela, N., Del Campo, R., & Silva, J. (2015). Molecular characterization and genetic diversity of ESBL-producing Escherichia coli colonizing the migratory Franklin's gulls (Leucophaeus pipixcan) in Antofagasta, North of Chile. Microbial drug resistance (Larchmont, N.Y.), 21(1), 111–116.
34
Bailey, J. S., Stern, N. J., Fedorka-Cray, P., Craven, S. E., Cox, N. A., & Cosby, D. E. (2001). Sources and movement of Salmonella through integrated poultry operations: A multistate epidemiological investigation. Journal of Food Protection, 64, 1690–1697.
Balali, G. I., Yar, D. D., Afua Dela, V. G., & Adjei-Kusi, P. (2020). Microbial contamination, an increasing threat to the consumption of fresh fruits and vegetables in today’s world. International Journal of Microbiology, 2020, Article ID: 3029295.
Baldi, M., Barquero Calvo, E., Hutter, S. E., & Walzer, C. (2019). Salmonellosis detection and evidence of antibiotic resistance in an urban raccoon population in a highly populated area, Costa Rica. Zoonoses and Public Health, 66(7), 852–860.
Barker, R. M., & Old, D.C. (1989). The usefulness of biotyping in studying the epidemiology and phylogeny of Salmonella. Journal of Medicinal Microbiology, 29, 81-88.
Benskin, C. M., Wilson, K., Jones, K., & Hartley, I. R. (2009). Bacterial pathogens in wild birds: A review of the frequency and effects of infection. Biological Reviews of the Cambridge Philosophical Society, 84(3), 349–373.
Bennett PM (2008). Plasmid encoded antibiotic resistance: Acquisition and transfer of antibiotic resistance genes in bacteria. British Journal of Pharmacology, 153(suppl. 1):S347–S357
Besser, J. M. (2018). Salmonella epidemiology: A whirlwind of change. Food Microbiology, 71, 55–59.
Bicudo, J. R., & Goyal, S. M. (2003). Pathogens and manure management systems: A review. Environmental Technology, 24(1), 115–130.
Blood D.C. & Radostits O.M. (1989). Diseases caused by bacteria, III. In Veterinary medicine: A textbook of the diseases of cattle, sheep, pigs, goats and horses (D.C. Blood, O.M. Radostits, J.H. Arundel & C.C. Gay, eds), 7th Ed. Ballière Tindall, London, 643-655.
Bondo, K.J. (2015). Epidemiology of Salmonella and antimicrobial resistant enteric bacteria in raccoons and the environment on swine farms and conservation areas in southern Ontario (Doctoral dissertation, University of Guelph; Guelph, Canada). Retrieved from https://atrium.lib.uoguelph.ca/xmlui/bitstream/handle/10214/9454/Bondo_Kristin_201601_Ph.D..pdf;sequence=3
Bondo, K. J., Weese, J. S., Rouseau, J., & Jardine, C. M. (2015). Longitudinal study of Clostridium difficile shedding in raccoons on swine farms and conservation areas in Ontario, Canada. BMC veterinary research, 11, 254.
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016a). Impact of season, demographic and environmental factors on Salmonella
35
occurrence in raccoons (Procyon lotor) from swine farms and conservation areas in southern Ontario. Plos One, 11(9), e0161497.
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016b). Epidemiology of antimicrobial resistance in Escherichia coli isolates from raccoons (Procyon lotor) and the environment on swine farms and conservation areas in Southern Ontario. Plos One, 11(11): e0165303.
Bonnedahl, J., & Järhult, J. D. (2014). Antibiotic resistance in wild birds. Upsala Journal of Medical Sciences, 119(2), 113–116.
Boerlin, P., Reid-Smith, R.J. (2008). Antimicrobial resistance: Its emergence and transmission. Animal Health Research Reviews, 9, 115-126.
Bojanic, K., Midwinter, A. C., Marshall, J. C., Rogers, L. E., Biggs, P. J., & Acke, E. (2017). Isolation of Campylobacter spp. from client-owned dogs and cats, and retail raw meat pet food in the Manawatu, New Zealand. Zoonoses and Public Health, 64(6), 438–449.
Broadfoot, J. D., Rosatte, R. C., & O'Leary, D. T. (2001). Raccoon and skunk population models for urban disease control planning in Ontario, Canada. Ecological Applications, 11, 295- 303.
Butzler, JP. (2004). Campylobacter, from obscurity to celebrity. Clinical Microbiology and Infection, 10, 868–876.
Cahill, N., O'Connor, L., Mahon, B., Varley, Á., McGrath, E., Ryan, P., Cormican, M., Brehony, C., Jolley, K. A., Maiden, M. C., Brisse, S., & Morris, D. (2019). Hospital effluent: A reservoir for carbapenemase-producing Enterobacterales? The Science of the Total Environment, 672, 618–624.
Carstens, C. K., Salazar, J. K, & Darkoh, C. (2019). Multistate outbreaks of foodborne illness in the United States associated with fresh produce from 2010 to 2017. Frontiers in Microbiology, 10, 2667.
CDC. (2008). "Salmonella". Centers for Disease Control and Prevention. Available at: www.cdc.gov/salmonella/ (accessed on 12 April 2020).
CDC. (2020). "Foodborne germs and illnesses". Centers for Disease Control and Prevention. Available at: https://www.cdc.gov/foodsafety/foodborne-germs.html (accessed 18 August 2020).
Cernicchiaro, N., Pearl, D. L., McEwen, S. A., Harpster, L., Homan, H. J., Linz, G. M., & Lejeune, J. T. (2012). Association of wild bird density and farm management factors with the prevalence of E. coli O157 in dairy herds in Ohio (2007-2009). Zoonoses and Public Health, 59(5), 320–329.
36
Chambers, G. K., Curtis, C., Millar, C. D., Huynen, L., & Lambert, D. M. (2014). DNA fingerprinting in zoology: Past, present, future. Investigative Genetics, 5(1), 3.
Champion, O. L., Best, E. L., & Frost, J. A. (2002). Comparison of pulsed-field gel electrophoresis and amplified fragment length polymorphism techniques for investigating outbreaks of enteritis due to Campylobacters. Journal of Clinical Microbiology, 40, 2263–2265.
Chen, H. M., Wang, Y., Su, L. H., & Chiu, C. H. (2013). Nontyphoid Salmonella infection: microbiology, clinical features, and antimicrobial therapy. Pediatrics and Neonatology, 54(3), 147–152.
Cheng, R. A., Eade, C. R., & Wiedmann, M. (2019). Embracing diversity: Differences in virulence mechanisms, disease severity, and host adaptations contribute to the success of nontyphoidal Salmonella as a foodborne pathogen. Frontiers in Microbiology, 10, 1368.
Chlebicz, A., & Śliżewska, K. (2018, May 1). Campylobacteriosis, salmonellosis, yersiniosis, and listeriosis as zoonotic foodborne diseases: A review. International Journal of Environmental Research and Public Health, 15(5), 863.
Christensen, J., Baggesen, D. L., Nielsen, B., & Stryhn, H. (2002). Herd prevalence of Salmonella spp. in Danish pig herds after implementation of the Danish Salmonella control program with reference to a pre-implementation study. Veterinary Microbiology, 88(2), 175-188.
Colavecchio, A., Cadieux, B., Lo, A., & Goodridge, L. D. (2017). Bacteriophages contribute to the spread of antibiotic resistance genes among foodborne pathogens of the Enterobacteriaceae family - A review. Frontiers in Microbiology, 8, 1108.
Cole, D., Drum, D. J. V., Stallknecht, D. E., White, D. G., Lee, M. D., Ayers S., Sobsey M., & Maurer, J. J. (2005). Free-living Canada geese and antimicrobial resistance. Emerging Infectious Diseases, 11(6), 935-938.
Cook, A. J. C., Miller, A., Snow, L., & Davies, R. H. (2005). Epidemiological studies of Salmonella infection in pigs. In Proc. Med-Vet-Net 1st General Scientific Meeting, Winchester, United Kingdom, 29 June – 1 July. Available at: www.medvetnet.org/cms/
Croucher, N. J., Harris, S. R., Grad, Y. H., & Hanage, W. P. (2013). Bacterial genomes in epidemiology--present and future. Philosophical Transactions of the Royal Society of London. Series B, Biological sciences, 368(1614), 20120202.
Crump J.A., Griffin P.M. & Angulo F.J. (2002). Bacterial contamination of animal feed and its relationship to human foodborne illness. Clinical Infectious Diseases, 35(7), 859-865.
37
CSPI. (2006). Closing the gaps in our federal food-safety net, 8th edition. Center for Science in the Public Interest. Washington, DC. www.cspinet.org.
Dargatz, D. A., Fedorka-Cray, P. J., Ladely, S. R., Kopral, C. A., Ferris, K. E., and Headrick, M. L. (2003). Prevalence and antimicrobial susceptibility of Salmonella spp. isolates from US cattle in feedlots in 1999 and 2000. Journal of Applied Microbiology, 95, 753–761.
Davies, R. H., & Wray, C. (1996). Persistence of Salmonella Enteritidis in poultry units and poultry food. British Poultry Science, 37, 589-596.
Davies, J., & Davies, D. (2010). Origins and evolution of antibiotic resistance. Microbiology and Molecular Biology Reviews, 74, 417-433.
Desai, P. T., Porwollik, S., Long, F., Cheng, P., Wollam, A., Clifton, S. W., Weinstock, G. M., & McClelland, M. (2013). Evolutionary genomics of Salmonella Enterica subspecies. MBio, 4(2).
Didelot, X., Bowden, R., Street, T., Golubchik, T., Spencer, C., McVean, G., Sangal, V., Anjum, M. F., Achtman, M., Falush, D., & Donnelly, P. (2011). Recombination and population structure in Salmonella Enterica. Plos Genetics, 7(7), e1002191.
Didelot, X., & Wilson, D. J. (2015). ClonalFrameML: Efficient inference of recombination in whole bacterial genomes. Plos Computational Biology, 11(2), e1004041.
Doane, C. A., Pangloli, P., Richards, H. A., Mount, J. R., Golden, D. A., and Draughon, F. A. (2007). Occurrence of Escherichia coli O157:H7 in diverse farm environments. Journal of Food Protection, 70, 6–10.
Duncan, C., Krafsur, G., Podell, B., Baeten, L. A., Levan, I., Charles, B., & Ehrhart, E. J. (2012). Leptospirosis and tularaemia in raccoons (Procyon lotor) of Larimer country, Colorado. Zoonoses and Public Health, 59(1), 29–34.
Duggal, P., Ladd-Acosta, C., Ray, D., & Beaty, T. H. (2019). The evolving field of genetic epidemiology: From familial aggregation to genomic sequencing. American Journal of Epidemiology, 188(12), 2069–2077.
Eberle, K. N. & Kiess, A. S. (2012). Phenotypic and genotypic methods for typing Campylobacter jejuni and Campylobacter coli in poultry. Poultry Science, 91, 255–64.
European Food Safety Authority (EFSA). (2006). Trends and sources of zoonoses, zoonotic agents and antimicrobial resistance in the European Union in 2004. EFSA Journal, 310(10) 23-95.
38
European Food Safety Authority (EFSA). (2016). The European Union summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in 2015. EFSA Journal, 14, 231.
Facciola, A., Riso, R., Avventuroso, E., Visalli, G., Delia, S. A., & Laganà, P. (2017). Campylobacter: From microbiology to prevention. Journal of Preventive Medicine and Hygiene, 58(2), E79–E92.
Ferrari, R. G., Panzenhagen, P., & Conte-Junior, C. A. (2017). Phenotypic and genotypic eligible methods for Salmonella Typhimurium source tracking. Frontiers in Microbiology, 8, 2587.
Fischer, J., Schmoger, S., Jahn, S., Helmuth, R., & Guerra, B. (2013). NDM-1 carbapenemase-producing Salmonella enterica subsp. enterica serovar Corvallis isolated from a wild bird in Germany. The Journal of Antimicrobial Chemotherapy, 68(12), 2954–2956.
Fischer, M. L., Hochkirch, A., Heddergott, M., Schulze, C., Anheyer-Behmenburg, H. E., Lang, J., Michler, F. U., Hohmann, U., Ansorge, H., Hoffmann, L., Klein, R., & Frantz, A. C. (2015). Historical invasion records can be misleading: Genetic evidence for multiple introductions of invasive raccoons (Procyon lotor) in Germany. Plos One, 10(5), e0125441.
Forshell, L. P., & Wierup, M. (2006). Salmonella contamination: A significant challenge to the global marketing of animal food products. Revue scientifique et technique (International Office of Epizootics) 25, 541-554.
Fredriksson-Ahomaa M. (2019). Wild Boar: A reservoir of foodborne zoonoses. Foodborne Pathogens and Disease, 16(3), 153–165.
French, S. K., Pearl, D. L., Peregrine, A. S., & Jardine, C. M. (2019). Baylisascaris procyonis infection in raccoons: A review of demographic and environmental factors influencing parasite carriage. Veterinary Parasitology, Regional Studies and Reports, 16, 100275.
French, N. P., Midwinter, A., Holland, B., Collins-Emerson, J., Pattison, R., Colles, F., & Carter, P. (2009). Molecular epidemiology of Campylobacter jejuni isolates from wild-bird fecal material in children's playgrounds. Applied and Environmental Microbiology, 75(3), 779–783.
Furukawa, M., Goji, N., Janzen, T. W., Thomas, M. C., Ogunremi, D., Blais, B., Misawa, N., & Amoako, K. K. (2018). Rapid detection and serovar identification of common Salmonella enterica serovars in Canada using a new pyrosequencing assay. Canadian Journal of Microbiology, 64(1), 75–86.
Gebreyes, W. A., Jackwood, D., de Oliveira, C. J. B., Lee, C.-W., Hoet, A. E., & Thakur, S. (2020). Molecular epidemiology of infectious zoonotic and livestock diseases. Microbiology Spectrum, 8(2).
39
Gillings MR. (2014). Integrons: Past, present, and future. Microbiology and Molecular Biology Reviews, 78, 257-277.
Goering, R. V. (2010). Pulsed field gel electrophoresis: A review of application and interpretation in the molecular epidemiology of infectious disease. Infection, Genetics and Evolution, 10, 866–875.
Government of Canada. (2019) "Causes of food poisoning". Available from: https://www.inspection.gc.ca/food-safety-for-industry/information-for-consumers/fact-sheets-and-infographics/food-poisoning/eng/1331151916451/1331152055552 (accessed on 18 August 2020).
Greig, J., Rajić, A., Young, I., Mascarenhas, M., Waddell, L., & LeJeune, J. (2015). A scoping review of the role of wildlife in the transmission of bacterial pathogens and antimicrobial resistance to the food chain. Zoonoses and Public Health, 62(4), 269–284.
Grover, A., & Sharma, P. C. (2016). Development and use of molecular markers: Past and present. Critical Reviews in Biotechnology, 36(2), 290-302.
Hale, C. R., Scallan, E., Cronquist, A. B., Dunn, J., Smith, K., Robinson, T., Lathrop, S., Tobin-D'Angelo, M., & Clogher, P. (2012). Estimates of enteric illness attributable to contact with animals and their environments in the United States. Clinical Infectious Diseases, 54 Suppl 5, S472–S479.
Hald, B., Skov, M. N., Nielsen, E. M., Rahbek, C., Madsen, J. J., Wainø, M., Chriél, M., Nordentoft, S., Baggesen, D. L., & Madsen, M. (2016). Campylobacter jejuni and Campylobacter coli in wild birds on Danish livestock farms. Acta Veterinaria Scandinavica, 58, 11.
Hansson, I., Sandberg, M., Habib, I., Lowman, R., & Engvall, E. O. (2018). Knowledge gaps in control of Campylobacter for prevention of campylobacteriosis. Transboundary and Emerging Diseases, 65, 30–48.
Hassan, A. Y., Lin, J. T., Ricker, N., & Anany, H. (2021). The age of phage: Friend or foe in the new dawn of therapeutic and biocontrol applications?. Pharmaceuticals, 14, 199.
Hassell, J. M., Ward, M. J., Muloi, D., Bettridge, J. M., Robinson, T. P., Kariuki, S., Ogendo, A., Kiiru, J., Imboma, T., Kang'ethe, E. K., Öghren, E. M., Williams, N. J., Begon, M., Woolhouse, M., & Fèvre, E. M. (2019). Clinically relevant antimicrobial resistance at the wildlife-livestock-human interface in Nairobi: An epidemiological study. The Lancet Planetary Health, 3(6), e259–e269.
Helms M, Ethelberg S, Mølbak K, DT104 Study Group. (2005). International Salmonella Typhimurium DT104 infections, 1992-2001. Emerging Infectious Diseases, 11, 859e 67
40
Hernandez, S. M., Keel, K., Sanchez, S., Trees, E., Gerner-Smidt, P., Adams, J. K., Cheng, Y., Ray, A., 3rd, Martin, G., Presotto, A., Ruder, M. G., Brown, J., Blehert, D. S., Cottrell, W., & Maurer, J. J. (2012). Epidemiology of a Salmonella enterica subsp. enterica serovar Typhimurium strain associated with a songbird outbreak. Applied and Environmental Microbiology, 78(20), 7290–7298.
Herikstad H., Motarjemi Y. & Tauxe R.V. (2002). Salmonella surveillance: A global survey of public health serotyping. Epidemiology and Infection, 129(1), 1-8
Hussein, H. S., & Bollinger, L. (2005). Prevalence of Shiga toxin-producing Escherichia coli in beef cattle. Journal of Food Protection, 68, 2224–2241.
Institute of Medicine (US) and National Research Council (US) Committee to ensure safe food from production to consumption. Ensuring safe food: From production to consumption. Washington (DC): National Academies Press (US; 1998). "The changing nature of food hazards: Cause for increasing concern". Available from: https://www.ncbi.nlm.nih.gov/books/NBK209107/
Iwu, C. D., & Okoh, A. I. (2019). Preharvest transmission routes of fresh produce associated bacterial pathogens with outbreak potentials: A review. International Journal of Environmental Research and Public Health, 16(22), 4407.
Jain, S., Mukhopadhyay, K., & Thomassin, P. J. (2019). An economic analysis of Salmonella detection in fresh produce, poultry, and eggs using whole genome sequencing technology in Canada. Food Research International, 116, 802–809.
Jamborova, I., Janecko, N., Halova, D., Sedmik, J., Mezerova, K., Papousek, I., Kutilova, I., Dolejska, M., Cizek, A., & Literak, I. (2018). Molecular characterization of plasmid-mediated AmpC beta-lactamase- and extended-spectrum beta-lactamase-producing Escherichia coli and Klebsiella pneumoniae among corvids (Corvus brachyrhynchos and Corvus corax) roosting in Canada. FEMS Microbiology Ecology, 94(11).
Jemmi, T., & Stephan, R. (2006). Listeria monocytogenes: Food-borne pathogen and hygiene indicator. Revue scientifique et technique (International Office of Epizootics), 25(2), 571–580.
Joensen, K. G., Tetzschner, A. M., Iguchi, A., Aarestrup, F. M., & Scheutz, F. (2015). Rapid and easy in silico serotyping of Escherichia coli isolates by use of whole-genome sequencing data. Journal of Clinical Microbiology, 53(8), 2410–2426.
Jones, J. L., Wang, L., Ceric, O., Nemser, S. M., Rotstein, D. S., Jurkovic, D. A., Rosa, Y., Byrum, B., Cui, J., Zhang, Y., Brown, C. A., Burnum, A. L., Sanchez, S., & Reimschuessel, R. (2019). Whole genome sequencing confirms source of pathogens associated with bacterial foodborne illness in pets fed raw pet food. Journal of Veterinary Diagnostic Investigation, 31(2), 235–240.
41
Jung, Y., Jang, H., & Matthews, K.R. (2014). Effect of the food production chain from farm practices to vegetable processing on outbreak incidence. Microbial Biotechnology, 7, 517–527.
Kaakoush, N. O., Castaño-Rodríguez, N., Mitchell, H. M., & Man, S. M. (2015). Global epidemiology of Campylobacter infection. Clinical Microbiology Reviews, 28(3), 687–720.
Karkman, A., Do, T. T., Walsh, F., & Virta, M. (2018). Antibiotic-resistance genes in wastewater. Trends in Microbiology, 26(3), 220–228.
Keen, J. E., Wittum, T. E., Dunn, J. R., Bono, J. L., & Durso, L. M. (2006). Shiga-toxigenic Escherichia coli O157 in agricultural fair livestock, United States. Emerging Infectious Diseases, 12, 780–786.
Kendall, P., Scharff, R., Baker, S., LeJeune, J., Sofos, J., & Medeiros, L. (2017). Food safety instruction improves knowledge and behavior risk and protection factors for foodborne illnesses in pregnant populations. Maternal and Child Health Journal, 21(8), 1686–1698.
Krishnasamy, V. P., Marshall, K., Dewey-Mattia, D., & Wise, M. (2020). Outbreak characteristics and epidemic curves for multistate outbreaks of Salmonella infections associated with produce: United States, 2009-2015. Foodborne Pathogens and Disease, 17(1), 15–22.
Lamas, A., Regal, P., Vázquez, B., Miranda, J. M., Cepeda, A., & Franco, C. M. (2018). Salmonella and Campylobacter biofilm formation: A comparative assessment from farm to fork. Journal of the Science of Food and Agriculture, 98(11), 4014–4032.
Lan, R., Reeves, P. R., & Octavia, S. (2009). Population structure, origins and evolution of major Salmonella enterica clones. Infection, Genetics and Evolution : Journal of Molecular Epidemiology and Evolutionary Genetics in Infectious Diseases, 9(5), 996–1005.
Lee, K., Iwata, T., Nakadai, A., Kato, T., Hayama, S., Taniguchi, T., & Hayashidani, H. (2011). Prevalence of Salmonella, Yersinia and Campylobacter spp. in feral raccoons (Procyon lotor) and masked palm civets (Paguma larvata) in Japan. Zoonoses and Public Health, 58(6), 424–431.
Leekitcharoenphon, P., Hendriksen, R. S., Le Hello, S., Weill, F. X., Baggesen, D. L., Jun, S. R., Ussery, D. W., Lund, O., Crook, D. W., Wilson, D. J., & Aarestrup, F. M. (2016). Global genomic epidemiology of Salmonella enterica Serovar Typhimurium DT104. Applied and Environmental Microbiology, 82(8), 2516–2526.
LeJeune, J.T., & Pearl, D.L. (2014). Wildlife and environmental contributions to the contamination of meats with antimicrobial-resistant bacteria. In Encyclopedia of Meat Sciences (2nd edition).
42
Leonard, E. K., Pearl, D. L., Janecko, N., Finley, R. L., Reid-Smith, R. J., Weese, J. S., & Peregrine, A. S. (2015). Risk factors for carriage of antimicrobial-resistant Salmonella spp. and Escherichia coli in pet dogs from volunteer households in Ontario, Canada, in 2005 and 2006. American Journal of Veterinary Research, 76(11), 959–968.
Liakopoulos, A., Mevius, D. J., Olsen, B., & Bonnedahl, J. (2016). The colistin resistance mcr-1 gene is going wild. The Journal of Antimicrobial Chemotherapy, 71(8), 2335–2336.
Mąka, Ł., & Popowska, M. (2016). Antimicrobial resistance of Salmonella spp. isolated from food. Roczniki Panstwowego Zakladu Higieny, 67(4), 343–358.
Majowicz, S. E., Musto, J., Scallan, E., Angulo, F. J., Kirk, M., O'Brien, S. J., Jones, T. F., Fazil, A., Hoekstra, R. M., & International Collaboration on Enteric Disease 'Burden of Illness' Studies (2010). The global burden of nontyphoidal Salmonella gastroenteritis. Clinical Infectious Diseases, 50(6), 882–889.
Maslow, J., Mulligan, M., & Arbeit, R. (1993). Molecular epidemiology: Application of contemporary techniques to the typing of microorganisms. Clinical Infectious Diseases, 17, 153–162.
McDermott, P. F., Zhao, S., Wagner, D. D., Simjee, S., Walker, R. D., & White, D. G. (2002). The food safety perspective of antibiotic resistance. Animal Biotechnology, 13(1), 71–84.
McEwen, S. A., & Collignon, P. J. (2018). Antimicrobial Resistance: A One Health Perspective. Microbiology spectrum, 6(2), 10.1128/microbiolspec.ARBA-0009-2017.
McManus M. C. (1997). Mechanisms of bacterial resistance to antimicrobial agents. American Journal of Health-System, 54(12), 1420–1446.
Mead P.S., Slutsker L., Dietz V., McCaig L.F., Bresee J.S., Shapiro C., Griffin P.M. & Tauxe R.V. (1999). Food-related illness and death in the United States. Emerging Infectious Diseases, 5(5), 607-625
Michael, G. B., & Schwarz, S. (2016). Antimicrobial resistance in zoonotic nontyphoidal Salmonella: an alarming trend? Clinical Microbiology and Infection, 22(12), 968–974.
Mintz E. (2003). A riddle wrapped in a mystery inside an enigma: Brainerd diarrhoea turns 20. Lancet, 362, 2037–2038.
Moser, S., Seth-Smith, H., Egli, A., Kittl, S., & Overesch, G. (2020). Campylobacter jejuni from canine and bovine cases of campylobacteriosis express high antimicrobial resistance rates against (fluoro)quinolones and tetracyclines. Pathogens (Basel, Switzerland), 9(9), 691.
Mutschall, S. K., Hetman, B. M., Bondo, K. J., Gannon, V., Jardine, C. M., & Taboada, E. N. (2020). Campylobacter jejuni strain dynamics in a raccoon (Procyon lotor) population in
43
southern Ontario, Canada: High prevalence and rapid subtype turnover. Frontiers in Veterinary Science, 7, 27.
Nadon, C., Van Walle, I., Gerner-Smidt, P., Campos, J., Chinen, I., Concepcion-Acevedo, J., Gilpin, B., Smith, A. M., Man Kam, K., Perez, E., Trees, E., Kubota, K., Takkinen, J., Nielsen, E. M., Carleton, H., & FWD-NEXT Expert Panel (2017). PulseNet International: Vision for the implementation of whole genome sequencing (WGS) for global food-borne disease surveillance. Euro surveillance, 22(23), 30544.
Neoh, H. M., Tan, X. E., Sapri, H. F., & Tan, T. L. (2019). Pulsed-field gel electrophoresis (PFGE): A review of the "gold standard" for bacteria typing and current alternatives. Infection, Genetics and Evolution, 74, 103935.
Newell, D. G., Mughini-Gras, L., Kalupahana, R., & Wagenaar, J.A. (2016). Campylobacter epidemiology—sources and routes of transmission for human infection. In Campylobacter: Features, detection, and prevention of foodborne disease. Klein, G., Ed., Long, UK; San Diego, CA, USA: Academic Press.
Nowakiewicz, A., Zięba, P., Ziółkowska, G., Gnat, S., Muszyńska, M., Tomczuk, K., Majer Dziedzic, B., Ulbrych, Ł., & Trościańczyk, A. (2016). Free-living species of carnivorous mammals in Poland: Red fox, beech marten, and raccoon as a potential reservoir of Salmonella, Yersinia, Listeria spp. and coagulase-positive Staphylococcus. Plos One, 11(5), e0155533.
Octavia, S., & Lan, R. (2006). Frequent recombination and low level of clonality within Salmonella enterica subspecies I. Microbiology (Reading, England), 152(Pt 4), 1099–1108.
Oravcová, V., Peixe, L., Coque, T. M., Novais, C., Francia, M. V., Literák, I., & Freitas, A. R. (2018). Wild corvid birds colonized with vancomycin-resistant Enterococcus faecium of human origin harbor epidemic vanA plasmids. Environment International, 118, 125–133.
Osaki, A., Sashika, M., Abe, G., Shinjo, K., Fujimoto, A., Nakai, M., Shimozuru, M., & Tsubota, T. (2019). Comparison of feeding habits and habitat use between invasive raccoons and native raccoon dogs in Hokkaido, Japan. BMC Ecology, 19(1), 35.
Park, S., Navratil, S., Gregory, A., Bauer, A., Srinath, I., Szonyi, B., Nightingale, K., Anciso, J., Jun, M., Han, D., Lawhon, S., & Ivanek, R. (2015). Multifactorial effects of ambient temperature, precipitation, farm management, and environmental factors determine the level of generic Escherichia coli contamination on preharvested spinach. Applied and Environmental Microbiology, 81(7), 2635–2650.
Pérez-Losada, M., Cabezas, P., Castro-Nallar, E., & Crandall, K. A. (2013). Pathogen typing in the genomics era: MLST and the future of molecular epidemiology. Infection, Genetics and Evolution, 16, 38–53.
44
Pintar, K. D., Christidis, T., Thomas, M. K., Anderson, M., Nesbitt, A., Keithlin, J., Marshall, B., & Pollari, F. (2015). A systematic review and meta-analysis of the Campylobacter spp. prevalence and concentration in household pets and petting zoo animals for use in exposure assessments. Plos One, 10(12), e0144976.
Pintar, K., Thomas, K. M., Christidis, T., Otten, A., Nesbitt, A., Marshall, B., Pollari, F., Hurst, M., & Ravel, A. (2017). A comparative exposure assessment of Campylobacter in Ontario, Canada. Risk Analysis, 37(4), 677–715.
Pogreba-Brown, K., Austhof, E., Armstrong, A., Schaefer, K., Villa Zapata, L., McClelland, D. J., Batz, M. B., Kuecken, M., Riddle, M., Porter, C. K., & Bazaco, M. C. (2020). Chronic gastrointestinal and joint-related sequelae associated with common foodborne illnesses: A scoping review. Foodborne Pathogens and Disease, 17(2), 67–86.
Poirel, L., Potron, A., De La Cuesta, C., Cleary, T., Nordmann, P., & Munoz-Price, L. S. (2012). Wild coastline birds as reservoirs of broad-spectrum-β-lactamase-producing Enterobacteriaceae in Miami Beach, Florida. Antimicrobial Agents and Chemotherapy, 56(5), 2756–2758.
Procter, T. D., Pearl, D. L., Finley, R. L., Leonard, E. K., Janecko, N., Reid-Smith, R. J., Weese, J. S., Peregrine, A. S., & Sargeant, J. M. (2014). A cross-sectional study examining Campylobacter and other zoonotic enteric pathogens in dogs that frequent dog parks in three cities in south-western Ontario and risk factors for shedding of Campylobacter spp. Zoonoses and Public Health, 61(3), 208–218.
Rolland, R. M., Hausfater, G., Marshall, B., & Levy, S. B. (1985). Antibiotic-resistant bacteria in wild primates: Increased prevalence in baboons feeding on human refuse. Applied and Environmental Microbiology, 49, 791-794.
Prange, S., Gehrt, S. D., & Hauver, S. (2011). Frequency and duration of contacts between free-ranging raccoons: Uncovering a hidden social system. Journal of Mammalogy, 92(6), 1331–1342.
Rabsch, W., Simon, S., & Humphrey, T. (2013). Public health aspects of Salmonella infections. In: Barrow PA, Methner U, editors. Salmonella in Domestic Animals (2nd edition). Wallingford, UK: CABI Publishing.
Ramey, A. M., & Ahlstrom, C. A. (2020). Antibiotic resistant bacteria in wildlife: Perspectives on trends, acquisition and dissemination, data gaps, and future directions. Journal of Wildlife Diseases, 56(1), 1–15.
Riley, L. W., Remis, R. S., Helgerson, S. D., McGee, H. B., Wells, J. G., Davis, B. R., Hebert, R. J., Olcott, E. S., Johnson, L. M., Hargrett, N. T., Blake, P. A., & Cohen, M. L. (1983). Hemorrhagic colitis associated with a rare Escherichia coli serotype. The New England Journal of Medicine, 308(12), 681–685.
45
Romero, J. L., Grande Burgos, M. J., Pérez-Pulido, R., Gálvez, A., & Lucas, R. (2017). Resistance to antibiotics, biocides, preservatives and metals in bacteria isolated from seafoods: Co-selection of strains resistant or tolerant to different classes of compounds. Frontiers in Microbiology, 8, 1650.
Rosatte, R. C. (2008). Hinterland Who's Who: Raccoon. Canadian Wildlife Federation, Environment and Climate Change Canada. Available at: https://www.hww.ca/en/wildlife/mammals/raccoon.html (accessed on 18 August 2020).
Rosatte, R., Ryckman, M., Ing, K., Proceviat, S., Allan, M., Bruce, L., Donovan, D., & Davies, J. C. (2010). Density, movements, and survival of raccoons in Ontario, Canada: Implications for disease spread and management. Journal of Mammalogy, 91, 122-135.
Rozwandowicz, M., Brouwer, M., Fischer, J., Wagenaar, J. A., Gonzalez-Zorn, B., Guerra, B., Mevius, D. J., & Hordijk, J. (2018). Plasmids carrying antimicrobial resistance genes in Enterobacteriaceae. The Journal of Antimicrobial Chemotherapy, 73(5), 1121–1137.
Ryan, K. J., Schainuck, L. I., Hickman, R. O., & Striker, G. E. (1969). Colistimethate toxicity: report of a fatal case in a previously healthy child. The Journal of the American Medical Association, 207(11), 2099–2101.
Santos, T., Silva, N., Igrejas, G., Rodrigues, P., Micael, J., Rodrigues, T., Resendes, R., Gonçalves, A., Marinho, C., Gonçalves, D., Cunha, R., & Poeta, P. (2013). Dissemination of antibiotic resistant Enterococcus spp. and Escherichia coli from wild birds of Azores Archipelago. Anaerobe, 24, 25–31.
Scallan, E., Griffin, P. M., Angulo, F. J., Tauxe, R. V., & Hoekstra, R. M. (2011). Foodborne illness acquired in the United States--unspecified agents. Emerging Infectious Diseases, 17(1), 16–22.
Schiaffino, F., Platts-Mills, J., & Kosek, M. N. (2019). A One Health approach to prevention, treatment, and control of campylobacteriosis. Current Opinion in Infectious Diseases, 32(5), 453–460.
Schindler, B. D., & Kaatz, G. W. (2016). Multidrug efflux pumps of Gram-positive bacteria. Drug Resistance Updates : Reviews and Commentaries in Antimicrobial and Anticancer Chemotherapy, 27, 1–13.
Schubert, S., & Kostrzewa, M. (2017). MALDI-TOF MS in the microbiology laboratory: Current trends. Current Issues in Molecular Biology, 23, 17–20.
Schumann, P., & Pukall, R. (2013). The discriminatory power of ribotyping as automatable technique for differentiation of bacteria. Systematic and Applied Microbiology, 36(6), 369–375.
46
Schürch, A. C., Arredondo-Alonso, S., Willems, R. J. L., & Goering, R. V. (2018). Whole genome sequencing options for bacterial strain typing and epidemiologic analysis based on single nucleotide polymorphism versus gene-by-gene–based approaches. Clinical Microbiology and Infection, 24(4), 350–354.
Sewell, A. M., & Farber, J.M. (2001). Foodborne outbreaks in Canada linked to produce. Journal of Food Protection, 64, 1863–1877.
Shen, Y., Shi, X., Shen, J., Wang, Y., & Wang, S. (2019). [Application of whole genome sequencing technology and bioinformatics analysis in antimicrobial resistance researches]. Chinese journal of biotechnology, 35(4), 541–557.
Shintani, M., Sanchez, Z. K., & Kimbara, K. (2015). Genomics of microbial plasmids: Classification and identification based on replication and transfer systems and host taxonomy. Frontiers in Microbiology, 6, 242.
Silva, C., Calva, E., & Maloy, S. (2014). One health and food-borne disease: Salmonella transmission between humans, animals, and plants. Microbiology Spectrum, 2(1), OH-0020-2013.
Singh, V. (2013). “Antimicrobial Resistance”. Microbial pathogens and strategies for combating them: Science, technology and education. Formatex Research Center, vol. 1, 291-296.
Singh, B. B., & Gajadhar, A. A. (2014). Role of India's wildlife in the emergence and re-emergence of zoonotic pathogens, risk factors and public health implications. Acta Tropica, 138, 67–77.
Skurnik, D., Clermont, O., Guillard, T., Launay, A., Danilchanka, O., Pons, S., Diancourt, L., Lebreton, F., Kadlec, K., Roux, D., Jiang, D., Dion, S., Aschard, H., Denamur, M., Cywes-Bentley, C., Schwarz, S., Tenaillon, O., Andremont, A., Picard, B., Mekalanos, J., … Denamur, E. (2016). Emergence of antimicrobial-resistant Escherichia coli of animal origin spreading in humans. Molecular Biology and Evolution, 33(4), 898–914.
Smillie, C., Garcillán-Barcia, M. P., Francia, M. V., Rocha, E. P., & de la Cruz, F. (2010). Mobility of plasmids. Microbiology and Molecular Biology Reviews, 74(3), 434–452.
Smith H.W. & Tucker J.F. (1978). The effect of antimicrobial feed additives on the colonization of the alimentary tract of chickens by Salmonella typhimurium. The Journal of Hygiene, (Lond.), 80(2), 217-231.
Spratt, B. G., & Maiden, M. C. (1999). Bacterial population genetics, evolution and epidemiology. Philosophical Transactions of the Royal Society of London. Series B, Biological sciences, 354(1384), 701–710.
47
Strassle, P. D., Gu, W., Bruce, B. B., & Gould, L. H. (2019). Sex and age distributions of persons in foodborne disease outbreaks and associations with food categories. Epidemiology and Infection, 147, e200.
Strawn, L. K., Gröhn, Y. T., Warchocki, S., Worobo, R. W., Bihn, E. A., & Wiedmann, M. (2013). Risk factors associated with Salmonella and Listeria monocytogenes contamination of produce fields. Applied and Environmental Microbiology, 79(24), 7618–7627.
Suginaka, H., Ichikawa, A., & Kotani, S. (1975). Penicillin-resistant mechanisms in Pseudomonas aeruginosa: Binding of penicillin to Pseudomonas aeruginosa KM 338. Antimicrobial Agents and Chemotherapy, 7(5), 629–635.
Tanwar, J., Das, S., Fatima, Z. & Hameed S. (2014). Multidrug resistance: An emerging crisis. Interdisciplinary Perspectives on Infectious Diseases, 541340, 1-7.
Tausova, D., Dolejska, M., Cizek, A., Hanusova, L., Hrusakova, J., Svoboda, O., Camlik, G., & Literak, I. (2012). Escherichia coli with extended-spectrum β-lactamase and plasmid-mediated quinolone resistance genes in great cormorants and mallards in Central Europe. The Journal of Antimicrobial Chemotherapy, 67(5), 1103–1107.
Thaller, M. C., Migliore, L., Marquez, C., Tapia, W., Cedeño, V., Rossolini, G. M., & Gentile, G. (2010). Tracking acquired antibiotic resistance in commensal bacteria of Galápagos land iguanas: No man, no resistance. Plos One, 5(2), e8989.
Threlfall, E. J. (2002). Antimicrobial drug resistance in Salmonella: Problems and perspectives in food- and water-borne infections. FEMS Microbiology Reviews, 26(2), 141–148.
Thomas, M. K., Murray, R., Flockhart, L., Pintar, K., Pollari, F., Fazil, A., Nesbitt, A., & Marshall, B. (2013). Estimates of the burden of foodborne illness in Canada for 30 specified pathogens and unspecified agents, circa 2006. Foodborne Pathogens and Disease, 10(7), 639–648.
Thomas, M., & Feng, Y. (2020). Risk of foodborne illness from pet food: Assessing pet owners' knowledge, behavior, and risk perception. Journal of Food Protection, 10.4315/JFP-20-108. Advance online publication.
Todd E. (1990). Epidemiology of foodborne illness: North America. Lancet (London, England), 336(8718), 788–790.
Toleman, M. A., & Walsh, T. R. (2011). Combinatorial events of insertion sequences and ICE in Gram-negative bacteria. FEMS microbiology reviews, 35(5), 912–935.
Urwin, R., & Maiden, M. C. (2003). Multi-locus sequence typing: A tool for global epidemiology. Trends in Microbiology, 11(10), 479–487.
48
Van Breda, L. K., & Ward, M. P. (2017). Evidence of antimicrobial and disinfectant resistance in a remote, isolated wild pig population. Preventive Veterinary Medicine, 147, 209–212.
Van Den Bogaard, A. E., & Stobberingh, E. E. (2000). Epidemiology of resistance to antibiotics: Links between animals and humans. International Journal of Antimicrobial Agents, 14, 327-335.
Van Goethem, M. W., Pierneef, R., Bezuidt, O., Van De Peer, Y., Cowan, D. A., & Makhalanyane, T. P. (2018). A reservoir of 'historical' antibiotic resistance genes in remote pristine Antarctic soils. Microbiome, 6(1), 40.
Veldman, K., van Tulden, P., Kant, A., Testerink, J., & Mevius, D. (2013). Characteristics of cefotaxime-resistant Escherichia coli from wild birds in the Netherlands. Applied and Environmental Microbiology, 79(24), 7556–7561.
Vogt, N. A., Stevens, C., Pearl, D. L., Taboada, E. N., & Jardine, C. M. (2020). Generalizability and comparability of prevalence estimates in the wild bird literature: Methodological and epidemiological considerations. Animal Health Research Reviews, 1–7.
Vranakis, I., Chochlakis, D., Sandalakis, V., Tselentis, Y., & Psaroulaki, A. (2012). Cost- and time-effectiveness of application of MALDI-TOF mass spectrometry methodology in a food and water microbiology laboratory. Archives of Hellenic Medicine, 29, 477– 479.
Wales A., & Davies, R. (2015). Co-selection of resistance to antibiotics, biocides and heavy metals, and its relevance to foodborne pathogens. Antibiotics, 4, 567-604.
Wang, J., Ma, Z. B., Zeng, Z. L., Yang, X. W., Huang, Y., & Liu, J. H. (2017). The role of wildlife (wild birds) in the global transmission of antimicrobial resistance genes. Zoological Research, 38(2), 55–80.
Wassenaar, T. M., Geilhausen, B., & Newell, D. G. (1998). Evidence of genomic instability in Campylobacter jejuni isolated from poultry. Applied and Environmental Microbiology, 64, 1816–1821.
Wegener, H. C., Hald, T., Lo Fo Wong, D., Madsen, M., Korsgaard, H., Bager, F., Gerner-Smidt, P., & Mølbak, K. (2003). Salmonella control programs in Denmark. Emerging Infectious Diseases, 9(7), 774–780.
Wheeler, E., Hong, P., Bedon, L. C., & Mackie, R. I. (2012). Carriage of antibiotic-resistant enteric bacteria varies among sites in Galapagos reptiles. Journal of Wildlife Diseases 48:56-67.
Whipps, J. M., Hand, P., Pink, D., & Bending, G. D. (2008). Phyllosphere microbiology with special reference to diversity and plant genotype. Journal of Applied Microbiology,105, 1744–1755.
49
World Health Organization. (2015). "WHO estimates of the global burden of foodborne diseases". Foodborne disease burden epidemiology reference group, 2007-2015. Available from: https://apps.who.int/iris/bitstream/handle/10665/199350/9789241565165_eng.pdf?sequence=1 (accessed on 10 August 2020).
World Health Organization. (2020). "Estimating the burden of foodborne diseases". Available at: https://www.who.int/activities/estimating-the-burden-of-foodborne-diseases (accessed on 10 August 2020).
Yoshida, C. E., Kruczkiewicz, P., Laing, C. R., Lingohr, E. J., Gannon, V. P., Nash, J. H., & Taboada, E. N. (2016). The Salmonella in silico typing resource (SISTR): An open web-accessible tool for rapidly typing and subtyping draft Salmonella genome assemblies. Plos One, 11(1), e0147101.
Zhu, Y. G., Zhao, Y., Zhu, D., Gillings, M., Penuelas, J., Ok, Y. S., Capon, A., & Banwart, S. (2019). Soil biota, antimicrobial resistance and planetary health. Environment International, 131, 105059.
50
CHAPTER 2
2 Epidemiology of Campylobacter jejuni in raccoons (Procyon lotor)
on swine farms and in conservation areas in southern Ontario
As published in: Zoonoses and Public Health, 2021 Feb; 68(1):19-28. doi: 10.1111/zph.12786.
Epub 2020 Nov 23.
Article authorship: N.A. Vogt, D.L. Pearl, E.N. Taboada, S.K. Mutschall, K.J. Bondo, C.M.
Jardine
2.1 Abstract
Campylobacter is a leading cause of foodborne illness in humans worldwide. Sources of
infection are often difficult to identify, and are, generally, poorly understood. Recent work
suggests that wildlife may represent a source of Campylobacter for human infections. Using a
repeated cross-sectional study design, raccoons were trapped on five swine farms and five
conservation areas in southern Ontario from 2011-2013. Our objectives were to: 1) assess the
impact of seasonal, climatic, location, annual and raccoon demographic factors on the occurrence
of Campylobacter jejuni in these animals; and 2) identify clusters of C. jejuni in space, time, and
space-time using spatial scan statistics. Multi-level multivariable logistic regression was used to
examine the odds of isolating C. jejuni, with site and animal modeled as random intercepts. The
following independent variables were examined: raccoon age and sex, year, location type,
season, temperature and rainfall. A total of 1096 samples were obtained from 627 raccoons;
51
46.3% were positive for C. jejuni. The following interactions and their main effects were
significant (p<0.05) and retained in the final model: seasonXtemperature, yearXrainfall,
yearXtemperature. Based on the results from our multivariable model and spatial scan statistics,
climatic variables (i.e., rainfall, temperature, and season) were associated with the carriage of C.
jejuni by raccoons, but the effects were not consistent, and varied by location and year. Although
raccoons may pose a zoonotic risk due to their carriage of Campylobacter, further work is
required to characterize the transmission and movement of this microorganism within the
ecosystem.
Keywords
Campylobacter; epidemiology; foodborne pathogens; wildlife; zoonoses
2.2 Introduction
Campylobacter is a leading cause of bacterial foodborne illness in humans worldwide and
represents a substantial public health burden (Cody et al., 2019). Although most human cases of
Campylobacter enteritis are self-limiting, some may require antimicrobial therapy and/or
hospitalization. In a small number of patients, serious immune-mediated complications can be
triggered by previous infection with Campylobacter and include such conditions as reactive
arthritis (1-5% of cases) and Guillain-Barré syndrome (<0.1% of cases; Acheson & Allos, 2001;
Pope et al., 2007). Certain aspects of the epidemiology of Campylobacter infections in humans
are well established: poultry has repeatedly been identified as an important source for human
infections, regardless of geographic region (Mughini-Gras et al., 2012; Lévesque et al., 2013;
52
Iglesias-Torrens et al., 2018). On the other hand, the contribution of environmental sources of
Campylobacter (e.g., contaminated water, wildlife) is less well understood, and these sources
have been estimated to contribute a range of anywhere from 4 to 20% of clinical cases (Sheppard
et al., 2009; Mughini-Gras et al., 2012; Lévesque et al., 2013; Berthenet et al., 2019). Although a
diversity of mammalian and avian wildlife species have been shown to harbour Campylobacter
as a commensal organism, their role in the epidemiology of human Campylobacter infections is
not well understood (Sippy et al., 2012; Bondo et al., 2019; Du et al., 2019).
Among wildlife species, wild birds have received the most attention as potential sources
of Campylobacter in previous source attribution research (Greig et al., 2015); not only is
Campylobacter widely recognized as a commensal organism in birds, but the highly mobile
nature of birds may facilitate the transmission and dissemination of Campylobacter between
different components of the ecosystem (Reed et al., 2003). Until recently, medium-sized
mammalian species such as raccoons (Procyon lotor) had received little attention in this area of
research (Lee et al., 2011; Rainwater et al., 2017; Mutschall et al., 2020). Due to their prevalent
populations within both urban and rural environments, and their propensity to interact with
domestic animals and humans, raccoons may represent an important link in the epidemiology of
Campylobacter within the ecosystem.
This work represents part of a larger longitudinal study examining the role of raccoons in
the maintenance and transmission of foodborne and other zoonotic bacteria (e.g., Salmonella,
Escherichia coli, Clostridium difficile) in a southern Ontario ecosystem (Bondo et al., 2015;
Bondo et al., 2016a; Bondo et al., 2016b; Mutschall et al., 2020). We believe this particular
ecosystem presents an ideal setting to study the intersection between different ecosystem
53
components; in addition to numerous conservation areas and a large human population (~ 1
million people), the Grand River watershed region contains many waterways and rivers within an
active agricultural region. Previous work examining Campylobacter genotypes (based on
comparative genomic fingerprinting) isolated from the raccoons in this ecosystem revealed a
high prevalence of Campylobacter (46.3%; Mutschall et al., 2020). Examination of the
longitudinal data from the subset of raccoons sampled multiple times over the three-year study
period demonstrated that Campylobacter was carried only for short periods of time (often 30
days or less; Mutschall et al., 2020). While some of the Campylobacter genotypes carried by
raccoons in this study had only ever been identified in raccoons, genotypes common to both
livestock sources and human clinical infections were prevalent (66.2% of genotypes; Mutschall
et al., 2020). To build on previous work examining Campylobacter in raccoons in this
longitudinal study, our objectives were to: 1) assess the potential impact of seasonal, climatic,
location, annual, and raccoon demographic factors on the isolation of Campylobacter from
raccoon fecal swabs; and 2) identify clusters of raccoons harboring Campylobacter in space,
time, and space-time using spatial scan statistics.
2.3 Methods
2.3.1 Sample collection
Procedures for trapping and handling of live raccoons were approved by the Animal Care
Committee at the University of Guelph following the guidelines of the Canadian Committee on
Animal Care (Permit number: 11R015). A detailed description of the study area and sites, as well
54
as a map of the study area, is available from Bondo et al. (2016b). Details regarding live-trapping
and processing of raccoons have been previously described (Bondo et al., 2016b). Briefly,
raccoons were live-trapped on five swine farms and five conservation areas near Guelph and
Cambridge in southern Ontario from May through November, 2011-2013. Each site was trapped
once every five weeks. Traps were set in slightly different locations between trapping sessions to
optimize capture rates; exact GPS coordinates for trapped animals were obtained using a
handheld GPS unit. Individual animals were sampled up to once per monthly trapping week;
however, additional samples were collected from the same raccoon if they were caught in
subsequent months. Fecal swabs were kept on ice in the field and refrigerated until further
processing.
2.3.2 Campylobacter isolation and speciation
Methods used to isolate thermophilic Campylobacter species are described elsewhere (Mutschall
et al., 2020). Samples which were collected prior to October 2011 were processed using the
‘conventional method’: enrichment followed by direct plating onto selective media. Samples
collected in subsequent years were processed using the ‘membrane method’: enrichment
followed by passive membrane filtration onto selective media (Mutschall et al., 2020).
Presumptive Campylobacter spp. colonies were identified based on growth and colony
morphology, as well as oxidase tests. DNA extraction from purified cultures was performed
using the EZ1 DNA tissue kit (Qiagen, Mississauga, Canada) according to manufacturer
instructions. Confirmation of presumptive Campylobacter spp. colonies was performed using
multiplex PCR targeting a Campylobacter genus-specific region of the 16S rRNA gene, as well
as mapA and ceuE genes for Campylobacter jejuni and Campylobacter coli identification,
55
respectively (Denis et al., 1999). Visualization of PCR amplicons was performed using a
QIAxcel capillary electrophoresis instrument with the DNA screening kit.
2.3.3 Statistical methods
Multi-level multivariable logistic regression was used to model the odds of isolating C. jejuni
from raccoon fecal swabs. Unspeciated Campylobacter and C. coli were reported descriptively
due to a low prevalence (<5%). The following independent variables were examined: raccoon
age (adult or juvenile), raccoon sex (male or female), location type (swine farm or conservation
area), year, season, and environmental factors (rainfall, temperature). Based on previous work
(Rosatte et al., 2010; Bondo et al., 2016b), season was grouped into two distinct categories
related to raccoon reproduction: rearing (May-July), and pre-denning/dispersal (August-
November). For environmental factors (total rainfall, mean air temperature), two different time
periods (previous 14 days, 30 days) were examined and modeled independently from one another
to avoid multicollinearity, thus, one model was built for each time period. These particular time
periods were selected to explore the potential impact of different periods of rainfall and
temperature on the odds of isolating C. jejuni from raccoons. Temperature and rainfall data were
downloaded from Environment Canada from the nearest weather station with complete data
(Fergus, Ministry of the Environment, ON). Data from the next nearest weather station in
Guelph, Ontario were used to fill in missing values. The Bayesian Information Criterion (BIC)
was used to select the best model from among the two models with different time periods used to
model environmental variables (previous 14 days rainfall and temperature vs. previous 30 days
rainfall and temperature).
56
Continuous independent variables which could not be appropriately modeled linearly or
as a quadratic were subsequently categorized into three quantiles (low, medium, high). The
correlation between mean temperature and total rainfall was examined for each model (14 days,
30 days) using the relevant correlation coefficient. If variables were highly correlated, │ρ │>
0.8, then each was modeled separately to avoid multicollinearity.
Univariable models were initially constructed using a liberal p-value (0.20). A manual
backwards model-building approach was used, and variables were retained in the final model if
they were significant or part of a significant interaction term. Variables were also retained in the
final model if their removal resulted in a >20% change in the coefficient of a significant potential
explanatory variable, as long as they were not identified a priori as an intervening variable for
that potential explanatory variable (Dohoo et al., 2014). A significance level of a=0.05 was used,
and all tests were two-tailed. Only biologically plausible two-way interactions were tested within
the following groups of variables: 1) location type, season, year; 2) year, season, rainfall,
temperature; and 3) age, sex, season. Contrasts were performed to interpret interactions
involving categorical variables.
In order to account for potential clustering due to repeated sampling from the same site
and/or animal, these were fitted as random intercepts. A random intercept was excluded from the
final model if a likelihood ratio test was non-significant, the variance component was very small,
and if model fit was not improved based on changes to the Akaike’s Information Criterion (AIC)
and BIC values with the addition of the random effect. Variance partition coefficients (VPCs)
were estimated using the latent variable technique (Dohoo et al., 2014). All statistical tests were
performed using STATA (STATA Intercooled 14.2; StataCorp, College Station, Texas, USA).
57
2.3.4 Assessing model fit
The fit of multi-level models was assessed by examining Pearson and Deviance residuals for
outliers. The model assumptions of normality and homogeneity of variance for the best linear
unbiased predictors (BLUPs) were examined graphically using a normal quantile plot, and by
plotting the predicted outcome against the BLUPs, respectively. If the assumptions were not met
by the BLUPs, the AICs and BICs from models with and without random effects were compared
in order to confirm that the addition of random effects improved model fit.
2.3.5 Spatial scan statistics
Potential clusters of C. jejuni-positive raccoons were identified retrospectively in space, time,
and space-time with spatial scan statistics using SatScanTM (Martin Kulldorff and Information
Management Services Inc., Boston, MA). A Bernoulli model was used, with cases defined as C.
jejuni-positive raccoons, and controls defined as C. jejuni-negative raccoons. For temporal scans,
month was used as the minimum time aggregation unit, and the maximum temporal scanning
window was set at 50% of the study period. For spatial scans, the maximum scanning window
was set at 50% of the population at risk. These maximum settings do not preclude the reporting
of smaller clusters; verification of smaller clusters using narrower scanning windows was not
performed since our goal in using spatial scan statistics was to validate the findings from our
statistical model, and to do so by maximizing specificity rather than sensitivity. Hierarchical
reporting of clusters with the ‘no pairs of centers both in each other’s clusters’ option was used
in space-time scans; the results of these scans were reviewed manually and only statistically
significant clusters not overlapping in space and time were reported. Statistical significance for
58
all scans was determined using 999 Monte Carlo replications, a two-sided hypothesis, and a
significance level of α = 0.05.
2.4 Results
2.4.1 Samples collected and isolates
A total of 1096 samples were collected from 627 unique raccoons. Of these samples, 508 were
positive for Campylobacter spp. (46.3%; 95%CI: 43.4-49.4%), 502 of which were positive for C.
jejuni. Unspeciated Campylobacter was isolated from 6 samples (one of which was also positive
for C. jejuni), and C. coli was isolated from 1 sample. After exclusion of samples containing only
non-jejuni species of Campylobacter, a total of 1086 samples from 625 unique raccoons were
included in our multi-level multivariable model of the odds of isolating C. jejuni: three
observations from three raccoons trapped at two different locations were randomly selected and
excluded, and one observation from a raccoon that was erroneously sampled twice within a week
was randomly selected and excluded. Age and sex were not recorded for three and one raccoon,
respectively. The proportion of C. jejuni-positive samples by age, sex, location type, year, and
season are presented in Table 2.1.
2.4.2 Univariable and multivariable analyses
Rainfall and temperature explanatory variables were categorized since their relationship with the
outcome was neither linear nor quadratic. In the univariable analyses, the following independent
variables were identified with significant unconditional associations with the outcome: age, year,
59
location type, season, rainfall, temperature (Table 2.2). The model examining climatic conditions
(rainfall, temperature) over the 14 days prior to sampling is presented (Table 2.3), as the BIC for
this model was lower than the model examining these variables over the 30 days prior to
sampling. Our final model included the following interactions and their simple effects:
seasonXtemperature, yearXtemperature, yearXrainfall. Both random intercepts (location, animal)
were retained in the final multivariable model based on the size of their variance components, a
significant likelihood ratio test, and an improvement in model fit based on a reduction in both
AIC and BIC with their inclusion. No outlying observations were identified, and the BLUPs met
the model assumptions of normality and homogeneity of variance. Most of the variance was
identified at the level of the sample (80.4%), followed by animal (13.2%), and site (5.4%; Table
2.3).
The prevalence of C. jejuni was significantly greater during the rearing period (Aug-Nov)
compared to pre-denning/dispersal (May-July) for both low and medium temperatures (Table
2.4). For high temperatures, the odds of isolating C. jejuni did not differ significantly between
seasons (Table 2.4). During the pre-denning/dispersal season, C. jejuni was significantly less
likely to be isolated with medium temperatures compared with low temperatures (Table 2.4).
During the rearing period (Aug-Nov), the odds of isolating C. jejuni were significantly lower for
all of the following temperature comparisons: high vs. low, medium vs. low, high vs. medium
(Table 2.4). For the interaction between rainfall and year, the following comparisons were
associated with a significantly greater odds of isolating C. jejuni, but only in 2013: medium vs.
low, high vs. low (Table 2.5). In 2012, a high total rainfall was associated with a significantly
greater odds of isolating C. jejuni compared to medium total rainfall (Table 2.5). For the
interaction between temperature and year, the medium category of temperature was associated
60
with a significantly lower odds of isolating C. jejuni compared with the low category of
temperature in 2011 (Table 2.6). In 2012, compared to a low temperature, both high and medium
temperatures were associated with a significantly greater odds of isolating C. jejuni (Table 2.6).
2.4.3 Spatial scan statistics
To protect the anonymity of farms, all location names were de-identified. Two statistically
significant spatial clusters were identified: a low prevalence of C. jejuni was noted at
conservation area 2, whereas a high prevalence was detected on farms 7, 9 and 10 (Table 2.7).
One statistically significant temporal cluster was identified: a high prevalence was observed from
September 2011 until October 2012 (Table 2.7). Four statistically significant space-time clusters
were detected: a high prevalence of C. jejuni on conservation area 4 from October 2011 until
October 2012, a low prevalence on conservation area 2 from May 2011 until September 2011, a
high prevalence on swine farms 6, 7, 9, and 10 from July 2012 until October 2012, and a low
prevalence on conservation areas 1 and 2 from July 2013 until August 2013 (Table 2.7).
2.5 Discussion
The overall prevalence of Campylobacter among raccoons sampled in our study was high
(46.3%) in comparison with previous research examining free-ranging raccoons in urban areas in
Japan (1.3%) and New York City (6%; Lee et al., 2011; Rainwater et al., 2017). However, a
similar prevalence was found among raccoons trapped on swine, beef, and dairy farms in the
same geographic region in 2010 (40.8%; Viswanathan et al., 2017a). In this work by
Viswanathan et al. (2017a, 2017b), there was little evidence of Campylobacter transmission
61
between livestock and raccoons and other small mammals trapped on these farms, based on
genotypes determined using 40-gene comparative genomic fingerprinting (CGF). In contrast,
previous work comparing Campylobacter genotypes isolated from raccoons in this study to a
larger database (the Canadian Campylobacter CGF database) containing samples from clinical
human cases, environmental sources (e.g., water), and poultry and livestock sources outside of
our study region, revealed that human- and livestock-associated genotypes were highly prevalent
among these animals (66.4%; Mutschall et al., 2020). Although this recent work by Mutschall et
al. (2020) suggests that transmission between raccoons and livestock occurs frequently, the type
of location (conservation area vs. swine farm) was not included in our final model since it was
not associated with Campylobacter carriage, was not part of a significant interaction term, and its
removal did not substantially alter the coefficients of any significant potential explanatory
variable. Together, these findings suggest that transmission of Campylobacter between raccoons
and livestock is occurring on a broad scale, but there is currently little evidence confirming that
direct transmission is occurring on farms within our study region. However, the identification
and distribution of different Campylobacter species, in particular, C. coli (commonly associated
with pigs), provides evidence of transmission from agricultural sources to wildlife, albeit perhaps
not to the extent that might be expected. Campylobacter coli is typically an uncommon
Campylobacter species in raccoons and other wildlife (<2% overall prevalence; Lee et al., 2011
Konicek et al., 2016; Krawiec et al., 2017; Kwon et al., 2017; Viswanathan et al., 2017a;
MacDonald et al., 2018; Bondo et al., 2019). In previous work, C. coli has been isolated from
wildlife in association with swine farm environments (Hald et al., 2016; Vogt et al., 2018).
Manure pit samples collected from swine farms on which raccoons in our study population were
sampled were almost exclusively C. coli (94.1%, 95%CI: 71.3-99.8%; Mutschall et al., 2020).
62
Correspondingly, the sole C. coli isolate in our current study was identified in a raccoon trapped
on a swine farm. Although these findings suggest that wildlife may be acquiring C. coli from
farm environments, the low overall prevalence of C. coli in our study provides evidence, in
conjunction with work by Viswanathan et al. (2017b), that direct transmission of this
Campylobacter species on farms is not occurring frequently, considering the proximity of
raccoons to these farms.
Unlike previous work with Salmonella in this raccoon population (Bondo et al., 2016b),
demographic factors such as sex and age were not significantly associated with the carriage of C.
jejuni. Our final multivariable model identified climatic, annual and seasonal factors as
significant predictors of the carriage of C. jejuni within this population of raccoons. Local
temperature and rainfall have previously been shown to influence the survival of Campylobacter
spp. in the environment (Moriarty et al., 2011; Moriarty et al., 2012; Smith et al., 2016). In
addition, these climatic factors were associated with the carriage of other bacterial organisms
(i.e., Salmonella, antimicrobial resistant E. coli) within this population of raccoons (Bondo et al.,
2016a; Bondo et al., 2016b). Apart from a seasonal trend (higher in the fall) with low and
moderate temperatures (4.7-19.4°C), the effects of rainfall, temperature and year were not
consistent in our study. There was some evidence that higher amounts of rainfall in the 14 days
prior to sampling was associated with a greater odds of isolating C. jejuni, but this was only
statistically significant for certain years. Overall, spatial scan statistics were suggestive of lower
prevalences of C. jejuni in conservation areas and higher prevalences in farm locations; however,
one high space-time cluster was identified in a conservation area, suggesting that Campylobacter
prevalence is variable at specific times, independent of location type. The results from both the
63
multivariable model and the spatial scan statistics suggest that certain combinations of seasonal
and climatic factors interact together in certain years and locations to produce favourable
conditions for C. jejuni, but no consistent or overarching trends could be identified with either
approach. It is also possible that other unmeasured factors may need to be present for the effects
of rainfall and temperature to have a consistent effect across time.
Our investigative approach was exploratory, and, as a result, carries certain limitations.
The population of C. jejuni carried by these raccoons was previously demonstrated to contain a
diversity of subtypes (Mutschall et al., 2020), which are known to possess differing capabilities
for survival in the environment outside of a host (e.g., water-adapted strains). Although strain-
related differences may have contributed to our findings, for the purposes of statistical
modelling, we treated all C. jejuni as equivalent, which is an oversimplification; the lack of clear
epidemiological patterns in our results is consistent with the notion that the pattern of C. jejuni
carriage by raccoons may vary by strain-type. In addition, it is possible that a change in the
isolation method during the study period impacted our findings. Although the sensitivity of both
isolation methods is comparable, the specificity of the membrane filtration method is greater,
yielding fewer false positive results (Jokinen et al., 2012). Results from our study did not reveal
any distinct annual patterns that could be explained by this change in methods, however the
increased test specificity in 2012 and 2013 may have impacted our findings to some extent,
potentially obscuring definitive patterns related to other potential explanatory variables.
Another limitation of our work is that we examined only two periods of time for the
impact of rainfall and temperature (i.e., 14 days, 30 days). It is possible that other time periods of
these climatic factors, or other characterizations of these factors (e.g., average daily rainfall) may
64
more accurately predict the occurrence of Campylobacter. In addition, certain predictors which
may be important for understanding why some raccoons were carriers of C. jejuni and others are
not, may be missing from our model. Although we characterized the type of environment on a
broad scale (i.e., farm vs. conservation area) and this variable was not identified as a risk factor,
there may be characteristics of the local environment (e.g., proximity to rivers) which facilitate
the transmission of Campylobacter, but which we could not account for due to insufficient
replication of such environments (i.e., only 10 locations). Such details were qualitatively
identified as relevant in previous work on antimicrobial resistant E. coli in this population of
raccoons; Bondo et al. (2016a) suggested that raccoons likely acquired resistant bacteria from
dumpsters within conservation areas. Another limitation of this work relates to our method of
sampling raccoons; since this was conducted with baited traps, the selection of study individuals
is likely biased to some degree (Vogt et al., 2020). As a result of sampling animals in this way, it
is possible that our study group represents a subpopulation of raccoons that may be either more
or less likely to carry Campylobacter due to differences in physiology, behaviour, and/or feeding
habits. Within this study group, many of the raccoons were captured on multiple occasions and
some of these animals might be characterized as “trap-happy”, with one individual having been
captured a total of eight times over the three-year period. Overall, 43% of our samples were
collected from animals captured on multiple occasions (Mutschall et al., 2020), therefore, the use
of spatial scan statistics should be interpreted with caution since multiple test results may be
contributed by an individual animal. However, we felt the value of performing the scans to
explore the distribution of C. jejuni-positive animals in space and time without pre-defined
categories of parameters and to corroborate our multivariable model outweighed its limitations.
65
The role of raccoons in the epidemiology of human Campylobacter infections is
unknown, but our current study suggests that climatic and seasonal factors are important
predictors of the carriage of Campylobacter by raccoons. Although raccoons in our study region
have been shown to carry human- and livestock-associated Campylobacter genotypes (Mutschall
et al., 2020), our study did not provide any strong indications about the source(s) for these
animals. Further work is needed to better understand the carriage of Campylobacter by this
potentially overlooked wildlife species. Given that Campylobacter has been shown to persist in
manure-amended soils for up to 4 months at a range of temperatures (Rogers et al., 2011), more
thorough sampling of farms is perhaps needed to uncover sources of this bacterium.
Campylobacter outbreaks in humans are uncommon, and although a previous human outbreak at
a wildlife rehabilitation centre was linked with raccoons (Saunders et al., 2017), it is more likely
that wildlife play a role in sporadic human cases, for which sources are generally poorly
understood (Ravel et al., 2016). Future research employing highly discriminatory methods (e.g.,
whole-genome sequencing) and the use of multiple sampling sources from farms and the
environment (e.g., soil, water), in tandem with epidemiological assessments of the importance of
environmental variables such as rainfall, temperature, and season, will provide valuable insights
concerning the role of wildlife in the epidemiology of Campylobacter within the ecosystem.
66
2.6 References
Acheson, D., & Allos, M. (2001). Campylobacter jejuni infections: update on emerging issues and trends. Clinical Infectious Diseases, 32, 1201–1206.
Berthenet, E., Thépault, A., Chemaly, M., Rivoal, K., Ducournau, A., Buissonnière, A., … Lehours, P. (2019). Source attribution of Campylobacter jejuni shows variable importance of chicken and ruminant reservoirs in non-invasive and invasive French clinical isolates. Scientific Reports, 9, 8098.
Bondo, K. J., Weese, J. S., Rousseau, J., & Jardine, C. M. (2015). Longitudinal study of Clostridium difficile shedding in raccoons on swine farms and conservation areas in Ontario, Canada. BMC Veterinary Research, 11, 1–6.
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016a). Epidemiology of antimicrobial resistance in Escherichia coli isolates from raccoons (Procyon lotor) and the environment on swine farms and conservation areas in southern Ontario. PLoS ONE, e0165303.
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016b). Impact of season, demographic and environmental factors on Salmonella occurrence in raccoons (Procyon lotor) from swine farms and conservation areas in southern Ontario. PLoS ONE, e0161497.
Bondo, K. J., Pearl, D. L., Janecko, N., Reid-Smith, R. J., Parmley, E. J., Weese, J. S., … & Jardine, C. M. (2019). Salmonella, Campylobacter, Clostridium difficile, and anti-microbial resistant Escherichia coli in the faeces of sympatric meso-mammals in southern Ontario, Canada. Zoonoses and Public Health, 66, 406–416.
Cody, A. J., Maiden, M. C., Strachan, N. J., & McCarthy, N. D. (2019). A systematic review of source attribution of human campylobacteriosis using multilocus sequence typing. Euro Surveillance, 24, pii=1800696.
Denis, M., Soumet, C., Rivoal, K., Ermel, G., Blivet, D., Salvat, G., & Colin, P. (1999). Development of a m‐PCR assay for simultaneous identification of Campylobacter jejuni and C. coli. Letters in Applied Microbiology, 29, 406–410.
Dohoo, I, Martin, W & Stryhn, H. (2014). Veterinary epidemiologic research, 3rd ed. Charlottetown, Prince Edward Island: VER Inc.
Du, J., Luo, J., Huang, J., Wang, C., Li, M., Wang, B., …& He, H. (2019). Emergence of genetic diversity and multi-drug resistant Campylobacter jejuni from wild birds in Beijing, China. Frontiers in Microbiology, 10, 2433.
67
Greig, J., Rajić, A., Young, I., Mascarenhas, M., Waddell, L., & Lejeune, J. (2015). A scoping review of the role of wildlife in the transmission of bacterial pathogens and antimicrobial resistance to the food chain. Zoonoses and Public Health, 62, 269–284.
Hald, B, Skov, MN, Nielsen, EM, Rahbek, C, Madsen, JJ, Wainø, M, …& Madsen, M. (2016). Campylobacter jejuni and Campylobacter coli in wild birds on Danish livestock farms. Acta Veterinaria Scandinavica, 58:11.
Iglesias-Torrens, Y., Miró, E., Guirado, P., Llovet, T., Muñoz, C., Cerdà-Cuéllar, M., …& Navarro, F. (2018). Population structure, antimicrobial resistance, and virulence-associated genes in Campylobacter jejuni isolated from three ecological niches: gastroenteritis patients, broilers, and wild birds. Frontiers in Microbiology, 9, 1676.
Jokinen, C. C., Koot, J. M., Carrillo, C. D., Gannon, V. P., Jardine, C. M., Mutschall, S. K., Topp, E., & Taboada, E. N. (2012). An enhanced technique combining pre-enrichment and passive filtration increases the isolation efficiency of Campylobacter jejuni and Campylobacter coli from water and animal fecal samples. Journal of Microbiological Methods, 91, 506–513.
Konicek, C., Vodrážka, P., Barták, P., Knotek, Z., Hess, C., Račka, K., … & Troxler, S. (2016). Detection of zoonotic pathogens in wild birds in the cross-border region Austria –Czech Republic. Journal of Wildlife Diseases, 52, 850–861.
Krawiec, M., Woźniak-Biel, A., Bednarski, M., & Wieliczko, A. (2017). Antimicrobial susceptibility and genotypic characteristic of Campylobacter spp. isolates from free-living birds in Poland. Vector-Borne and Zoonotic Diseases, 17, 755–763.
Kwon, Y. K., Oh, J. Y., Jeong, O. M., Moon, O. K., Kang, M. S., Jung, B. Y.,…& Lee, H. S. (2017). Prevalence of Campylobacter species in wild birds of South Korea. Avian Pathology, 46, 474–480.
Lee, K., Iwata, T., Nakadai, A., Kato, T., Hayama, S., Taniguchi, T., & Hayashidani, H. (2011). Prevalence of Salmonella, Yersinia and Campylobacter spp. in feral raccoons (Procyon lotor) and masked palm civets (Paguma larvata) in Japan. Zoonoses and Public Health, 58, 424–431.
Levesque, S., Fournier, E., Carrier, N., Frost, E., D. Arbeit, R., & Michaud, S. (2013). Campylobacteriosis in urban versus rural areas: a case-case study integrated with molecular typing to validate risk factors and to attribute sources of infection. PLoS ONE, 8.
MacDonald, A. M., Jardine, C. M., Susta, L., Slavic, D., & Nemeth, N. M. (2018). Survey for bacteria and antimicrobial resistance in wild turkeys (Meleagris gallopavo) in Ontario, Canada. Avian Diseases, 62, 184–188.
68
Moriarty, E. M., Mackenzie, M. L., Karki, N., & Sinton, L. W. (2011). Survival of Escherichia coli, Enterococci, and Campylobacter spp. in sheep feces on pastures. Applied and Environmental Microbiology, 77, 1797–1803.
Moriarty, E. M., Weaver, L., Sinton, L. W., & Gilpin, B. (2012). Survival of Escherichia coli, Enterococci and Campylobacter jejuni in Canada Goose faeces on pasture. Zoonoses and Public Health, 59, 490–497.
Mughini-Gras, L., Smid, J. H., Wagenaar, J. A., de Boer, A. G., Havelaar, A. H., Friesema, I. H. M., … & van Pelt, W. (2012). Risk factors for campylobacteriosis of chicken, ruminant, and environmental origin: a combined case-control and source attribution analysis. PLoS ONE, 7.
Mutschall, S. K., Hetman, B. M., Bondo, K. J., Gannon, V. P. J., Jardine, C. M., & Taboada, E. N. (2020). Campylobacter jejuni strain dynamics in a raccoon (Procyon lotor) population in southern Ontario, Canada: high prevalence and rapid subtype turnover. Frontiers in Veterinary Science, 7, 27.
Pope, J. E., Krizova, A., Garg, A. X., Thiessen-Philbrook, H., & Ouimet, J. M. (2007). Campylobacter reactive arthritis: a systematic review. Seminars in Arthritis and Rheumatism, 37, 48–55.
Rainwater, K. L., Marchese, K., Slavinski, S., Humberg, L. A., Dubovi, E. J., Jarvis, J. A., … & Calle, P. P. (2017). Health survey of free-ranging raccoons (Procyon lotor) in central park, New York, New York, USA: implications for human and domestic animal health. Journal of Wildlife Diseases, 53, 272–284.
Ravel, A., Pintar, K., Nesbitt, A., & Pollari, F. (2016). Non-food-related risk factors of campylobacteriosis in Canada: A matched case-control study. BMC Public Health, 16, 1016.
Reed, K. D., Meece, J. K., Henkel, J. S., & Shukla, S. K. (2003). Birds, migration and emerging zoonoses: West nile virus, lyme disease, influenza A and enteropathogens. Clinical Medicine & Research, 1, 5–12.
Rogers, S. W., Donnelly, M., Peed, L., Kelty, C. A., Mondal, S., Zhong, Z., & Shanks, O. C. (2011). Decay of bacterial pathogens, fecal indicators, and real-time quantitative PCR genetic markers in manure-amended soils. Applied and Environmental Microbiology, 77, 4839–4848.
Rosatte, R., Ryckman, M., Ing, K., Proceviat, S., Allan, M., Bruce, L., ... & Davies J.C. (2010). Density, movements, and survival of raccoons in Ontario, Canada: implications for disease spread and management. Journal of Mammalogy, 91, 122-135.
69
Saunders, S., Smith, K., Schott, R., Dobbins, G., & Scheftel, J. (2017). Outbreak of campylobacteriosis associated with raccoon contact at a wildlife rehabilitation centre, Minnesota, 2013. Zoonoses and Public Health, 64, 222–227.
Sheppard, S. K., Dallas, J. F., Strachan, N. J., MacRae, M., McCarthy, N. D., Wilson, D. J., Gormley, F. J., Falush, D., Ogden, I. D., Maiden, M. C., & Forbes, K. J. (2009). Campylobacter genotyping to determine the source of human infection. Clinical Infectious Diseases, 48, 1072–1078.
Sippy, R., Sandoval-Green, C. M. J., Sahin, O., Plummer, P., Fairbanks, W. S., Zhang, Q., & Blanchong, J. A. (2012). Occurrence and molecular analysis of Campylobacter in wildlife on livestock farms. Veterinary Microbiology, 157, 369–375.
Smith, S., Meade, J., Gibbons, J., McGill, K., Bolton, D., & Whyte, P. (2016). The impact of environmental conditions on Campylobacter jejuni survival in broiler faeces and litter. Infection Ecology & Epidemiology, 6, 31685.
Viswanathan, M., Pearl, D. L., Taboada, E. N., Parmley, E. J., Mutschall, S., & Jardine, C. M. (2017a). Molecular and statistical analysis of Campylobacter spp. and antimicrobial-resistant Campylobacter carriage in wildlife and livestock from Ontario farms. Zoonoses and Public Health, 64, 194–203.
Viswanathan, M., Pearl, D. L., Taboada, E. N., Parmley, E. J., Mutschall, S. K., & Jardine, C. M. (2017b). Cluster analysis of Campylobacter jejuni genotypes isolated from small and medium-sized mammalian wildlife and bovine livestock from Ontario farms. Zoonoses and Public Health, 64, 185–193.
Vogt, N.A., Stevens, C.P.G., Pearl, D.L., Taboada, E.N., & Jardine, C.M. (2020). Generalizability and comparability of prevalence estimates in the wild bird literature: methodological and epidemiological considerations. Animal Health Research Reviews, 1-7.
Vogt, N. A., Pearl, D. L., Taboada, E. N., Mutschall, S. K., Janecko, N., Reid-Smith, R., …& Jardine, C. M. (2018). Epidemiology of Campylobacter, Salmonella and antimicrobial resistant Escherichia coli in free-living Canada geese (Branta canadensis) from three sources in southern Ontario. Zoonoses and Public Health, 65, 873–886.
70
2.7 Tables
Table 2.1: Proportion (%) of raccoon fecal swabs testing positive for Campylobacter jejunia overall and by
age, sex, location type, season, and year in southern Ontario from May-November 2011-2013 (n=1086b
samples)
Independent variable Category (n) % Positive (95%CI)
Agec Adult (n=751) 69.1 (66.3-71.9) Juvenile (n=332) 30.6 (27.8-33.4) Sexc Female (n=579) 53.3 (50.3-56.2)
Male (n=506) 46.6 (43.6-49.6) Location type Swine farm (n=401) 36.9 (34.0-39.9)
Conservation area (n=685) 63.1 (60.1-65.9) Season May to July (n=581) 53.5 (50.5-56.5) Aug. to Nov. (n=505) 46.5 (43.5-49.5) Year 2011 (n=346) 31.9 (29.1-34.7) 2012 (n=443) 40.8 (37.9-43.8) 2013 (n=297) 27.3 (24.7-30.1) Overall 46.0 (43.0-49.0)
a Other species of Campylobacter (Campylobacter coli, unspeciated Campylobacter) were uncommonly identified
(<1% total prevalence) and reported descriptively in the results section of this manuscript. b Not included in this total are four samples which were randomly selected and excluded: three observations from
three raccoons trapped at two different locations, and another observation from a raccoon that was erroneously
sampled twice within one week. cAge was not recorded for 3 raccoons and sex was not recorded for 1 raccoon.
71
Table 2.2: Results from univariable multi-levela logistic regression models showing associations between the occurrence of Campylobacter jejunib in
raccoon fecal swabs and raccoon sex and age, year, season, location type, mean air temperature, and total rainfall in Ontario, Canada (n=1086c)
Independent variable Sub-Category Odds
Ratio 95% CI p-value Variance [VPCd]
(95% CI) Site-level Animal-level Sample-level
Raccoon sex Female REF
Male 1.18 0.88-1.57 0.273 0.24 [5.9]
(0.08-0.74) 0.51 [12.6] (0.19-1.40)
[81.4]
Raccoon age Adult REF
Juvenile 1.37 1.00-1.87 0.054 0.25 [6.2] (0.08-0.76)
0.50 [12.4] (0.18-1.38)
[81.4]
Season Rearing period (May-July) REF Pre-denning/dispersal
(Aug.-Nov.) 2.60 1.96-3.48 <0.001 0.28 [7.0] (0.10-0.84)
0.39 [9.8] (0.11-1.33) [83.0]
Location type Conservation area REF
Swine farm 1.72 0.93-3.15 0.083 0.18 [4.5] (0.06-0.57)
0.51 [12.8] (0.19-1.40)
[82.7]
Year 2011 REF
2012 1.97 1.38-2.82 <0.001 0.19 [4.8] (0.06-0.60)
0.48 [12.1] (0.17-1.39)
[83.1]
2013 0.64 0.44-0.94 0.023 0.19 [4.8]
(0.06-0.60) 0.48 [12.1] (0.17-1.39) [83.1]
Temperature (over 14 days prior) low (4.7-16.5°C) REF
medium (16.6-19.4 °C) 0.59 0.42-0.82 0.002 0.28 [6.8]
(0.09-0.83) 0.54 [13.1] (0.20-1.45) [80.0]
high (19.5-22.9°C) 0.62 0.44-0.86 0.005 0.28 [6.8]
(0.09-0.83) 0.54 [13.1] (0.20-1.45)
[80.0]
Temperature (over 30 days prior) low (8.4-16.4°C) REF
medium (16.5-19.1°C) 0.71 0.51-0.99 0.042 0.26 [6.4]
(0.09-0.79) 0.53 [13.0] (0.20-1.43) [80.6]
high (19.2-22°C) 0.71 0.51-0.98 0.040 0.26 [6.4]
(0.09-0.79) 0.53 [13.0] (0.20-1.43)
[80.6]
72
Total rainfall (over 14 days prior) low (0-27.6mm) REF
medium (28.4-60.8mm) 1.63 1.17-2.28 0.004 0.24 [6.0]
(0.08-0.74) 0.46 [11.5] (0.15-1.36)
[82.4]
high (62.8-130.1mm) 1.02 0.72-1.44 0.906 0.24 [6.0]
(0.08-0.74) 0.46 [11.5] (0.15-1.36) [82.4]
Total rainfall (over 30 days prior) low (3-82.8mm) REF
medium (83.2-111.8mm) 1.24 0.90-1.72 0.186 0.22 [5.5]
(0.07-0.68) 0.46 [11.6] (0.16-1.35) [82.9]
high (112.2-273.8mm) 0.61 0.44-0.85 0.004 0.22 [5.5]
(0.07-0.68) 0.46 [11.6] (0.16-1.35)
[82.9]
a Site and animal were modeled as random intercepts. Significant associations are indicated in bold (based on liberal p-value of 0.20), REF = referent group, and
CI = confidence interval. b Other species of Campylobacter (Campylobacter coli, unspeciated Campylobacter) were uncommonly identified (<1% total prevalence) and reported
descriptively.
c Not included in this total are four samples which were randomly selected and excluded: three observations from three raccoons trapped at two different
locations, and another observation from a raccoon that was erroneously sampled twice within one week.
d VPC - variance partition coefficient.
73
Table 2.3: Results from multi-levela multivariable logistic regression models showing associations between the occurrence of Campylobacter jejunib in
raccoon fecal swabs and year, season, air temperature, total rainfall and interaction effects in Ontario, Canada (n=1086c)
Independent variable Sub-Category Odds Ratio 95%CI p-value Season Rearing period (May-July) REF Pre-denning/dispersal (Aug.-Nov.) 9.77 5.17-18.45 <0.001 Year 2011 REF 2012 0.48 0.15-1.51 0.211 2013 0.13 0.03-0.46 0.002 Air temperatured low (4.7-16.5°C) REF medium (16.6-19.4 °C) 0.40 0.16-0.98 0.046 high (19.5-22.9°C) 0.49 0.15-1.55 0.225
Total rainfalld low (0-27.6mm) REF medium (28.4-60.8mm) 0.79 0.30-2.11 0.642 high (62.8-130.1mm) 0.44 0.16-1.21 0.111 Interactionse Season*Air temperature (Aug.-Nov.)*medium 0.43 0.19-0.97 0.043
(Aug.-Nov.)*high 0.08 0.03-0.22 <0.001
Year*Total rainfall 2012*medium 1.04 0.33-3.27 0.948 2012*high 4.25 1.30-13.95 0.017 2013*medium 5.95 1.46-24.20 0.013 2013*high 5.48 1.48-20.25 0.011 Year*Air temperature 2012*medium 6.76 2.28-20.08 0.001 2012*high 9.32 2.88-30.09 <0.001 2013*medium 1.23 0.44-3.44 0.688 2013*high 2.52 0.67-9.52 0.173 Variance [VPCf] Site-level 0.22 [5.4] 0.06-0.73 Animal-level 0.54 [13.2] 0.18-1.61 Sample-level [80.4]
a Site and animal were modeled as random intercepts. Significant associations are indicated in bold, REF = referent group, and CI = confidence interval. b Other species of Campylobacter (Campylobacter coli, unspeciated Campylobacter) were uncommonly identified (<1% total prevalence) and reported
descriptively.
74
c Not included in this total are four samples which were randomly selected and excluded: three observations from three raccoons trapped at two different
locations, and another observation from a raccoon that was erroneously sampled twice within one week. d The mean air temperature and total rainfall over the 14 days prior to sampling was examined. e To interpret season, year, temperature, rainfall, and their interaction effects, refer to contrasts in Tables 2.4, 2.5, and 2.6.
f VPC - variance partition coefficient.
75
Table 2.4: Contrasts derived from the multi-level logistic regression model for the occurrence of Campylobacter jejuni in raccoon fecal swabs (Table 2.3)
in Ontario, Canada to interpret interaction effects between season and mean air temperature in the 14 days prior to sampling (n=1086)
Contrast Sub-category Odds Ratiob 95% CI p-value
Aug.-Nov. vs. May-July (REF)
Low temperature (4.7-16.5°C) 9.77 5.17-18.45 <0.001
Medium temperature (16.6-19.4 °C) 4.20 2.25-7.84 <0.001
High temperature (19.5-22.9°C) 0.76 0.36-1.61 0.472
Medium vs. low temperature (REF) Rearing period (May-July) 0.40 0.16-0.98 0.046
Pre-denning/dispersal (Aug.-Nov.) 0.17 0.08-0.38 <0.001
High vs. low temperature (REF) Rearing period (May-July) 0.49 0.15-1.55 0.225
Pre-denning/dispersal (Aug.-Nov.) 0.04 0.01-0.11 <0.001
High vs. medium temperature (REF) Rearing period (May-July) 1.23 0.40-3.77 0.716
Pre-denning/dispersal (Aug.-Nov.) 0.22 0.07-0.67 0.007
76
Table 2.5: Contrasts derived from the multi-level logistic regression model for the occurrence of Campylobacter jejuni in raccoon fecal swabs (Table 2.3)
in Ontario, Canada to interpret interaction effects between year and total rainfall in the 14 days prior to sampling (n=1086)
Contrast Sub-category Odds Ratiob 95% CI p-value
Medium vs. low rainfalla (REF)
2011 0.79 0.30-2.11 0.642
2012 0.82 0.42-1.61 0.573
2013 4.72 1.73-12.89 0.002
High vs. low rainfalla (REF)
2011 0.44 0.16-1.21 0.111
2012 1.88 0.84-4.19 0.124
2013 2.42 1.03-5.68 0.042
High vs. medium rainfalla (REF)
2011 0.55 0.24-1.28 0.169
2012 2.28 1.02-5.12 0.046
2013 0.51 0.24-1.10 0.088
a Rainfall categories were: low (0-27.6mm), medium (28.4-60.8mm), high (62.8-130.1mm). b Significant associations are indicated in bold, REF = referent group, and CI = confidence interval.
77
Table 2.6: Contrasts derived from the multi-level logistic regression model for the occurrence of Campylobacter jejuni in raccoon fecal swabs (Table 2.3)
in Ontario, Canada to interpret interaction effects between year and mean air temperature in the 14 days prior to sampling (n=1086)
Contrast Sub-category Odds Ratiob 95% CI p-value
Medium vs. low temperaturea (REF)
2011 0.40 0.16-0.98 0.046
2012 2.70 1.15-6.31 0.022
2013 0.49 0.20-1.23 0.129
High vs. low temperaturea (REF)
2011 0.49 0.15-1.55 0.225
2012 4.57 2.04-10.25 <0.001
2013 1.24 0.51-3.01 0.641
High vs. medium temperaturea (REF)
2011 1.23 0.40-3.77 0.716
2012 1.69 0.84-3.40 0.137
2013 2.51 0.91-6.97 0.077
a Temperature categories were: low (4.7-16.5°C), medium (16.6-19.4 °C), high (19.5-22.9°C). b Significant associations are indicated in bold, REF = referent group, and CI = confidence interval.
78
Table 2.7: Statistically significant spatial, temporal and space-time clusters of Campylobacter jejunia positive raccoon fecal swabs in Ontario, Canada
using spatial scan statistics with Bernoulli models (n=1086b)
Type of cluster
Number of raccoon samples
High or low cluster Location(s)c/Month(s) No. expected
cases Observed/Expected Relative risk p-value
Spatial 125 low Conservation area 3 57.44 0.40 0.37 <0.001
Spatial 220 high Farms 7, 9, 10 101.09 1.25 1.33 0.003
Temporal 559 high Sept. 2011– Oct. 2012 256.85 1.32 2.02 0.001
Space-time 124 high Conservation area 1 from Oct. 2011–Oct. 2012
56.98 1.68 1.85 0.001
Space-time 120 high Farms 6, 7, 9, 10 from July 2012 – Oct. 2012
55.14 1.67 1.82 0.001
Space-time 49 low Conservation area 3 from May 2011 – Sept. 2011
22.51 0.13 0.13 0.001
Space-time 49 low Conservation areas 3, 4 from July 2013 – Aug. 2013
22.51 0.13 0.13 0.001
a Other species of Campylobacter (Campylobacter coli, unspeciated Campylobacter) were uncommonly identified (<2% overall prevalence) and reported
descriptively in the results section of this manuscript. b Not included in this total are four samples which were randomly selected and excluded: three observations from three raccoons trapped at two different
locations, and another observation from a raccoon that was erroneously sampled twice within one week. c To protect the anonymity of farm locations, all location names were de-identified, and exact locations are not provided.
79
CHAPTER 3
3 Genomic epidemiology of Salmonella and Escherichia coli and
associated antimicrobial resistance in raccoons (Procyon lotor),
swine manure pits, and soil samples on swine farms in southern
Ontario: A risk factor analysis
3.1 Abstract
Foodborne illness caused by Salmonella and the increasing prevalence of antimicrobial
resistance continue to pose a major threat to human and animal health. In this work, we
examined whole-genome sequencing data from Salmonella and E. coli isolates collected from
raccoons and environmental sources on farms in southern Ontario to better understand the
contribution of wildlife to the dissemination of Salmonella and antimicrobial resistance. All
Salmonella and phenotypically resistant E. coli previously collected from live-trapped raccoons,
soil, and manure pits on five swine farms as part of a previous longitudinal study were included.
We assessed for evidence of potential transmission of Salmonella and E. coli between different
sources and farms using a combination of population structure assessments (using core-genome
multi-locus sequence typing), direct comparisons of multi-drug resistant isolates, and
epidemiologically-informed statistical modeling of antimicrobial resistance (AMR) genes and
plasmid replicons identified in silico. Mixed univariable logistic regression models were fitted to
80
assess the impact of source type, farm location, and sampling year on the occurrence of select
resistance genes and plasmid types. A total of 159 Salmonella and 96 resistant E. coli isolates
were included. We found evidence of potential transmission of both Salmonella strains and
resistance genes between raccoon, soil, and swine manure pit samples locally on swine farms,
and between different swine farms. Salmonella sequence types of international importance were
identified in raccoons (e.g., ST19 Typhimurium), and a handful of these isolates demonstrated
AMR, but AMR among Salmonella isolates was uncommon overall (<5%). Several plasmid
types and resistance genes were significantly associated with source type and farm location, with
a higher odds on different farms, depending on the outcome examined. Those plasmid replicons
and resistance genes associated with source type were consistently more likely to be identified in
raccoons than swine manure pits, suggesting that manure pits are not a primary source of those
resistance determinants for raccoons. More comprehensive sampling of farms, and assessment of
farms with other livestock species, as well as additional environmental sources (e.g., rivers) may
help to further elucidate primary sources of resistance genes in this region. Since most plasmid
types identified in Salmonella were not associated with resistance genes and given our selective
inclusion of phenotypically resistant E. coli, a more detailed examination of plasmids in
susceptible and resistant bacterial populations is also warranted.
Keywords
Antimicrobial resistance, Procyon lotor, source attribution, wildlife, whole-genome sequencing
81
3.2 Introduction
The rise of antimicrobial resistance is a major global threat to the health of humans and animals
alike (De Oliveira et al., 2020; Graham et al., 2019). There is mounting evidence of widespread
movement of antimicrobial resistance (AMR) determinants (e.g., genes and the plasmids
associated with their movement) within natural environments (Iwu et al., 2020; Arnold et al.,
2016; Greig et al., 2015), and genes conferring resistance to high-priority antimicrobials (e.g.,
mcr-1) have consistently been identified in avian and mammalian wildlife across the world
(Anyanwu et al., 2020; Fischer et al., 2013; Wang et al., 2017). It is generally recognized that
wild animals represent sentinels of environmental AMR pollution, but recent work suggests that
wildlife may also physically disseminate AMR determinants from one location to another
through their feces (Ahlstrom et al., 2018; Ahlstrom et al., 2019). With recognition of the
importance of One Health approaches that consider different sampling sources, there is a need to
integrate epidemiological investigations with new technologies such as whole-genome
sequencing, which permit a higher resolution assessment of the genetic basis of AMR (Collineau
et al., 2019; Croucher et al., 2013).
A number of investigations combining genomics with epidemiology to examine
foodborne pathogens and/or antimicrobial resistance in wildlife have recently been performed
(Bloomfield et al., 2017; Toro et al., 2016; Mather et al., 2016; Ahlstrom et al., 2018; Ahlstrom
et al., 2019). With much of the literature on antimicrobial resistance and wildlife focused on wild
birds (Torres et al., 2020), mammalian wildlife such as raccoons (Procyon lotor), arguably merit
further examination in this context due to their prevalent populations, tendency to forage in
anthropogenic environments, and general proximity to human and domestic animal settings
82
(French et al., 2019). This work is part of a larger longitudinal study of wildlife (primarily
raccoons) on swine farms and conservation areas in southern Ontario that was conducted
between 2011 and 2013 (Bondo et al., 2016a; Bondo et al., 2016b; Bondo et al., 2019). The study
region which includes the Grand River Watershed (6800km2), and a population of ~1 million
people in an area with intensive agriculture represents a unique opportunity to examine the
intersection between wildlife, livestock, and environmental sources in a heavily populated region
of southern Ontario, Canada. Although the prevalence of phenotypic resistance in E. coli and
Salmonella isolated from wildlife and soil samples in this previous longitudinal study did not
differ significantly between swine farms and conservation areas, certain strains (e.g.,
Typhimurium var. DT 104 Copenhagen), serovars (e.g., Salmonella Agona) and resistance
patterns only appeared in samples obtained on swine farms (Bondo et al., 2016a; Bondo et al.,
2016b).
The objective of this study was to examine in detail the subset of samples from swine
farms in this previous study, utilizing whole-genome sequencing data from Salmonella and
resistant E. coli isolates from these different sources to assess for potential transmission of these
organisms and associated AMR determinants. To assess for potential transmission between
different sources, a combination of population structure assessments (using core-genome multi-
locus sequence typing based on Enterobase schemes; https://enterobase.warwick.ac.uk/),
epidemiologically-informed statistical modeling of select AMR determinants (i.e., genes,
plasmid replicons), and direct comparisons of multidrug resistant isolates was performed. For
assessments of potential transmission utilizing statistical modeling, our aim was to determine the
impact of source type, farm location, and sampling year on the occurrence of AMR determinants
83
(identified in silico), while accounting for clustering of samples obtained from the same animal
or swine manure pit.
3.3 Methods
3.3.1 Sample collection
Samples for this study were previously collected as part of a longitudinal study of wildlife on
swine farms and conservation areas in southern Ontario (2011-2013; Bondo et al., 2016a; Bondo
et al., 2016b). For this study, we included only samples originating from swine farms. Sampling
sources included raccoons, swine manure pits, and soil samples. The study region and sampling
methods have been previously described and are available in Bondo et al. (2016a; 2016b).
Briefly, samples were obtained from monthly sampling of five swine farms in the Grand River
watershed, near the cities of Guelph and Cambridge in Ontario, Canada. Raccoons were live-
trapped, and animals were anesthetized to obtain a rectal fecal swab using a Cary-Blair
applicator (BBL CultureSwab, Bd; Becton, Dickinson and Company, Annapolis, Maryland,
USA). Individual animals were ear-tagged and microchipped for subsequent identification, and
animals were sampled up to once monthly, but animals caught within the same trapping month
were released. Swine manure pit and soil samples were obtained from each site at the beginning
of each trapping week. For soil samples, 10 g of soil was collected from within a 2-m radius of
each animal trap and stored in a sterile container. Swine manure pits were sampled by pooling
samples from two different depths (i.e., top 1/3, and mid-depth) at three different locations
around the pit. All samples were kept on ice in the field until further processing. All Salmonella
84
isolates originating from swine farms were selected for sequencing and inclusion in this study.
For E. coli isolates, only isolates demonstrating phenotypic resistance to at least one of 15
antimicrobials examined (see details below) were selected for sequencing and included in the
present study. During 2011, three different E. coli isolates were cultured from each sample; for
samples with more than one isolate demonstrating phenotypic resistance, a random number
generator was used to select one resistant isolate for sequencing.
3.3.2 Culture and susceptibility testing
Samples were cultured for Salmonella and E. coli within three days of collection at the McEwen
Group Research Lab at the Canadian Research Institute for Food Safety, University of Guelph
(Guelph, Ontario, Canada) as previously described (Bondo et al., 2016a; Bondo et al., 2016b).
Phenotypic susceptibility testing was performed at the Antimicrobial Resistance Reference
Laboratory (National Microbiology Laboratory at Guelph, Public Health Agency of Canada,
Guelph, Ontario, Canada) in accordance with methods outlined by the Canadian Integrated
Program for Antimicrobial Resistance Surveillance (CIPARS; Government of Canada, 2015).
Isolates were tested using the National Antimicrobial Resistance Monitoring System (NARMS;
Sensititre, Thermo Scientific), and antimicrobial panel CMV3AGNF, which includes the
following 15 antimicrobials: gentamicin (GEN), kanamycin (KAN), streptomycin (STR),
amoxicillin-clavulanic acid (AMC), cefoxitin (FOX), ceftiofur (TIO), ceftriaxone (CRO),
ampicillin (AMP), chloramphenicol (CHL), sulfisoxazole (SOX), trimethoprim-
sulfamethoxazole (SXT), tetracycline (TCY), nalidixic acid (NAL), ciprofloxacin (CIP), and
azithromycin (AZM).
85
3.3.3 DNA extraction, whole-genome sequencing and genome assembly
Cultures of Salmonella and E. coli were grown on Mueller Hinton Agar and incubated overnight
at 35 °C. Cultures were then distributed to the National Microbiology Laboratory (NML) in
Winnipeg for DNA extraction and sequencing, or these steps were performed on site, at the
University of Guelph and the NML in Guelph, Ontario, respectively. Genomic DNA extraction
was performed using 1 ml of culture as input to the Qiagen DNEasy plant and tissue 96 kit,
according to manufacturer protocols (Qiagen, Hilden, Germany). Sequencing was then
performed at the NML in Guelph or in Winnipeg, using Nextera XT libraries and Illumina
MiSeq version 3 (600-cycle kit) or NextSeq550 platforms, according to manufacturer protocols.
Assembly of raw reads was performed using Spades (Bankevich et al., 2012), as part of the
Shovill pipeline (version 1.0.1; https://github.com/tseemann/shovill) using the following settings:
"--minlen 200 --mincov 2; --assembler spades; --trim".
3.3.4 Analysis of whole-genome assemblies
MLST and core genome MLST analysis
Prediction of legacy multi-locus sequence types was performed using MLST (version 2.19.0;
https://github.com/tseemann/mlst) according to the Achtman 7-loci scheme for Salmonella
enterica and E. coli (https://pubmlst.org/mlst/). Isolates were also typed using fsac (version
1.2.0; https://github.com/dorbarker/fsac) according to the core-genome multi-locus sequence
typing (cgMLST) schemes available from Enterobase (https://enterobase.warwick.ac.uk/) with
3002- and 2513-loci schemes for Salmonella and E. coli, respectively. Isolates with 25 or more
missing cgMLST loci were considered poor quality and excluded from any further analyses.
86
Visualization of population structure
The standalone GrapeTree software package (version 1.5; Zhou et al., 2018) was used to create
minimum spanning trees to visualize population structure, using the "MSTreeV2" algorithm. A
cluster threshold of 50 allelic differences for visualizations of E. coli. For Salmonella, cluster
thresholds of 30 and 5 allelic differences were used for visualizations of more than 4, and 4 or
fewer serovars, respectively.
In silico prediction of acquired AMR genes
Acquired resistance genes were identified using Abricate (version 0.8.13;
https://github.com/tseemann/abricate) and the Resfinder database (current as of May-17-2020),
with settings of 90% identity and 60% coverage. Identity and coverage settings were increased to
100% and 90% to identify acquired beta-lactamases. The sensitivity and specificity of in silico
AMR prediction were calculated by antimicrobial class and overall (i.e., pooling all individual
test results); phenotypic AMR results were considered the gold standard, and the presence of
genotypic resistance was considered a positive test result. Isolates with intermediate
susceptibility were categorized as susceptible. We elected not to assess test sensitivity or
specificity of drug classes for which chromosomal mutations are known to confer a considerable
proportion of expressed resistance (i.e., quinolones, aminoglycosides; Chen et al., 2007; Vetting
et al., 2003). As a quality control measure, samples with missing genotypes for resistant
phenotypic test results for three or more of the seven antimicrobial classes were examined and
excluded from further analyses if they were also missing greater than 20 loci based on cgMLST.
Plasmid replicon prediction
87
Identification of plasmid replicons was performed using Abricate (version 0.8.13;
https://github.com/tseemann/abricate) and the Plasmidfinder database (current as of May-17-
2020). Settings of 98% identity and 70% coverage were used.
In silico serotyping
Serotyping of E. coli isolates was performed using ECTyper (version 1.0.0, database version 1.0;
https//github.com/phac-nml/ecoli_serotyping) and default settings. Serotyping of Salmonella
isolates was performed using SISTR (version 1.1.1; https://github.com/phac-nml/sistr_cmd), and
default settings with the "centroid" allele database.
3.3.5 Statistical analyses
Univariable multi-level logistic regression was used to model the odds of identifying select
plasmids and antimicrobial resistance genes found in E. coli and Salmonella from different
sources. All statistical analyses were performed using STATA (STATA Intercooled 14.2;
StataCorp, College Station, Texas, USA). Only plasmid types and resistance genes with a
prevalence greater than 10% and less than 90% were modeled. The following independent
variables were examined: year of sampling, farm location (farms 6-10, as in Bondo et al., 2016a),
and source type (i.e., raccoon, swine manure pit, soil). Due to low effective samples sizes,
univariable logistic regression was performed, with a random intercept to account for clustering
of samples obtained from the same raccoon or swine manure pit. For models that did not
converge using the ‘melogit’ command, the model was subsequently fit using the ‘meqrlogit’
command, which uses QR decomposition of the variance-components matrix. Variance
components were used to calculate intraclass correlation coefficients (ICCs) using the latent
88
variable technique (Dohoo et al., 2014). The fit of multi-level models was assessed by examining
the best linear unbiased predictions (BLUPS) for normality and homoscedasticity, and Pearson's
residuals were examined for outliers. If variance components were very small (<1x10-3), the
Bayesian information criterion (BIC) was used to compare the fit of the multi-level logistic
regression model with an ordinary logistic regression; the better fitting model was reported
(Dohoo et al., 2014). If low effective sample sizes posed estimation issues for univariable
models, exact logistic regression was used, and the score method was used to calculate p-values
for these models. A significance level of a=0.05 was used, and all tests were two-tailed.
3.4 Results
3.4.1 Dataset
Salmonella
Based on our study criteria, a total of 159 Salmonella isolates from the following sources were
included: raccoon (n=92), soil (n=46), and swine manure pit (n=21). Most of these isolates were
obtained from samples collected in 2012 (n=82, 52%) and in 2011 (n=50, 31%), with fewer
isolates in 2013 (n=27, 17%). Isolates originated from 80 unique raccoons; among animals
captured multiple times, eight individuals contributed two isolates, and two raccoons contributed
three isolates from different trapping dates. The majority of Salmonella isolates were
phenotypically pan-susceptible: 96.7% of raccoon isolates (95%CI: 90.8-99.3%), 95.6% of soil
isolates (95%CI: 85.2-99.5), and 95.3% of swine manure pit isolates (95%CI: 76.2-99.9%). Six
of the 159 isolates demonstrated phenotypic resistance (Table 3.1), and the overall prevalence of
89
multiclass resistance (3+ drug classes) was 1.9% (n=3/159, 95%CI: 0.4-5.4%). These multiclass
resistant isolates were identified in two raccoon samples and one manure sample (Table 3.1).
E. coli
A total of 96 resistant E. coli isolates were included, with the following source distribution:
raccoon (n=20), soil (n=45), and swine manure pit (n=31). Most resistant isolates were obtained
from samples collected in 2013 (n=39, 41%), followed by 2011 (n=37, 39%), and 2012 (n=20,
21%). Resistant raccoon isolates were obtained from 20 unique individuals, with no repeated
sampling. Overall, 26.0% of resistant isolates were multiclass resistant (3+ drug classes) based
on phenotype (n=25/96), with most of these isolates identified in soil samples (n=13), followed
by raccoon samples (n=7), and swine manure pit samples (n=5). The corresponding prevalence
of multiclass resistance was highest among resistant raccoon isolates (35.0%, 95%CI: 15.4-
59.2%) and resistant soil isolates (28.9%, 95%CI: 16.4-44.3%), and lowest in resistant swine
manure pit isolates (16.1%, 95%CI: 5.4-33.7%).
3.4.2 Distribution of serovars and MLST types
Salmonella
In total, 21 sequence types representing 21 different serovars were identified (Figure 3.1A, Table
3.2). Three isolates were not typeable with MLST (two raccoon, one soil). The four most
common serovars identified among all source types, in descending order, were Newport (28%),
Agona (18%), Infantis (11%), and Typhimurium (9%; Figure 3.1A). A number of international
lineages were also identified (Figure 3.1B); seventeen ST32 (Infantis) isolates were identified on
all farms from all source types (i.e., manure, raccoon and soil), and, apart from one sample, all
90
were collected in 2011 and 2012. A total of fourteen ST19 (Typhimurium) isolates were
identified in all sources on four of five farms; nine of these samples (n=9/14; 64%) were
obtained from the same farm (farm 6) in 2012. Finally, one isolate each of ST96
(Schwarzengrund), ST15 (Heidelberg), and ST65 (Brandenburg) were isolated from two
raccoons and one manure sample, respectively (Figure 3.1B).
E. coli
This population of resistant E. coli was comprised of 49 sequence types, of which two samples
were not typeable with MLST (two manure isolates; Table 3.3). Serovars identified are available
in Table 3.4.
3.4.3 Population structure based on cgMLST
Salmonella
The following Salmonella serovars were identified in both swine manure pit samples and
raccoon samples: Agona, Infantis, Poona, Typhimurium (Figure 3.2). Identical or nearly identical
cgMLST subtypes were identified from all sources for both Agona and Poona serovars. The 29
Agona isolates had between 0 and 12 allelic differences; isolates were identified in 2011 and
2012 on three of five farms. Four Poona isolates that differed by a maximum of 2 loci were
isolated on farm 8 in 2013 from all sources. Salmonella Infantis and Typhimurium isolates
clustered into single groups at thresholds of 51 and 329 allelic differences, respectively. Less
commonly identified serovars that were isolated from both raccoon and soil samples differed by
a variety of minimum and maximum allelic differences between the two sources: Hadar (39
allelic differences), Enteritidis (270 allelic differences), Hartford (1-15 allelic differences),
91
Thompson (0-36 allelic differences), Paratyphi B var. Java (0-28 allelic differences), and
Newport (0-68 allelic differences). Hadar and Hartford isolates were isolated from the same farm
in different months of the same year. The majority of Thompson isolates were obtained from
farm 8 (n=8/10), and most were collected in 2012 (n=7/10).
E. coli
The population structure of E. coli based on cgMLST is presented in Figure 3.3. Similar or
identical subtypes were identified from all sources, regardless of farm location, year of sampling,
or the presence of phenotypic multiclass antimicrobial resistance.
3.4.4 In silico determination of acquired AMR genes and plasmid replicons
Salmonella
Eighteen different plasmids types and nine different antimicrobial resistance genes were
identified in this Salmonella population (Tables 3.1, 3.5). AMR genes identified were aadA2,
aadA4, aph(3'')-Ib, aph(6)-Id, fosA7, sul1, tet(A), tet(B), and blaTEM-1. Gene fosA7 was only
identified in phenotypically susceptible isolates, with an overall prevalence of 19.5% (n=31/159).
All six phenotypically resistant Salmonella isolates were isolated in 2013 (Table 3.1). A manure
pit isolate and a raccoon isolate with identical phenotypic resistance patterns (SOX-STR-TCY)
previously identified as ST19 (Typhimurium) DT104 by Bondo et al., (2016b) contained
identical resistance genes, and some of the same plasmid types. These two DT104 isolates
differed at only 12 cgMLST loci, and both were isolated in July 2013, but on different farms. A
single ST96 (Schwarzengrund) isolate with SOX-STR-TCY from a raccoon contained resistance
genes sul1, tet(B), and aadA4, and several plasmid types. Two ST33 (Hadar) isolates, one from a
92
raccoon and another from soil displayed identical phenotypic resistance (STR-TCY) mediated by
the same genes (tet[A], aph[6]-Id), but carried different plasmid replicons. These two Hadar
isolates differed at 39 cgMLST loci, and both were isolated in different months from farm 6. The
remaining resistant isolate was a Salmonella Kiambu isolate from soil with AMP-TCY resistance
conferred by blaTEM-1 and tet(A), and also contained an IncX1 replicon.
E. coli
A total of 27 resistance genes and 21 plasmid types were identified among resistant E. coli
isolates (Tables 3.5, 3.6). The distribution of resistance genes among different source types is
presented in Table 3.6. The majority of genes identified confer resistance to aminoglycosides,
tetracyclines, and folate pathway inhibitors. Genes conferring resistance to phenicols were
uncommonly identified, and no macrolide resistance genes were identified (Table 3.6). Besides
blaTEM-1 (26.0% prevalence), only one other type of beta-lactamase, blaCMY-2, was identified, and
occurred in a single raccoon isolate (Table 3.6). The raccoon isolate containing the sole blaCMY-2
displayed phenotypic resistance to five of seven drug classes examined (AMC-AMP-FOX-TIO-
CRO-CHL-STR-SOX-TCY-SXT), and also contained genes aadA2, sul2 and dfrA12, as well as
a single plasmid type (IncC). Two additional isolates were phenotypically resistant to five of the
seven drug classes examined: a soil isolate (AMP-CHL-KAN-STR-SOX-TCY-SXT), and a
manure isolate (AMP-CHL-STR-SOX-TCY-SXT). Despite having nearly identical phenotypic
resistance patterns, these soil and manure isolates represented different sequence types,
contained different plasmid replicons, and, apart from the presence of tet(A) and blaTEM-1,
contained different genes responsible for conferring resistance.
93
3.4.5 Comparison of Salmonella and E. coli
Among plasmid types commonly identified in this population of Salmonella and E. coli, few
incompatibility types were identified in both organisms (Table 3.5). The majority of plasmid
types were restricted to either Salmonella or E. coli, but not found in both (e.g.,
IncFIB[AP001918] in E. coli, IncFiip96a in Salmonella). We also evaluated whether resistance
genes may be shared between E. coli and Salmonella isolates within the same animal; of the
three raccoon samples positive for resistant Salmonella (Table 3.1), no corresponding resistant E.
coli were isolated from the same animal during the study period, either on the same capture date
or on another capture date. Along with a single resistant Salmonella manure sample originating
from farm 9, three resistant E. coli manure isolates were obtained from the same farm, with one
collected in the same year; however, apart from two genes in common between two of the E. coli
isolates and the Salmonella isolate (i.e., sul1, tet[A]), there was no overlap with regards to
resistance genes, resistance patterns or plasmid types between these resistant E. coli isolates and
the resistant Salmonella isolate.
3.4.6 Sensitivity and specificity of in silico AMR prediction
Test sensitivity could not be assessed for certain drug classes where no phenotypic resistance
was identified (e.g., macrolides, phenicols; Table 3.7). Test sensitivity and specificity were 89%
or greater for all drug classes in both Salmonella and E. coli. The overall test sensitivity and
specificity were 97% or greater for both organisms.
94
3.4.7 Statistical results
Salmonella
All five plasmid types and the one resistance gene analyzed (fosA7) were significantly associated
with at least one independent variable (Table 3.8). Among the four plasmid types associated with
source type (i.e., IncX1, IncFIIS, IncX3, IncFiip96a), the odds of identifying plasmids was
consistently greater in raccoons compared to swine manure samples, and in some cases, the odds
were also greater in soil samples compared to swine manure (i.e., IncFIIS, IncFiip96a; Table
3.8). Two plasmid types (i.e., IncX3, Colye4449) and fosA7 were significantly associated with
farm location. Colye4449 was the only outcome associated with year of sampling and was
significantly lower in 2013 compared to 2011 and 2012 (Table 3.8). The random intercept was
not retained in any model with a statistically significant fixed effect, since it did not improve the
fit of the model, the variance component was negligible (<1x10-3), and/or coefficients could not
be estimated without exact logistic regression. Contrasts from logistic regression analyses are
available in Table 3.9.
E. coli
Of the eight plasmid types and four resistance genes examined statistically, only two plasmid
types (i.e., IncI1[alpha], IncFIB[AP001918]) and one gene (sul2) were significantly associated
with source or farm location, and none were associated with the year of sampling (Table 3.10).
IncI1(alpha) was significantly associated with farm location, whereas sul2 and
IncFIB(AP001918) were associated with source type, and both were higher in raccoons
compared to swine manure samples. Although model assumptions were met for sul2 and
95
IncFIB(AP001918) models, the random intercept was not retained since the variance components
were negligible (<1x10-3), and the BIC favoured models without the random intercept.
3.5 Discussion
Examination of population structure of the Salmonella and E. coli isolates in this study provides
evidence suggesting local transmission of these organisms between wildlife, livestock, and
environmental sources on swine farms or exposure to a common source within this region of
southern Ontario. Salmonella serovars with a broad-host affinity such as Typhimurium and
Infantis (Forshell & Wierup, 2006) were identified in all sampling sources (i.e., raccoons, swine
manure pits, soil); on average, these serovars displayed greater diversity in cgMLST profiles
(>50 allelic differences) than serovars such as Poona and Agona, which were also identified in
all sampling sources, and typically had very similar cgMLST profiles (differed at <15 loci).
Additionally, we identified “clusters” of certain Salmonella serovars that were found in multiple
sources but restricted to certain years and farms (Salmonella Poona on farm 8 in 2013;
Salmonella Typhimurium ST19 on farm 6 in 2012; Salmonella Thompson on farm 8), albeit
these observations are based on small sample sizes (n<20) and differences could not be tested
statistically. All sources were found to contain at least one or more internationally important
Salmonella sequence types (e.g., ST19 Typhimurium, ST32 Infantis), but, overall, very few
isolates exhibited phenotypic AMR (<5%), as previously observed by Bondo et al. (2016b). As a
result of this low prevalence of AMR among Salmonella, and apart from one gene (fosA7), we
were unable to examine patterns in the distribution of resistance genes or assess risk factors
statistically. Conversely, our inclusion of untyped E. coli based on demonstrated phenotypic
AMR unquestionably resulted in a sampling bias (Lucassen et al., 2019), but this approach
96
enabled statistical assessments which contribute to a preliminary understanding of the dynamics
and movement of AMR in enteric bacterial populations in these different sources.
Similar to other studies (Prasertsee et al., 2019; Pornsukarom et al., 2018), the use of in
silico tools for the identification of resistance genes in this study was generally reliable, although
we did not assess drug classes for which chromosomal resistance plays an important role (i.e.,
quinolones, aminoglycosides). For the purposes of validating phenotypic results and
investigating possible transmission of resistance genes between different sources, these tools
have provided insights about the movement of AMR determinants in a southern Ontario
agroecosystem. In some cases, resistant Salmonella with identical phenotypic AMR patterns
contained different genes responsible for conferring resistance. Our findings of similar cgMLST
profiles (<40 allelic differences), along with the presence of identical resistance genes and
different plasmid replicons among isolates that were spatially or temporally linked were highly
suggestive of clonal dissemination. Such cases included two Salmonella ST33 Hadar (one soil,
one raccoon) from the same farm in the same year, and two Salmonella DT104 ST19
Typhimurium (one raccoon, one swine manure pit) from different farms in the same year; these
findings highlight the potential occurrence of on-farm as well as between-farm transmission of
Salmonella between different sources. Similar or identical cgMLST profiles were also frequently
identified among raccoon and soil isolates for a variety of serovars, including, but not limited to,
the following: Newport, Agona, Thompson, Hartford, Paratyphi B var. Java. Examination of
these findings using approaches with greater discriminatory power (e.g., SNPs) is needed,
however, in conjunction with previous work (Bondo et al., 2016c), these particular findings
97
suggest that mechanical transmission between raccoons and their immediate environment is
frequently occurring.
Among E. coli isolates, only a handful of extensively drug resistant isolates were
identified. One such E. coli isolate identified in a raccoon (AMC-AMP-FOX-TIO-CRO-CHL-
STR-SOX-TCY-SXT) contained many resistance genes, including a blaCMY-2, and an IncC
plasmid replicon. In general, the types of resistance identified in these populations of Salmonella
and E. coli conferred resistance to aminoglycosides, tetracyclines, and folate pathway inhibitors;
other more concerning types of resistance to macrolides, phenicols, and fluoroquinolones were
rarely identified, or absent altogether in this study. In spite of the low prevalence of AMR among
Salmonella in our study, the top serovars identified here (Newport, Agona, Infantis,
Typhimurium, Paratyphi B var. Java) have also appeared among the top five serovars responsible
for human Salmonella infections in the same geographic region (Enteritidis, Typhimurium,
Heidelberg, Newport, Thompson; Flockhart et al., 2017). Interestingly, Flockhart et al. (2017)
observed that serovars Newport and Thompson were infrequently isolated from agricultural
sources, in contrast with other serovars that were commonly isolated from poultry and cattle in
the region (e.g., Enteritidis, Typhimurium, Heidelberg). Together, Newport and Thompson were
previously implicated in only 6.2% (n=29/465) of reported human Salmonella cases in our study
region (Flockhart et al., 2017), but their near absence in a collective 3214 samples from beef,
broilers, dairy, swine and water sources in Flockhart et al. (2017), and their common occurrence
in soil and raccoons in the present study suggests a potential role for raccoons in the
epidemiology of human infections with uncommon Salmonella serovars.
98
Not all plasmid replicons found in Salmonella were associated with resistance genes; in
fact, two plasmid types were present in almost half of all Salmonella isolates, but with only six
resistant Salmonella isolates in total, it is apparent that plasmids are freely circulating among
susceptible isolates. The presence of virulence genes (not evaluated here) may have contributed
to the persistence and carriage of these plasmids among susceptible Salmonella isolates (Pilla &
Tang et al., 2018). The types of plasmids appearing in Salmonella were very distinct from those
identified in the resistant E. coli population; few plasmid types were commonly identified in both
organisms. Previous work by Varga et al. (2008) demonstrated a lack of association between
phenotypic resistance patterns for Salmonella and generic E. coli isolates originating from the
same swine manure sample. We made a similar observation based on our examination of a small
number of raccoons that carried both resistant Salmonella and resistant E. coli; we identified no
similarities among AMR determinants between organisms originating from the raccoon, manure
pit, or soil sample. However, examination of only one respective isolate per fecal sample may
not capture or represent important and relevant aspects of the microbiome and related resistome
(Penders et al., 2013).
Major differences in the distribution of AMR genes and plasmid replicons between E.
coli and Salmonella were identified. The majority of AMR genes and plasmid replicons in E. coli
analyzed statistically were not significantly associated with any of the predictor variables,
whereas all of the plasmids and the single gene analyzed in Salmonella (i.e., fosA7) were
associated with source type or farm location, or both. The lack of association for most of the
AMR determinants in E. coli with source, farm, or year of sampling suggests widespread sharing
of E. coli AMR determinants between sources in this region. However, these findings should be
99
interpreted in light of our selection of a population of resistant E. coli, and future examination of
both susceptible and resistant isolates will provide important context for these findings. For those
plasmid types and genes associated with farm location, the farm with the highest odds of these
outcomes was not consistent and varied depending on the particular plasmid replicon or gene
under examination. In contrast, plasmid replicons and genes associated with source type were
consistently more likely to be identified in raccoons compared to swine manure pit samples, and,
for certain outcomes, they were also more likely to be isolated from soil samples compared to
swine manure. These findings suggest that, for certain AMR determinants, limited exchange
between raccoons and soil with swine manure pits is occurring, similar to findings of previous
work in this study area examining Campylobacter isolates from raccoons and livestock (swine,
dairy, beef; Viswanathan et al., 2017). Carriage of a moderate prevalence of certain AMR
determinants by raccoons in this region suggests that they may play a role in disseminating
resistance genes, but their importance in the movement of plasmids not associated with AMR
genes (e.g., IncFiip96a in Salmonella) requires further study. Unfortunately, we did not
characterize plasmid mobility, so it is unknown whether these plasmids have the potential to be
an important source of antimicrobial resistance genes. However, mobility is likely based on the
appearance of most non-AMR associated plasmids in multiple serovars.
The identification of sequences types of international importance in all source types
sampled in swine farm environments, in particular ST19 (the most prevalent sequence type
among Typhimurium isolates globally [Achtman et al., 2012]) suggests widespread circulation of
these strains in this region of Ontario. To date, there are few studies that have examined
subtyping data and gene-level antimicrobial resistance data in raccoons (Bondo et al., 2016a;
100
Bondo et al., 2016b; Very et al., 2016; Compton et al., 2008); this study contributes new data to
this literature regarding common serovars, sequence types, microbial population structure,
resistance genes, and plasmid replicons carried by a rural raccoon population. In wild birds,
where genomic investigations are becoming increasingly common, Enterobacteriaceae
containing resistance genes to high-priority antimicrobials (e.g., blaCTX-M-65, blaIMP-4, mcr-1), and
international clones have been identified (Wang et al., 2020; Fuentes-Castillo et al., 2019; Wang
et al., 2017; Dolejska et al., 2016). The types of resistance identified in raccoons and other
sources on swine farms in our study mirror that found in swine in other parts of the world (e.g.,
sulfonamides, aminoglycosides, tetracyclines; Nguyen Thi et al., 2020; Joaquim et al., 2021;
Ikwap et al., 2021). Our findings of widespread tetracycline resistance genes (i.e., tetA, tetB) that
were not associated with particular sources, locations, or years are plausibly driven by the swine
farm environment, since tetracyclines, among other antimicrobials, are frequently used in the
swine industry (Abubakar et al., 2019). These findings are also consistent with recent mounting
evidence that the use of antimicrobials in agriculture is a major driver of resistance in intensive
farm environments (Turcotte et al., 2020; Abubakar et al., 2019; Mencía-Ares et al., 2010). With
the increasing application of genomics to investigations involving raccoons and other wildlife
species, we suspect that international clones and genes conferring resistance to critically
important antimicrobials will be identified in these wildlife populations. In the future,
exploration of enteric bacteria and AMR carried by raccoons in different types of environments,
including cities, may contribute valuable information to this field of work.
Limitations
101
The low overall prevalence of resistance among Salmonella isolates on swine farms presented an
obvious challenge for the study of the movement of AMR determinants, and, in many cases,
precluded statistical assessments of these data. Our low effective sample sizes did not permit
multivariable modelling, or the ability to account for potential confounding by serovar (due to
the sheer number of serovars identified, models could not converge). Unfortunately, our
univariable analyses do not adequately capture the complexities of AMR in the ecosystem, but
they represent an important first step in this process. Similarly, our analysis of plasmids should
be interpreted with caution since we did not reconstruct plasmids or characterize mobility, thus,
future work would be strengthened with confirmation of these aspects of plasmid biology
(Robertson et al., 2020). Identification of the precise location of resistance genes either on
plasmids or within chromosomes in future work will provide further insights about AMR
transmission and movement. As previously alluded to, the true relationships between plasmids
and the risk factors examined here are potentially obscured by our inclusion of only
phenotypically resistant E. coli. Moreover, inclusion of only one E. coli isolate per fecal sample
limits our understanding of the transmission of AMR determinants within the greater gut
microbiome (Penders et al., 2013), and of the true population structure of E. coli. The lack of
similarities between Salmonella and E. coli obtained from the same animal therefore do not
constitute definitive evidence that AMR gene transmission is not commonly occurring in the
gastrointestinal tract of raccoons. Finally, the measures of association reported in our analyses do
not account for the initial probability of isolating the organism in the sample. To illustrate this
point, consider that the prevalence of resistant E. coli was highest in raccoons, and lowest in
swine manure samples (35% vs 16%); the odds of certain plasmids were higher in raccoons
102
compared to swine manure samples, but these odds were always relative to samples that already
contained a resistant E. coli isolate.
3.6 Conclusions
A diversity of Salmonella serovars were isolated from the raccoon population in this longitudinal
study, some of which have been implicated in human clinical cases in the study region (e.g.,
Thompson, Newport). Findings from this work suggest that local transmission of Salmonella, E.
coli and related AMR determinants is likely occurring between raccoons and environmental
sources (i.e., soil, swine manure pits), both locally on farms, and between farms in the region.
Several plasmid types and resistance genes analyzed statistically were associated with source
type or farm location; among source type associations, these outcomes were consistently higher
in raccoons compared to swine manure pits, suggesting a lack of exchange between
environmental livestock sources and raccoons for those outcomes. Other resistance genes and
plasmid replicons were not associated with either farm location or source type, suggesting
widespread dissemination of those AMR determinants in this region, regardless of source. The
majority of plasmid replicons in Salmonella were not associated with AMR, thus, the importance
of these plasmid types commonly occurring in pan-susceptible Salmonella requires further
examination, with consideration of virulence factors. More comprehensive sampling of farm
environments, and additional environmental sources, as well as a thorough examination of
plasmids in resistant, as well as susceptible, isolates is warranted.
103
3.7 References
Abubakar, R. H., Madoroba, E., Adebowale, O., Fasanmi, O. G., & Fasina, F. O. (2019). Antimicrobial usage in pig production: Effects on Escherichia coli virulence profiles and antimicrobial resistance. The Onderstepoort Journal of Veterinary Research, 86(1), e1–e11.
Achtman, M., Wain, J., Weill, F. X., Nair, S., Zhou, Z., Sangal, V., Krauland, M. G., Hale, J. L., Harbottle, H., Uesbeck, A., Dougan, G., Harrison, L. H., Brisse, S., & S. Enterica MLST Study Group. (2012). Multilocus sequence typing as a replacement for serotyping in Salmonella enterica. PLoS Pathogens, 8(6), e1002776.
Ahlstrom, C. A., Bonnedahl, J., Woksepp, H., Hernandez, J., Reed, J. A., Tibbitts, L., Olsen, B., Douglas, D. C., & Ramey, A. M. (2019). Satellite tracking of gulls and genomic characterization of faecal bacteria reveals environmentally mediated acquisition and dispersal of antimicrobial-resistant Escherichia coli on the Kenai Peninsula, Alaska. Molecular Ecology, 28(10), 2531–2545.
Ahlstrom, C. A., Bonnedahl, J., Woksepp, H., Hernandez, J., Olsen, B., & Ramey, A. M. (2018). Acquisition and dissemination of cephalosporin-resistant E. coli in migratory birds sampled at an Alaska landfill as inferred through genomic analysis. Scientific Reports, 8(1), 7361.
Anyanwu, M. U., Jaja, I. F., & Nwobi, O. C. (2020). Occurrence and characteristics of mobile colistin resistance (mcr) gene-containing isolates from the environment: A review. International Journal of Environmental Research and Public Health, 17(3), 1028.
Arnold, K. E., Williams, N. J., & Bennett, M. (2016). 'Disperse abroad in the land': The role of wildlife in the dissemination of antimicrobial resistance. Biology Letters, 12(8), 20160137.
Bankevich, A., Nurk, S., Antipov, D., Gurevich, A. A., Dvorkin, M., Kulikov, A. S., Lesin, V. M., Nikolenko, S. I., Pham, S., Prjibelski, A. D., Pyshkin, A. V., Sirotkin, A. V., Vyahhi, N., Tesler, G., Alekseyev, M. A., & Pevzner, P. A. (2012). SPAdes: A new genome assembly algorithm and its applications to single-cell sequencing. Journal of Computational Biology, 19(5), 455–477.
Bloomfield, S. J., Benschop, J., Biggs, P. J., Marshall, J. C., Hayman, D., Carter, P. E., Midwinter, A. C., Mather, A. E., & French, N. P. (2017). Genomic analysis of Salmonella enterica serovar Typhimurium DT160 associated with a 14-Year outbreak, New Zealand, 1998-2012. Emerging Infectious Diseases, 23(6), 906–913.
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016a). Epidemiology of antimicrobial resistance in Escherichia coli isolates from raccoons (Procyon lotor) and the environment on swine farms and conservation areas in southern Ontario. PLoS ONE, 11(11), e0165303.
104
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016b). Impact of season, demographic and environmental factors on Salmonella occurrence in raccoons (Procyon lotor) from swine farms and conservation areas in southern Ontario. PLoS ONE, e0161497.
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016c). Epidemiology of Salmonella on the paws and in the faeces of free-ranging raccoons (Procyon Lotor) in southern Ontario, Canada. Zoonoses and Public Health, 63(4), 303–310.
Bondo, K. J., Pearl, D. L., Janecko, N., Reid-Smith, R. J., Parmley, E. J., Weese, J. S., Rousseau, J., Taboada, E., Mutschall, S., & Jardine, C. M. (2019). Salmonella, Campylobacter, Clostridium difficile, and anti-microbial resistant Escherichia coli in the faeces of sympatric meso-mammals in southern Ontario, Canada. Zoonoses and Public Health, 66(4), 406–416.
Chen, S., Cui, S., McDermott, P. F., Zhao, S., White, D. G., Paulsen, I., & Meng, J. (2007). Contribution of target gene mutations and efflux to decreased susceptibility of Salmonella enterica serovar Typhimurium to fluoroquinolones and other antimicrobials. Antimicrobial Agents and Chemotherapy, 51(2), 535–542.
Collineau, L., Boerlin, P., Carson, C. A., Chapman, B., Fazil, A., Hetman, B., McEwen, S. A., Parmley, E. J., Reid-Smith, R. J., Taboada, E. N., & Smith, B. A. (2019). Integrating whole-genome sequencing data into quantitative risk assessment of foodborne antimicrobial resistance: A review of opportunities and challenges. Frontiers in Microbiology, 10, 1107.
Compton, J. A., Baney, J. A., Donaldson, S. C., Houser, B. A., San Julian, G. J., Yahner, R. H., Chmielecki, W., Reynolds, S., & Jayarao, B. M. (2008). Salmonella infections in the common raccoon (Procyon lotor) in western Pennsylvania. Journal of Clinical Microbiology, 46(9), 3084–3086.
Croucher, N. J., Harris, S. R., Grad, Y. H., & Hanage, W. P. (2013). Bacterial genomes in epidemiology--present and future. Philosophical Transactions of the Royal Society of London. Series B, Biological sciences, 368(1614), 20120202.
De Oliveira, D., Forde, B. M., Kidd, T. J., Harris, P., Schembri, M. A., Beatson, S. A., Paterson, D. L., & Walker, M. J. (2020). Antimicrobial resistance in ESKAPE pathogens. Clinical Microbiology Reviews, 33(3), e00181-19.
Dohoo, I, Martin, W & Stryhn, H. (2014). Veterinary epidemiologic research, 3rd ed. Charlottetown, Prince Edward Island: VER Inc.
Dolejska, M., Masarikova, M., Dobiasova, H., Jamborova, I., Karpiskova, R., Havlicek, M., Carlile, N., Priddel, D., Cizek, A., & Literak, I. (2016). High prevalence of Salmonella and IMP-4-producing Enterobacteriaceae in the silver gull on Five Islands, Australia. The Journal of Antimicrobial Chemotherapy, 71(1), 63–70.
105
Fischer, J., Schmoger, S., Jahn, S., Helmuth, R., & Guerra, B. (2013). NDM-1 carbapenemase-producing Salmonella enterica subsp. enterica serovar Corvallis isolated from a wild bird in Germany. The Journal of Antimicrobial Chemotherapy, 68(12), 2954–2956.
Flockhart, L., Pintar, K., Cook, A., McEwen, S., Friendship, R., Kelton, D., & Pollari, F. (2017). Distribution of Salmonella in humans, production animal operations and a watershed in a FoodNet Canada sentinel site. Zoonoses and Public Health, 64(1), 41–52.
Forshell, L. P., & Wierup, M. (2006). Salmonella contamination: A significant challenge to the global marketing of animal food products. Revue Scientifique et Technique (International Office of Epizootics), 25(2), 541–554.
French, S. K., Pearl, D. L., Peregrine, A. S., & Jardine, C. M. (2019). Baylisascaris procyonis infection in raccoons: A review of demographic and environmental factors influencing parasite carriage. Veterinary Parasitology, Regional Studies and Reports, 16, 100275.
Fuentes-Castillo, D., Farfán-López, M., Esposito, F., Moura, Q., Fernandes, M. R., Lopes, R., Cardoso, B., Muñoz, M. E., Cerdeira, L., Najle, I., Muñoz, P. M., Catão-Dias, J. L., González-Acuña, D., & Lincopan, N. (2019). Wild owls colonized by international clones of extended-spectrum β-lactamase (CTX-M)-producing Escherichia coli and Salmonella Infantis in the southern cone of America. The Science of the Total Environment, 674, 554–562.
Graham, D. W., Bergeron, G., Bourassa, M. W., Dickson, J., Gomes, F., Howe, A., Kahn, L. H., Morley, P. S., Scott, H. M., Simjee, S., Singer, R. S., Smith, T. C., Storrs, C., & Wittum, T. E. (2019). Complexities in understanding antimicrobial resistance across domesticated animal, human, and environmental systems. Annals of the New York Academy of Sciences, 1441(1), 17–30.
Greig, J., Rajić, A., Young, I., Mascarenhas, M., Waddell, L., & LeJeune, J. (2015). A scoping review of the role of wildlife in the transmission of bacterial pathogens and antimicrobial resistance to the food Chain. Zoonoses and Public Health, 62(4), 269–284.
Government of Canada (2015). Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS) 2015 annual report. Guelph, ON: Public Health Agency of Canada.
Ikwap, K., Gertzell, E., Hansson, I., Dahlin, L., Selling, K., Magnusson, U., Dione, M., & Jacobson, M. (2021). The presence of antibiotic-resistant Staphylococcus spp. and Escherichia coli in smallholder pig farms in Uganda. BMC Veterinary Research, 17(1), 31.
Iwu, C. D., Korsten, L., & Okoh, A. I. (2020). The incidence of antibiotic resistance within and beyond the agricultural ecosystem: A concern for public health. Microbiology Open, 9(9), e1035.
106
Joaquim, P., Herrera, M., Dupuis, A., & Chacana, P. (2021). Virulence genes and antimicrobial susceptibility in Salmonella enterica serotypes isolated from swine production in Argentina. Revista Argentina de Microbiologia, S0325-7541(20)30097-3.
Lucassen, R., Rehberg, L., Heyden, M., & Bockmühl, D. (2019). Strong correlation of total phenotypic resistance of samples from household environments and the prevalence of class 1 integrons suggests for the use of the relative prevalence of intI1 as a screening tool for multi-resistance. PloS ONE, 14(6), e0218277.
Mather, A. E., Lawson, B., de Pinna, E., Wigley, P., Parkhill, J., Thomson, N. R., Page, A. J., Holmes, M. A., & Paterson, G. K. (2016). Genomic analysis of Salmonella enterica serovar Typhimurium from wild passerines in England and Wales. Applied and Environmental Microbiology, 82(22), 6728–6735.
Mencía-Ares, O., Argüello, H., Puente, H., Gómez-García, M., Álvarez-Ordóñez, A., Manzanilla, E. G., Carvajal, A., & Rubio, P. (2020). Effect of antimicrobial use and production system on Campylobacter spp., Staphylococcus spp. and Salmonella spp. resistance in Spanish swine: A cross-sectional study. Zoonoses and Public Health, 10.1111/zph.12790. Advance online publication.
Nguyen Thi, H., Pham, T. T., Turchi, B., Fratini, F., Ebani, V. V., Cerri, D., & Bertelloni, F. (2020). Characterization of Salmonella spp. isolates from swine: Virulence and antimicrobial resistance. Animals, 10(12), 2418.
Penders, J., Stobberingh, E. E., Savelkoul, P. H., & Wolffs, P. F. (2013). The human microbiome as a reservoir of antimicrobial resistance. Frontiers in Microbiology, 4, 87.
Pilla, G., & Tang, C. M. (2018). Going around in circles: Virulence plasmids in enteric pathogens. Nature Reviews. Microbiology, 16(8), 484–495.
Pornsukarom, S., Van Vliet, A. H. M., & Thakur, S. (2018). Whole genome sequencing analysis of multiple Salmonella serovars provides insights into phylogenetic relatedness, antimicrobial resistance, and virulence markers across humans, food animals and agriculture environmental sources. BMC Genomics, 19(1), 801.
Prasertsee, T., Chuammitri, P., Deeudom, M., Chokesajjawatee, N., Santiyanont, P., Tadee, P., Nuangmek, A., Tadee, P., Sheppard, S. K., Pascoe, B., & Patchanee, P. (2019). Core genome sequence analysis to characterize Salmonella enterica serovar Rissen ST469 from a swine production chain. International Journal of Food Microbiology, 304, 68–74.
Robertson, J., Bessonov, K., Schonfeld, J., & Nash, J. (2020). Universal whole-sequence-based plasmid typing and its utility to prediction of host range and epidemiological surveillance. Microbial Genomics, 6(10), mgen000435.
107
Toro, M., Retamal, P., Ayers, S., Barreto, M., Allard, M., Brown, E. W., & Gonzalez-Escalona, N. (2016). Whole-genome sequencing analysis of Salmonella enterica serovar Enteritidis isolates in Chile provides insights into possible transmission between gulls, poultry, and humans. Applied and Environmental Microbiology, 82(20), 6223–6232.
Torres, R. T., Carvalho, J., Cunha, M. V., Serrano, E., Palmeira, J. D., & Fonseca, C. (2020). Temporal and geographical research trends of antimicrobial resistance in wildlife - A bibliometric analysis. One Health, 11, 100198.
Turcotte, C., Thibodeau, A., Quessy, S., Topp, E., Beauchamp, G., Fravalo, P., Archambault, M., & Gaucher, M. L. (2020). Impacts of short-term antibiotic withdrawal and long-term judicious antibiotic use on resistance gene abundance and cecal microbiota composition on commercial broiler chicken farms in Québec. Frontiers in Veterinary Science, 7, 547181.
Varga, C., Rajić, A., McFall, M. E., Reid-Smith, R. J., Deckert, A. E., Pearl, D. L., Avery, B. P., Checkley, S. L., & McEwen, S. A. (2008). Comparison of antimicrobial resistance in generic Escherichia coli and Salmonella spp. cultured from identical fecal samples in finishing swine. Canadian Journal of Veterinary Research, 72(2), 181–187.
Very, K. J., Kirchner, M. K., Shariat, N., Cottrell, W., Sandt, C. H., Dudley, E. G., Kariyawasam, S., & Jayarao, B. M. (2016). Prevalence and spatial distribution of Salmonella infections in the Pennsylvania raccoon (Procyon lotor). Zoonoses and Public Health, 63(3), 223–233.
Vetting, M., Roderick, S. L., Hegde, S., Magnet, S., & Blanchard, J. S. (2003). What can structure tell us about in vivo function? The case of aminoglycoside-resistance genes. Biochemical Society Transactions, 31(Pt 3), 520–522.
Viswanathan, M., Pearl, D. L., Taboada, E. N., Parmley, E. J., Mutschall, S., & Jardine, C. M. (2017). Molecular and statistical analysis of Campylobacter spp. and antimicrobial-resistant Campylobacter carriage in wildlife and livestock from Ontario farms. Zoonoses and Public Health, 64(3), 194–203.
Wang, X., Zhai, Z., Zhao, X., Zhang, H., Jiang, H., Wang, X., Wang, H., & Chang, W. (2020). Occurrence and characteristics of Escherichia coli mcr-1-like in rabbits in Shandong, China. Veterinary Medicine and Science, 10.1002/vms3.340.
Wang, J., Ma, Z. B., Zeng, Z. L., Yang, X. W., Huang, Y., & Liu, J. H. (2017). The role of wildlife (wild birds) in the global transmission of antimicrobial resistance genes. Zoological Research, 38(2), 55–80.
Zhou, Z., Alikhan, N. F., Sergeant, M. J., Luhmann, N., Vaz, C., Francisco, A. P., Carriço, J. A., & Achtman, M. (2018). GrapeTree: Visualization of core genomic relationships among 100,000 bacterial pathogens. Genome Research, 28(9), 1395–1404.
108
3.8 Tables
Table 3.1: Resistance genes and plasmid replicons identified in silico in antimicrobial resistant Salmonella enterica from raccoons, swine manure pits,
and soil samples on swine farms in southern Ontario, Canada 2011-2013 (n=159)
Sequence type
Serovar Source Resistance
patterna Resistance genes Plasmid replicons Year
Farm location
ST309 Kiambu Soil AMP-TCY blaTEM-1, tet(A) IncX1 2013 6
ST19 Typhimurium Raccoon STR-SOX-TCY aadA2, sul1, tet(A) ColRNAI, IncFIIS, IncFIBS, Col156,
Col440I
2013 7
ST19 Typhimurium Manure STR-SOX-TCY aadA2, sul1, tet(A) ColRNAI, IncFIIS, IncFIBS, Col156
2013 9
ST96 Schwarzengrund Raccoon STR-SOX-TCY aadA4, sul1, tet(B) IncHI2, pkpccav1321, IncHI2A
2013 6
ST33 Hadar Soil STR-TCY aph(3')-Ib, aph(6)-Id, tet(A)
ColRNAI, ColpVC, Col156
2013 6
ST33 Hadar Raccoon STR-TCY aph(3')-Ib, aph(6)-Id, tet(A)
Col440 2013 6
a AMP = ampicillin; SOX = sulfisoxazole; STR = streptomycin; SXT = trimethoprim sulfamethoxazole; TCY = tetracycline.
109
Table 3.2: Frequency of Salmonella enterica legacy multi-locus sequence types of isolates obtained from
raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013
(n=159)
Sequence typea (Serovar)
Source type
Totalb (%)
Raccoon (n=92)
Swine manure pit (n=21)
Soil (n=46)
ST350 (Newport) 31 0 11 42 (26.4%) ST13 (Agona) 16 5 8 29 (18.2%) ST32 (Infantis) 9 3 5 17 (10.7%) ST19 (Typhimurium) 7 1 6 14 (8.8%) ST26 (Thompson) 6 0 5 11 (6.9%) ST404 (Paratyphi B var. Java) 9 0 2 11 (6.9%) ST638 (Livingstone) 0 9 1 10 (6.3%) ST15 (Heidelburg) 1 0 0 1 (0.6%) ST96 (Schwarzengrund) 1 0 0 1 (0.6%) ST65 (Brandenburg) 0 1 0 1 (0.6%)
Bolded STs represent internationally recognized sequence types. a Sequence types (STs) determined using 7-loci Achtman scheme. b Other STs identified within 5 or fewer samples were: ST413 (Mbandaka; n=1), ST2848 (IIIb 11:k:z53; n=2), ST22
(Braenderup; n=1), ST23 (Oranienburg; n=4), ST33 (Hadar; n=2), ST309 (Kiambu; n=1), ST308 (Poona; n=4),
ST405 (Hartford; n=3), ST469 (Rissen; n=1), ST11 (Enteritidis; n=2), ST544 (Molade; n=1). e Three isolates (2 raccoon, 1 soil) were not typeable.
110
Table 3.3: Multi-locus sequence types identified using whole-genome sequencing data from antimicrobial
resistant Escherichia coli isolates obtained from raccoons, swine manure pits, and soil samples on swine farms
in southern Ontario, Canada 2011-2013 (n=96)
Sequence type Count (%)
10 11 (11.5%)
101 9 (8.7%)
58 5 (5.2%)
542 4 (3.9%)
1633 4 (3.9%)
155 3 (2.9%)
34 3 (2.9%)
*Sequence types identified in fewer than 3 samples were: 43, 898, 349, 345, 295, 3714, 7324, 1844, 1721, 641,
1727, 847, 1771, 48, 547, 1406, 106, 973, 871, 218, 117, 1324, 2178, 9982, 1086, 2077, 206, 3531, 388, 716, 6777,
6975, 4085, 2562, 1079, 212, 602, 5614, 4429, 1112, 362, 1304.
111
Table 3.4: Serovars identified using whole-genome sequencing data from antimicrobial resistant Escherichia
coli isolates obtained from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario,
Canada 2011-2013 (n=96)
Serovar Count (%)
O82:H8 6 (6.3%) O6:H10 2 (2.1%) O108:H2 2 (1.9%) O8:H49 2 (1.9%) O8:H25 2 (1.9%) O83:H14 2 (1.9%) O62:H30 2 (1.9%) O86:H51 2 (1.9%) O109:H45 2 (1.9%)
O42/O28:H21 2 (1.9%) O46/O8/O134 2 (1.9%)
O96:H5 2 (1.9%) O166:H15 2 (1.9%)
*Additional serovars not listed here were identified in fewer than two samples.
112
Table 3.5: Frequencies of plasmid replicon types identified using whole-genome sequencing data from
Salmonella enterica and antimicrobial resistant Escherichia coli isolates obtained from raccoons, swine
manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013
Salmonella (n=159) E. coli (n=96) Rep-type Count (%) Count (%) IncFIB(AP001918) 0 (0%) 41 (42.7%) IncI1(alpha) 0 (0%) 12 (12.5%) IncFiip96a 59 (37.1%) 0 (0%) IncFII 0 (0%) 15 (15.6%) IncFIIS 64 (40.2%) 0 (0%) IncFIA 0 (0%) 9 (9.4%) IncY 3 (1.9%) 11 (11.5%) IncX1 27 (17.0%) 7 (7.3%) IncX3 25 (15.7%) 0 (0%) IncQ1 0 (0%) 5 (5.2%) IncI1(I-gamma) 5 (3.1%) 0 (0%) IncR 0 (0%) 8 (8.3%) ColYE4449 28 (17.6%) 0 (0%) ColRNAI 6 (3.8%) 0 (0%) ColpHAD28 7 (4.4%) 0 (0%) p0111 0 (0%) 8 (8.3%)
a Plasmid replicon types identified in fewer than five E. coli samples included: ColBS512 (n=2), ColE10 (n=2),
ColpVC (n=1), IncC (n=2), IncB/O/K/Z (n=1), IncFIA(HI1) (n=3), IncFIB(K) (n=2), IncFIB(pB171) (n=1),
IncFIC(FII) (n=3), IncFII(pHN7A8) (n=1), IncHI2A (n=2), IncHI2 (n=2). b Plasmid replicon types identified in fewer than five Salmonella samples included: ColpVC (n=2), IncFIBS (n=5),
Col440ii (n=3), IncHI2 (n=1), pkpccav1321 (n=1), IncHI2A (n=1), Col156 (n=3), Col440i (n=2), IncFIBphcm2
(n=1).
113
Table 3.6: Frequencies of acquired antimicrobial resistance genes identified using whole-genome sequencing
data from antimicrobial resistant Escherichia coli isolates obtained from raccoons, swine manure pits, and
soil samples on swine farms in southern Ontario, Canada 2011-2013 (n=96)
Antimicrobial group
Resistance gene
Accession no.†
Raccoon(n=20)
Swine manure pit
(n=31) Soil
(n=45) Total (%)
Aminoglycoside aac(3)-IVa NC_009838 1 0 0 1 (1.0%)
aadA2 JQ364967 3 1 1 5 (4.8%)
ant(3'')-Ia X02340 4 7 8 19 (19.8%)
aph(3')-Ia V00359/
EF015636 1 0 2 3 (3.1%)
aph(3')-IIa V00618 0 1 0 1 (1.0%)
aph(3")-Ib AF321551/
AF024602 11 13 20 44 (45.8%)
aph(6)-Ic X01702 0 1 0 1 (1.0%)
aph(6)-Id M28829 11 13 20 44 (45.8%) Beta-lactam blaCMY-2 X91840 1 0 0 1 (1.0%) blaTEM-1 AY458016/
HM749966/FJ560503
7 7 11 25 (26.0%)
Lincosamide lnuC AY928180 0 0 1 1 (1.0%) lnuF EU118119 0 0 1 1 (1.0%) Folate pathway inhibitors
dfrA1 AF203818/ X00926
1 1 4 6 (6.2%)
dfrA5 X12868 0 1 3 4 (4.2%) drfA12 AM040708 2 0 0 2 (2.0%) dfrA14 DQ388123 2 0 1 3 (3.1%) dfrA23 AJ746361 1 0 0 1 (1.0%) sul1 EU780013 5 2 4 11 (11.5%) sul2 HQ840942/
AY034138 6 1 9 16 (16.7%)
sul3 AJ459418 1 1 2 4 (4.2%)
Phenicol floR AF118107 3 1 3 7 (7.3%) cmlA1 M64556 1 1 1 3 (3.1%) Fosfomycin fosA7 LAPJ010000
14 1 0 2 3 (3.1%)
Tetracycline tet(A) AF534183 9 11 25 45 (46.9%) tet(B) AF326777/
AP000342 5 16 15 36 (37.5%)
tet(C) AY046276/AF055345
0 0 2 2 (2.1%)
tet(M) X04388 0 0 1 1 (1.0%) † Values from Resfinder database.
114
Table 3.7: Test sensitivity and specificitya for in silico identification of acquired antimicrobial resistance genes in Salmonella enterica and Escherichia
coli isolates obtained from raccoons, swine manure pits, and soil samples on swine farms in southern Ontario, Canada 2011-2013
Antimicrobial class Salmonella (n=159) Escherichia coli (n=96)
Test sensitivity (95%CI) Test specificity (95%CI) Test sensitivity (95%CI) Test specificity (95%CI)
Beta-lactam 100% (2.5-100%*) 100% (97.7-100%*) 89.65% (72.7-97.8%) 100% (94.6-100%*)
Macrolide —c 100% (97.7-100%*) —c 100% (96.2-100%*)
Sulfonamide 100% (29.2-100%*) 100% (97.7-100%*) 100% (86.8-100%*) 98.57% (92.3-99.9%)
Phenicol —c 100% (97.7-100%*) 100% (66.4-100%*) 100% (95.8-100%*)
Tetracycline 100% (54.1-100%*) 100% (97.6-100%*) 100% (95.5-100%*) 93.33% (68.1-99.8)
Overallb 100% (69.2-100%*) 100% (99.5-100*%) 97.9% (94.1-99.6%) 99.4% (97.9-99.9%)
a Phenotypic antimicrobial resistance test results were considered the gold standard. Detection of 15 antimicrobials performed using the CMV3AGNF panel from
National Antimicrobial Resistance Monitoring System (Sensititre, Thermo Scientific). In silico acquired resistance genes detected using Abricate and the
Resfinder database. b Raw counts for all samples and antimicrobials were pooled together. c Not applicable since no phenotypic resistance was identified.
* One-sided, 97.5% confidence interval.
115
Table 3.8: Univariable logistic regression modelsa,b,c assessing the association between source type, farm location, and year of sampling and the
occurrence of select antimicrobial resistance genes and plasmid replicons in Salmonella enterica isolates from raccoons, swine manure pits, and soil
samples on swine farms in southern Ontario, Canada 2011-2013 (n=159)
IncX1 IncFIIS IncX3
OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Source type
Swine manure pit REF 0.012
(global)a REF <0.001 (global)b REF 0.012
(global)a
Raccoon 9.06* (1.47-∞) 0.012 20.00 (2.57-155.28) 0.004 8.52* (1.37-∞) 0.025
Soil 3.30* (0.42-∞) 0.173 11.72 (1.44-95.33) 0.021 2.54* (0.30-∞) 0.300
Farm
6 REF 0.288 (global)c REF 0.089
(global)c REF <0.001 (global)a
7 6.33 (0.45-99.89) 0.171 30.27 (1.63-562.91) 0.022 9.36* (1.39-∞)
0.011
8 0.73 (0.05-9.83) 0.811 71.63 (2.98-1721.31) 0.008 2.12* (0.22-∞) 0.310
9 0.71 (0.02-22.64) 0.849 7.65 (0.36-163.68) 0.193 1.41* (0.04-∞) 0.415
10 25.34 (0.80-797.52) 0.066 242.30
(4.50-13042.56) 0.007 18.9* (2.80-∞) 0.001
Year
2011 REF 0.182 (global)b REF 0.420
(global)c REF 0.230 (global)c
2012 2.53 (0.87-7.32) 0.086 0.46 (0.14-1.48) 0.190 5.72 (0.55-59.64) 0.145
2013 1.56 (0.38-6.39) 0.533 0.69 (0.15-3.19) 0.641 0.66 (0.05-9.13) 0.755
116
IncFiip96a Colye4449 fosA7
OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Source type
Swine manure pit REF <0.001
(global)a REF 0.758 (global)c REF 0.831
(global)b
Raccoon 26.39* (4.37-∞) <0.001 0.62 (0.17-2.19) 0.457 0.78 (0.25-2.40) 0.663
Soil 13.54 (2.10-∞) 0.003 0.68 (0.17-2.67) 0.580 0.67 (0.19-2.38) 0.539
Farm
6 REF 0.117 (global)c REF <0.001
(global)a REF <0.001 (global)b
7 95.69 (1.58-5800.67) 0.029 13.40* (2.06-∞) 0.003 10.73 (1.31-87.67) 0.027
8 325.49 (3.35-31575.21) 0.013 1.00 (0-∞) NE 0.52 (0.03-8.75) 0.652
9 20.59 (0.31-1363.38) 0.157 40.00* (5.54-∞) <0.001 32.86 (3.56-303.41) 0.002
10 1735.86 (6.29-478596) 0.009 6.32* (0.81-∞) 0.056 4.79 (0.52-44.20) 0.167
Year
2011 REF 0.319 (global)c REF 0.026
(global)a REF 0.241 (global)c
2012 0.41 (0.10-1.65) 0.209 0.77 (0.30-1.98) 0.661 0.81 (0.33-2.00) 0.644
2013 0.25 (0.03-1.82) 0.173 0.09* (0-0.57) 0.006 0.23 (0.04-1.27) 0.093
REF = referent group, CI = confidence interval, NE = not estimated.
* Median unbiased estimates obtained with exact logistic regression. a Exact logistic regression was used.
117
b Ordinary logistic regression model. c Multi-level model. A random intercept to account for repeated sampling of animals and swine manure pits was retained: IncX1 farm intraclass correlation
coefficient (ICC): 55.7% (95%CI: 8.1-94.7%); IncFIIS farm ICC: 61.9% (95%CI: 24.8-88.9%); IncFIIS year ICC: 53.6% (95%CI: 19.2-84.9%); IncX3 year
ICC: 54.4% (95%CI: 18.2-86.5%); IncFiip96a farm ICC: 70.1% (95%CI: 32.6-91.9%); IncFiip96a year ICC: 64.5% (95%CI: 29.4-88.8%); Colye4449 source
ICC: 6.8% (95%CI: 0.0-99.2%); fosA7 year ICC: 10.3% (95%CI: 0.0-96.1%).
** Contrasts are available in Table 3.9.
118
Table 3.9: Contrasts from univariable logistic regression modelsb,c assessing the association between source
type, farm location, and year of sampling and the occurrence of select antimicrobial resistance genes and
plasmid replicons in Salmonella enterica isolates from raccoons, swine manure pits, and soil samples on swine
farms in southern Ontario, Canada 2011-2013 (n=159)
* Median unbiased estimates.
** None of the source type contrasts for IncFiip96a, IncX1, IncFIIS, or IncX3 were statistically significant.
a The odds of Colye4449 were significantly lower in 2013 compared to 2012 (OR: 0.11*, 95%CI: 0-0.70).
b Exact logistic regression model.
c Ordinary logistic regression model.
IncX3 b Colye4449 a,b fosA7 c
Contrast OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Farm 8 vs. 7 0.25
(0.04-1.06) 0.039
0.04* (0-0.25)
<0.001 0.05
(0.01-0.39) 0.004
Farm 9 vs. 7 0.22
(0.04-1.06) 0.158
3.33 (0.92-12.89)
0.044 3.06
(0.96-9.72) 0.058
Farm 10 vs. 7 2.06
(0.65-6.59) 0.192
0.50 (0.12-1.76)
0.277 0.45
(0.14-1.41) 0.171
Farm 9 vs. 8 0.88
(0.02-11.87) 0.999
75.39* (10.64-∞)
<0.001 62.86
(6.93-570.05) <0.001
Farm 10 vs. 8 8.28
(1.89-51.75) 0.001
11.92* (1.55-∞)
0.007 9.17
(1.01-83.05) 0.049
Farm 10 vs. 9 9.39
(1.13-446.61) 0.034
0.15 (0.03-0.68)
0.008 0.14
(0.04-0.57) 0.006
119
Table 3.10: Univariable logistic regression modelsa,b,c assessing the association between source type, farm location, and year of sampling, and the
occurrence of select antimicrobial resistance genes and plasmid replicons in Escherichia coli isolates obtained from raccoons, swine manure pits, and
soil samples on swine farms in southern Ontario, Canada 2011-2013 (n=96)
tet(A) tet(B) blaTEM-1 sul1
OR (95% CI) p-value OR
(95% CI) p-value OR (95% CI) p-value OR
(95% CI) p-value
Source type
Swine manure
pit REF 0.220
(global)a REF 0.117 (global)a REF 0.594
(global)a REF 0.133
(global)a
Raccoon 1.49
(0.47-4.69) 0.498
0.47
(0.18-1.20) 0.113
1.85
(0.53-6.41) 0.335
4.83
(0.84-27.93) 0.078
Soil 2.27
(0.89-5.83) 0.088
0.31
(0.09-1.07) 0.064
1.11
(0.38-3.27) 0.851
1.41
(0.24-8.24) 0.700
Farm
6 REF 0.798
(global)a REF 0.165
(global)a REF 0.060
(global)a
REF 0.676
(global)a
7 1.98
(0.63- 6.26) 0.245
0.41
(0.12-1.41) 0.157
0.49
(0.10-2.34) 0.372
0.58
(0.04-5.66) 0.660
8 1.56
(0.35-6.94) 0.563
1.30
(0.29-5.76) 0.730
0.90
(0.14-5.66) 0.911
0.56*
(0-5.65) 0.536
9 1.09
(0.30-3.91) 0.896
0.40
(0.10-1.61) 0.197
4.05
(1.02-16.00) 0.046
1.41
(0.16-12.20) 0.999
10 1.41
(0.43-4.68) 0.571
1.43
(0.43-4.69) 0.555
1.44
(0.36-5.67) 0.602
1.11
(0.13-9.37) 0.999
Year
2011 REF 0.077
(global)a REF 0.417
(global)a REF 0.836
(global)a REF 0.688
(global)a
2012 0.32
(0.09-1.05) 0.060
2.08
(0.68-6.35) 0.197
1.16
(0.35-3.84) 0.812
1.99
(0.36-10.98) 0.425
2013 1.10
(0.45-2.72) 0.828
1.17
(0.45-3.02) 0.750
0.81
(0.29-2.29) 0.691
1.67
(0.37-7.53) 0.507
120
sul2 ant(3’’)-Ia aph(3’’)-Ib aph(6)-Id
OR (95% CI) p-value
OR (95% CI) p-value OR
(95% CI) p-value OR
(95% CI) p-value
Source type
Swine manure
pit REF
0.017 (global)a REF 0.876
(global)a REF 0.638
(global)a REF 0.638
(global)a
Raccoon 12.86
(1.41-117.20) 0.024
0.86
(0.21-3.41) 0.827
1.69
(0.54-5.26) 0.363
1.69
(0.54-5.26) 0.363
Soil 7.5
(0.90-62.61) 0.063
0.74
(0.24-2.31) 0.606
1.11
(0.44-2.79) 0.828
1.11
(0.44-2.79) 0.828
Farm
6 REF 0.636
(global)b REF 0.987 (global)a REF
0.283 (global)a REF
0.283 (global)a
7 0.48
(0.02-13.29) 0.666
1.19
(0.28-5.10) 0.817
0.61
(0.19-1.92) 0.399
0.61
(0.19-1.92) 0.399
8 0.22
(0.00-31.73) 0.555
1.19
(0.18-7.84) 0.858
2.14
(0.44-10.39) 0.346
2.14
(0.44-10.39) 0.346
9 0.05
(0.00-4.42) 0.190
1.02
(0.19-5.29) 0.983
0.38
(0.10-1.44) 0.155
0.38
(0.10-1.44) 0.155
10 2.97
(0.10-90.37) 0.532
1.48
(0.34-6.48) 0.599
0.83
(0.25-2.72) 0.763
0.83
(0.25-2.72) 0.763
Year
2011 REF 0.788
(global)a REF 0.922
(global)a REF 0.354
(global)a REF 0.354
(global)a
2012 1.60
(0.38-6.79) 0.524
0.91
(0.23-3.48) 0.886
1.97
(0.65-5.95) 0.230
1.97
(0.65-5.95) 0.230
2013 1.40
(0.40-4.87) 0.597
0.79
(0.25-2.46) 0.688
0.91
(0.37-2.27) 0.845
0.91
(0.37-2.27) 0.845
121
IncFIB(AP001918) IncI1(alpha) IncFII IncY
OR (95% CI) p-value
OR (95% CI) p-value OR
(95% CI) p-value OR
(95% CI) p-value
Source type
Swine manure
pit REF
0.013 (global)a REF 0.907
(global)a REF 0.196
(global)a REF 0.998
(global)b
Wildlife 5.14
(1.50-17.57) 0.009
1.19
(0.24-5.99) 0.832
0.18
(0.02-1.60) 0.124
1.37 (1.8e-
08-1.04e+08) 0.973
Soil 3.27
(1.18-9.14) 0.023
0.84
(0.21-3.43) 0.812
0.63
(0.20-2.03) 0.440
1.14 (1.9e-
08-6.5e+07) 0.989
Farm
6 REF 0.807
(global)b REF 0.028 (global)c REF
0.343 (global)a REF
0.049 (global)a
7 1.95
(0.35-10.81) 0.446
3.97
(0.64-43.91) 0.140
4.19
(0.43-40.62) 0.216
0.92
(0.01-75.19) 0.999
8 2.38
(0.22-25.77) 0.476
0.93
(0.00-12.47) 0.564
9.43
(0.84-105.79) 0.069
2.3*
(0.00-89.70) 0.999
9 1.26
(0.23-7.02) 0.787
2.20
(0.22-29.58) 0.634
4.71
(0.44-49.94) 0.198
6.46
(0.56-348.13) 0.144
10 0.59
(0.09-3.69) 0.574 0.44 (0-5.80) 0.489
5.18
(0.53-50.65) 0.158
6.60
(0.65-339.59) 0.088
Year
2011 REF 0.512
(global)a REF 0.227
(global)b REF 0.865
(global)a REF 0.518
(global)b
2012 1.18
(0.39-3.50) 0.770
0.54
(0.08-3.74) 0.529
1.13
(0.24-5.31) 0.878
0.07
(0.00-29.73) 0.395
2013 0.66
(0.26-1.65) 0.374
0.13
(0.01-1.34) 0.087
1.40
(0.40-4.88) 0.597
0.23
(0.01-4.03) 0.318
REF = referent group, OR = odds ratio.
* Median unbiased estimates obtained with exact logistic regression.
a Ordinary logistic model.
122
b Multi-level model. A random intercept to account for repeated sampling of animals and swine manure pits retained in the following models: sul2 farm intraclass
correlation coefficient (ICC): 74.3% (95%CI: 27.8-95.5%); IncFIB(AP001918) farm ICC: 32.1% (95%CI: 2.9-88.2%); IncI1(alpha) year ICC: 35.2% (95%CI:
2.4-92.1%); IncY source type ICC: 99.8% (95%CI: 98.9-99.9%); IncY year ICC: 99.8% (95%CI: 97.6-99.9%).
c Exact logistic regression model.
** The following contrasts were statistically significant (p<0.05): the odds of IncFIB(AP001918) were lower in swine compared to soil (OR: 0.30, 95%CI: 0.11-
0.85); the odds of IncI1(alpha) were lower on farm 10 versus farm 7 (OR: 0.10*, 95%CI: 0-0.70).
123
3.9 Figures
124
Figure 3.1. Population structure of 159 Salmonella enterica isolates from raccoons, swine manure pits, and soil samples on swine farms in southern
Ontario based on 3002-loci cgMLST scheme from Enterobase. Minimum spanning tree created using k=30 clustering threshold in GrapeTree. (A)
Distribution of 21 serovars determined using SISTR. (B) Distribution of internationally recognized multi-locus sequence types. Three isolates (2
raccoon, 1 soil) were not typeable with MLST. Frequency counts are in square brackets.
125
126
Figure 3.2. Population structure of 64 isolates of Salmonella enterica isolates (serovars Agona, Infantis, Typhimurium, Poona) from raccoons, swine
manure pits, and soil on swine farms in southern Ontario based on 3002-loci cgMLST scheme from Enterobase. Serovars depicted here were those
found to be overlapping in raccoon and swine manure pit sources. Minimum spanning tree created using k=5 clustering threshold in GrapeTree. (A)
Population structure with serovars determined using SISTR. (B) Distribution by source type. (C) Distribution by farm. (D) Distribution by year of
sampling. Frequency counts are in square brackets.
127
128
Figure 3.3. Population structure of 96 antimicrobial resistant Escherichia coli isolates from wildlife, soil, and swine manure on swine farms in southern
Ontario based on 2513-loci cgMLST scheme from Enterobase. Minimum spanning tree created using k=50 clustering threshold in GrapeTree. (A)
Distribution of source types. (B) Distribution by farm location. (C) Distribution by year of sampling. (D) Antimicrobial resistance by number of drug
classes. Frequency counts are in square brackets.
129
CHAPTER 4
4 Genomic epidemiology of antimicrobial resistant Escherichia coli
isolates from wildlife, livestock, and environmental sources in
southern Ontario, and the potential for transmission: A risk
factor analysis
4.1 Abstract
Antimicrobial resistance threatens the health of humans and animals and has repeatedly been
detected in wild animal species across the world. This cross-sectional study integrates whole-
genome sequencing data from Escherichia coli isolates with demonstrated phenotypic resistance
that originated from a previous longitudinal wildlife study in southern Ontario, and from resistant
E. coli water isolates previously collected as part of public health surveillance programs. The
objective of this work was to assess for possible transmission of antimicrobial resistance
determinants between wildlife, water sources, livestock, and the environment on a broad scale,
and at a local scale for the subset of samples collected on both swine farms and conservation
areas in the previous wildlife study. Multi-level logistic regression was used to assess potential
associations between source type, location type (swine farm vs. conservation area), and the
occurrence of select resistance genes and plasmid replicons (determined using in silico tools),
while accounting for clustering by sampling site and animal/dumpster/swine manure pit. In total,
200 samples were included from the following sources: water (n=20), wildlife (n=72), swine
130
manure pit (n=31), and other environmental samples (soil and dumpster samples; n=72). Some
resistance genes were significantly associated with source type, whereas other genes and plasmid
replicons were associated with location type and were significantly more likely to be identified
on swine farms compared to conservation areas. Conversely, internationally distributed sequence
types (e.g., ST131), extended-spectrum beta-lactamase and AmpC-producing E. coli isolated in
lower prevalences (<10%) were almost exclusively isolated from water samples and from
raccoon and soil samples on conservation areas. Differences in the odds of detecting different
resistance genes and plasmids among various sources and location types suggest different
primary sources for particular genes and plasmid types, but, broadly, our findings suggest that
raccoons in this region may be acquiring AMR determinants from water, agriculture, and
anthropogenic sources in conservation areas. Based on our findings, apparently healthy wildlife
may act as sentinels of environmental contamination with antimicrobial resistance and may differ
in their carriage of certain sequence types, genes and plasmid replicons depending on their local
environment.
Keywords
Antimicrobial resistance, Procyon lotor, raccoons, source attribution, whole-genome sequencing
4.2 Introduction
The emergence and persistence of antimicrobial resistance is a major challenge for human health
worldwide (De Oliveira et al., 2020). Antimicrobial resistant infections directly result in
increased morbidity and mortality, and also represent a substantial burden to health care in terms
131
of both cost and efficacy (Cassini et al., 2019; MacKinnon et al., 2020). There is growing
evidence to suggest that wild animals play a role in disseminating determinants of antimicrobial
resistance (AMR) within the environment (Wang et al., 2017; Dolejska & Papagiannitsis, 2018).
Exploratory work, primarily in the form of cross-sectional surveys, has demonstrated that avian
and mammalian wildlife may carry a variety of zoonotic organisms (e.g., Salmonella,
Campylobacter) some of which have been shown to be resistant to antimicrobials considered
critical to human health (Cummins et al., 2020; Boonkusol et al., 2020; Marotta et al., 2020).
Furthermore, bacterial clones of international importance (e.g., E. coli ST131), extended
spectrum beta-lactamase-producers (ESBLs), and organisms resistant to last-resort
antimicrobials (e.g., colistin, vancomycin, carbapenems) mediated by mobile resistance genes
have all been isolated from and documented in wildlife (Tausova et al., 2012; Jamborova et al.,
2018; Veldman et al., 2013; Poirel et al., 2012; Oracová et al., 2018; Báez et al., 2015; Fischer et
al., 2013; Liakopoulos et al., 2016).
Much of the work examining the epidemiology of antimicrobial resistance in wildlife has
been focused on wild birds (Wang et al., 2017; Bonnedahl & Järhult, 2014). However, there is
mounting evidence that mammalian wildlife such as raccoons (Procyon lotor) may also acquire
antimicrobial resistant bacteria and might therefore represent a potential source of resistant
clones or resistance genes for humans and other animals (Bondo et al., 2016a; Bondo et al.,
2016b, Baldi et al., 2019). The use of whole-genome sequencing data in recent epidemiological
investigations examining the role of wildlife in the maintenance and dissemination of
antimicrobial resistance has provided valuable insights; Ahlstrom et al. (2018, 2019) found that
gulls in Alaska acquired resistance genes of public health importance from landfills in that region
and could contribute to the spread of these genes while foraging at distant locations. To address
132
current research gaps concerning antimicrobial resistance and wildlife, highly discriminatory
data, such as whole-genome sequence data, are valuable tools that allow researchers to
simultaneously identify strains, resistance genes, plasmids, and other genetic markers, such as
virulence genes, using a single laboratory processing method (Croucher et al., 2013).
Findings from a previous three-year longitudinal study of raccoons and environmental
samples on swine farms and conservation areas in southern Ontario demonstrated that the overall
prevalence of antimicrobial resistance among untyped Escherichia coli isolates from raccoon
fecal samples did not differ significantly between these two types of environments (Bondo et al.,
2016a). However, a comparison of resistance phenotypes and genotypes (determined using PCR)
of resistant E. coli isolates revealed similar phenotypes and resistance genes among isolates
obtained in conservation areas that were altogether absent from samples obtained on swine farms
(e.g., blaCMY-2; Bondo et al., 2016a), suggesting that there may, indeed, be differences in the
types of antimicrobial resistance carried by raccoons, depending on the local environment. The
aim of this work was to use whole-genome sequencing to characterize in greater detail the
population of phenotypically resistant E. coli isolates collected through the previous longitudinal
wildlife study (Bondo et al., 2016a), along with water samples obtained by routine public health
surveillance in the same study region and time period, to better understand the potential role of
raccoons in the ecology of antimicrobial resistance. More specifically, our objectives were to: 1)
assess for possible transmission of antimicrobial resistance determinants between water, wildlife,
livestock, and the environment at a broad scale; and 2) assess for possible transmission of
antimicrobial resistance determinants at a local scale for the subset of wildlife and environmental
samples replicated on both swine farms and conservation areas. Transmission between sampling
sources was assessed using a combination of population structure assessments (based on core-
133
genome multi-locus sequence typing using Enterobase's 2513-loci scheme;
https://enterobase.warwick.ac.uk), epidemiologically-informed statistical modeling of select
AMR determinants (i.e., genes, plasmid replicons), and direct comparisons of multidrug resistant
isolates. The aim of our statistical modeling was to infer sources and transmission of AMR
determinants (determined in silico), by assessing the impact of source type and location type (if
applicable) on the occurrence of select genes and plasmid replicons, while accounting for
clustering by sampling site and sampling source, and the potential confounding effect of
sampling year.
4.3 Methods
4.3.1 Dataset
Isolates examined within this study were obtained from samples collected previously for two
different projects/programs. Wildlife, livestock, and environmental samples, excluding water,
were obtained from a previous longitudinal wildlife study on swine farms and conservation areas
in southern Ontario (2011-2013; Bondo et al., 2016a). Water samples, collected in the same
geographic region and time period as the longitudinal study, were obtained through the FoodNet
Canada surveillance program. The study region was within the Grand River Watershed
(6800km2), and the surrounding region, which includes Guelph, Kitchener, Waterloo and
Cambridge, is a populous region of southern Ontario (~1 million people) that overlaps with
intensive agriculture and an abundance of natural ecosystems (~11 conservation areas).
Wildlife, livestock, and environmental samples
134
Selection of isolates for sequencing from the previous wildlife study was based on demonstrated
phenotypic resistance to at least one of 15 antimicrobials (see details below), as previously
reported by Bondo et al. (2016a). Samples from this previous wildlife study included: paw and
fecal swabs from live-trapped raccoons, skunks (Mephitis mephitis), and opossums (Didelphis
virginiana), soil samples, swine manure pit samples (from swine farms only), and dumpster
samples (from conservation areas only). Live trapping was focused on raccoons; however,
skunks and opossums that were successfully trapped were also sampled (Bondo et al., 2019).
Methods used to trap, and process wildlife have been previously described (Bondo et al. 2016b).
Briefly, animals were live-trapped at study sites once every five weeks, trapped animals were
anesthetized, and rectal fecal swabs were obtained using Cary-Blair applicators (BBL
CultureSwab, Bd; Becton, Dickinson and Company, Annapolis, Maryland, USA). In 2012 only,
the paws of captured wildlife were also sampled to assess for surface transmission of
microorganisms (Bondo et al., 2016c). Paw surfaces were sampled by wiping a Swiffer®
(Armstrong, Proctor and Gamble, Cincinnati, Ohio, USA) soaked in 25 ml of sterile saline over
the surfaces of both front and back paw pads with a gloved hand, and gloves were changed
between different animals. Ear tagging and microchipping was performed for all wildlife species
for subsequent identification. Individual raccoons were sampled up to once monthly, but samples
were obtained from the same animal if caught in subsequent trapping months. For soil samples,
approximately 10 g of soil was collected from within a 2-m radius of each animal trap and stored
in a sterile container; samples were collected on the first day of each trapping week at each site.
Manure pit samples were obtained on swine farms at the beginning of each trapping week, by
pooling samples from two different depths (i.e., top 1/3, and mid-depth) at three locations around
the pit. In 2013, dumpsters that were present in three of the five conservation areas were sampled
135
within a day of the dumpster being emptied by swabbing the entire bottom surface of the
dumpster using a Swiffer® (Armstrong, Proctor and Gamble, Cincinnati, Ohio, USA) wipe
attached to a 1-2 m extension pole. Swiffer® wipes used to sample animal paws and dumpsters
were kept moist before and after sampling by placing them in Whirl-pak bags (Nasco, For
Atkinson, Wisconsin, United States) that contained approximately 20 ml of sterile 0.85% saline.
All samples were kept on ice in the field and refrigerated until further processing. In 2011, three
E. coli isolates were cultured from each sample; for samples with more than one isolate
demonstrating phenotypic resistance, selection of one resistant isolate for sequencing was
performed using a random number generator.
FoodNet Canada water samples
Phenotypically resistant water isolates obtained as part of the FoodNet Canada surveillance
program were included in the present study if they were collected in the Region of Waterloo
sentinel site; sampling of water for E. coli in this region was initiated in 2012 and continued in
2013. Water sampling was performed bi-weekly at five core water sites in the Grand River
watershed, and three recreational areas in the study region; one of these sites was a conservation
area that was also sampled for the wildlife study. A map and description of water sampling sites
is available from Kadykalo et al. (2020). For river sites, water samples were obtained using an
extendable sampling pole and a sterile 1-litre bottle to collect water from a fast-flowing portion
of the river. Water samples were obtained from beach sites from June to September using a 1-
litre sterile sampling bottle, 15 cm below the surface. Samples were kept on ice in the field and
refrigerated until further processing.
136
4.3.2 Culture and susceptibility testing
Samples were cultured within three days of collection. Isolation of untyped E. coli from the
wildlife subset was performed at the McEwen Group Research Lab at the Canadian Research
Institute for Food Safety, University of Guelph, (Guelph, Ontario, Canada) as previously
described (Bondo et al., 2016a). Isolation of untyped E. coli from FoodNet Canada water
samples was performed at the Ontario Ministry of the Environment, Conservation and Park’s
laboratory (Toronto, Ontario) following Standard Method 9222D (APHA, 2012), except with
modified plating onto mFC-BCIG agar (Cieben et al., 1995; Lee et al., 2014). Susceptibility to
15 antimicrobials (CMV3AGNF panel) was performed using the National Antimicrobial
Resistance Monitoring System (NARMS; Sensititre, Thermo Scientific): azithromycin (AZM),
gentamicin (GEN), kanamycin (KAN), streptomycin (STR), amoxicillin-clavulanic acid (AMC),
cefoxitin (FOX), ceftiofur (TIO), ceftriaxone (CRO), ampicillin (AMP), chloramphenicol (CHL),
sulfisoxazole (SOX), trimethoprim-sulfamethoxazole (SXT), tetracycline (TCY), nalidixic acid
(NAL), and ciprofloxacin (CIP). Susceptibility testing for all samples was performed by the
Antimicrobial Resistance Reference Laboratory (National Microbiology Laboratory at Guelph,
Public Health Agency of Canada, Guelph, ON, Canada) using methods outlined by CIPARS
(Canadian Integrated Program for Antimicrobial Resistance Surveillance; Government of
Canada, 2015). Isolates with only intermediate resistance were considered susceptible, both in
selection for sequencing and for interpretation alongside in silico resistance results.
137
4.3.3 DNA extraction, whole-genome sequencing, and genome assembly
Genomic DNA extractions were performed at the University of Guelph, or at the National
Microbiology Laboratory (NML) in Winnipeg, Manitoba. Briefly, cultures of 2 ml E. coli were
grown overnight, and 1 ml was used for the Qiagen DNEasy plant and tissue 96 kit, using
manufacturer protocols (Qiagen, Hilden, Germany). Samples were then sequenced at the NML in
Guelph, Ontario or at the NML in Winnipeg, Manitoba. Nextera XT libraries and Illumina
MiSeq version 3 or NextSeq550 platforms were used according to manufacturer protocols. Raw
reads were assembled using Spades (Bankevich et al., 2012), as part of the Shovill pipeline
(version 1.0.1; https://github.com/tseeman/shovill) with the following settings: "--minlen 200 --
mincov 2; --assembler spades; --trim".
4.3.4 Analysis of whole-genome assemblies
MLST and core genome MLST analysis
Prediction of legacy multi-locus sequence types was performed using MLST (version 2.19.0;
https://github.com/tseemann/mlst), which uses the 7-loci Achtman scheme
(https://pubmlst.org/mlst/). Core-genome multi-locus sequence typing (cgMLST) of isolates was
performed using the ‘fairly simple allele calling’ tool fsac (version 1.2.0;
https://github.com/dorbarker/fsac) and the 2513-loci Enterobase scheme
(https://enterobase.warwick.ac.uk/). Isolates with 30 or more missing loci were considered poor
quality and were excluded from any further analyses.
Visualization of population structure
138
Minimum spanning trees generated by the standalone Grapetree software package (version 1.5;
Zhou et al., 2018) were used to visualize population structure, using the "MSTreeV2" algorithm.
A cluster threshold of 50 allelic differences was used for all visualizations.
In silico prediction of acquired AMR genes
Identification of acquired resistance genes was performed using Abricate (version 0.8.13;
https://github.com/tseemann/abricate) and the Resfinder database (updated May-17 2020).
Settings used were 90% identity and 60% coverage. For the identification of acquired beta-
lactamases, the identity and coverage settings were increased to 100% identity and 90%
coverage. Sensitivity and specificity of in silico AMR prediction were calculated for each
antimicrobial class, and overall (i.e., all individual test results pooled); phenotypic test results
were considered as the gold standard, and resistance was considered a positive test result. Test
sensitivity and specificity were not assessed for drug classes for which chromosomal mutations
are responsible for a considerable proportion of expressed resistance (i.e., quinolones,
aminoglycosides; Chen et al., 2007; Vetting et al., 2003). Samples with missing genotypes for
resistant phenotypic test results for three or more of the seven antimicrobial classes assessed
were examined and excluded from further analyses if the number of missing cgMLST loci
exceeded 20.
Plasmid incompatibility (Inc) type prediction
Plasmid incompatibility types were identified using Abricate (version 0.8.13;
https://github.com/tseemann/abricate) and the Plasmidfinder database (updated May-17 2020).
Settings used were 98% identity and 70% coverage.
139
In silico serotyping
Serotyping was performed using ECTyper (version 1.0.0, database version 1.0;
https://github.com/phac-nml/ecoli_serotyping) with default settings.
4.3.5 Statistical analyses
Multi-level logistic regression was used to model the odds of select plasmid replicons and
antimicrobial resistance genes being identified in E. coli from different sources. All statistical
tests were performed using STATA (STATA Intercooled 14.2; StataCorp, College Station,
Texas, USA). Only plasmid Inc types and resistance genes present in at least 10% and less than
90% of samples were modeled. In addition to examination of the overall dataset (A), analyses
were also performed for a subset of the data (B) consisting of soil and wildlife samples from
conservation areas and swine farms (Figure 4.1). Source type was examined as an independent
variable for both datasets (A) and (B), and the categorization of sampling sources for each
dataset is provided in Figure 4.1. For the subset of data (B), location type (swine farm vs.
conservation area) was also examined as an independent variable for sources replicated in both
location types (i.e., soil and wildlife samples). For the full dataset, we did not evaluate location
type as an independent variable since water samples originated from a variety of location types
across the watershed that could not be characterized by a single group distinct from swine farms
and conservation areas. A causal diagram illustrating the relationships between different
variables is provided in Figure 4.2. The mixed logistic regression models included random
intercepts to account for clustering at the site-level, and at the level of the individual sampling
source (i.e., animal, swine manure pit, dumpster). Models that did not converge using the
‘melogit’ command were fitted using ‘meqrlogit’, which uses QR decomposition methods.
140
Sampling year was included in these mixed models if it confounded the association between
source or location type (i.e., its addition caused a >20% change in the coefficient of these
variables; Dohoo et al., 2014), or if it was statistically significant (p ≤ 0.05). Model fit was
assessed by examining the best linear unbiased predictions for normality and homoscedasticity,
and Pearson's residuals were examined for potential outliers. If variance components were very
small (<1x10-3), the Bayesian Information Criterion (BIC) value was used to compare the fit of
the multi-level logistic regression model to an ordinary logistic regression, and the better fitting
model was reported (Dohoo et al., 2014). All tests were two-tailed, and a significance level of
a=0.05 was used for all analyses.
4.4 Results
4.4.1 Description of dataset
A total of 223 isolates were sequenced. Following exclusion based on 30 or greater missing loci
based on the 2513-loci cgMLST scheme, 200 isolates with the following source distribution were
available for subsequent analyses: water (n=20), swine manure pit (n=31), wildlife (n=73), and
other environmental sources (n=76; 73 soil samples, 3 dumpster samples). Wildlife isolates
originated from 58 unique raccoons, two opossums, and four skunks. In several cases, isolates
were obtained from the same animal captured on different occasions; two isolates from one
skunk, two isolates from five different raccoons, and four isolates from one raccoon. For the
subset of isolates (B) collected from the same source type on both swine farm and conservation
areas (i.e., wildlife and soil, n=146), the following distribution was noted in conservation areas:
141
wildlife (n=46; 31 raccoon fecal, 12 raccoon paw, 2 opossum fecal, 1 skunk fecal), and soil
(n=28). For the subset of isolates collected on swine farms (n=103), the distribution was as
follows: swine manure pit (n=31), soil (n=45), and wildlife (n=27; 20 raccoon fecal, 3 raccoon
paw, 3 skunk fecal, 1 skunk paw). Isolates were obtained from a total of 15 sites in the Grand
River watershed: five swine farms, six conservation areas, and four river sites in the region.
Isolates were almost evenly distributed across different years (2011, n=63; 2012, n=58; 2013,
n=79), but for water isolates, the majority were collected in 2013 (n=17/20), and the remaining
three samples were obtained in 2012.
4.4.2 Distribution of serovars and MLST types
In total, 113 serovars representing 94 sequence types were identified. Eight isolates were not
typeable with MLST (1 water, 2 manure, 1 dumpster, 1 skunk, 1 raccoon, 2 soil), and 39 isolates
could not be serotyped. Pathogenic serovars and sequence types of international importance were
identified, among which several isolates also demonstrated antimicrobial resistance. Based on
identified serotypes, two non-O157 Shiga-toxin producing E. coli isolates were identified: one
O103:H21 ST2354 isolate with phenotypic resistance to streptomycin from a skunk, and one
O103:H2 ST2307 isolate with phenotypic resistance to ampicillin, sulfisoxazole, and
trimethoprim/sulfamethoxazole from a soil sample. A number of water and wildlife isolates were
also identified as internationally important sequence types ST69, ST95, and ST131, all with
antimicrobial resistance; most of these isolates were identified in the Grand River (drinking
water intake), and in conservation area 1 (Table 4.1). None of the isolates from swine manure
pits contained sequence types of international importance or pathogenic E. coli serovars. A
summary of the sequence types and serovars identified are available in Tables 4.2 and 4.3.
142
4.4.3 Population structure based on cgMLST
The population structure of E. coli for the overall dataset (A), and for subset (B) are presented in
Figures 4.3 and 4.4, respectively. Similar or identical E. coli subtypes were identified from
diverse sources (Figures 4.3A, 4.4A), regardless of location type (for wildlife and soil samples;
Figure 4.4B). Internationally distributed sequence types are presented in Figures 4.3C and 4.4C.
The degree of antimicrobial resistance did not appear to cluster based on cgMLST subtypes
(Figures 4.3D, 4.4D).
4.4.4 In silico determination of acquired AMR genes and plasmid Inc types
In total, 43 resistance genes, and 28 plasmid incompatibility types were identified (Tables 4.4,
4.5). The most commonly identified resistance genes, with >10% prevalence overall were:
blaTEM-1, tet(A), tet(B), sul1, sul2, aph(6)-Id, ant(3’’)-Ia, aph(3’’)-Ib (Table 4.4). Three plasmid
Inc types were identified with a >10% overall prevalence: IncFIB(001918), IncI1(1-alpha), and
IncFII (Table 4.5). Resistance genes of human health importance were also identified (e.g.,
AmpC-producers). Isolates from ten samples contained blaCMY-2, and one raccoon isolate
contained a blaTEM-35. Of the isolates containing blaCMY-2, two were multi-drug resistant (MDR)
isolates from raccoons, and all but two of these isolates were obtained in conservation areas. One
of the two MDR raccoon isolates had phenotypic resistance pattern AMC-AMP-FOX-TIO-CRO-
CHL-GEN-NAL-STR-SOX-TCY and intermediate resistance to CIP and contained resistance
genes aac(3)-IVa, ant(3’’)-Ia, aph(3’’)-Ib, aph(6)-Id, floR, sul1, sul2, and tet(A) in addition to
blaCMY-2; this particular sample also contained two Inc types, IncC and IncFIBAP001918. The
other MDR raccoon isolate from a raccoon on a swine farm was resistant to AMR-AMP-FOX-
TIO-CRO-CHL-STR-SOX-TCY-SXT, and, in addition to blaCMY-2, contained genes aadA2,
143
aph(3’’)-Ib, dfrA12, floR, sul1, sul2, tet(A), and an IncC replicon. A water isolate from the Grand
River was identified with an OXA-1 beta-lactamase; this isolate demonstrated phenotypic
resistance to AMP-CHL-STR-SOX-TCY, as well as intermediate resistance to AMC, and
contained genes ant(3’’)-Ia, aph(3’’)-Ib, aph(6)-Id, floR, tet(A), tet(B), and an IncX1-1. Genes
conferring quinolone resistance were also identified: qnrS1 was identified in a raccoon in a
conservation area with IncI1(1-alpha) and IncX1-1, and qnrB19 was identified in a water sample
that displayed phenotypic resistance to AMP-CHL-GEN-STR-SOX-TCY-SXT and intermediate
resistance to CIP, and also contained genes aac(3)-IId, aadA5, aph(3’’)-Ib, aph(6)-Id, catA1,
dfrA1, sul2, blaTEM-1, and an IncFIBAP001918 replicon. Other resistance genes such as blaCARB-2
and catA1 were identified in fewer than five samples in the wildlife subset and were found only
in samples obtained from conservation areas.
4.4.5 Sensitivity and specificity of in silico AMR prediction
The overall sensitivity and specificity of in silico identification of resistance genes were 96.0%
and 97.6%, respectively (Table 4.6). Apart from a moderate test sensitivity for beta-lactamases
(90.2%), the sensitivity and specificity for all other drug classes were greater than 95% (Table
4.6).
4.4.6 Associations between source type, location type, and the carriage of resistance genes
and plasmid replicons
Genes tet(B), blaTEM-1, and sul2 were significantly associated with source type in the overall
dataset (Table 4.7). The odds of tet(B) were significantly greater in swine manure pit samples
compared to wildlife and other environmental samples, but there were no significant differences
144
between other sources (Tables 4.7, 4.8). The odds of identifying sul2 were greater in water
samples compared to all other sources, and they were significantly greater in wildlife and other
environmental samples compared to swine manure pit samples (Tables 4.7, 4.8). The odds of
blaTEM-1 were higher in water samples compared to swine manure and other environmental
samples, and higher in wildlife compared to other environmental samples (Table 4.7). No
associations with source type were identified for the remaining genes and plasmid Inc types
examined (Table 4.7). Among statistically significant models in dataset (A), clustering by
sampling site was noted for both the tet(B) and blaTEM-1 models, and model assumptions were
met.
In the subset (B) of samples that were collected in both location types, only blaTEM-1 was
significantly associated with source, with the odds being significantly greater in wildlife
compared to soil samples, and sampling site was retained as a random intercept for this model
(Table 4.9). The following genes and Inc types had a significantly greater odds of identification
on swine farms compared to conservation areas: IncFIB(AP001918), aph(3’’)-Ib, tet(A), and
aph(6)-Id (Table 4.6). IncFII, IncI1(1-alpha), ant(3’’)-Ia, sul1, sul2, and tet(B) were not
associated with either source type or location type (Table 4.9). All model assumptions were met.
4.5 Discussion
Previous work examining the role of wildlife in the maintenance and transmission of
antimicrobial resistance has often, but not always, shown that wild animals living in close
proximity to humans are more likely to carry organisms displaying resistance (Allen et al., 2011;
145
Cole et al., 2005; Ahlstrom et al., 2018; Thaller et al., 2010; Kozak et al., 2009). It is apparent
that regional approaches are most useful in understanding complex ecological issues such as the
movement of antimicrobial resistance between different sources. Selection of isolates for this
study was based on previously demonstrated antimicrobial resistance, and, although biased, this
work provides valuable insights about the nature of AMR transmission between different
wildlife, livestock and environmental sources in the Grand River watershed. The majority of
isolates in this study were collected from raccoons and soil (~70%), followed by swine manure
pits and water samples, with very few samples from additional sources (i.e., dumpsters, skunks,
opossums). In part, this distribution was related to sampling methods of previous work (e.g.,
fewer samples were obtained from swine manure pits than from raccoons; see Bondo et al.
2016a). Interestingly, the number of resistant E. coli obtained from wildlife and soil samples in
the previous wildlife study was nearly identical between conservation areas (n=74) and swine
farms (n=72), and the overall prevalence of resistant E. coli among raccoon fecal samples did not
significantly differ by location type (Bondo et al., 2016a). With location type as a proxy for the
impact of local environment (i.e., farm vs. conservation) on the occurrence of phenotypic
antimicrobial resistance in wildlife isolates, it appeared that the presence of antimicrobial
resistance in general does not vary within this region, regardless of proximity to agriculture. Our
analysis of population structure also suggests that mixing of isolates from different sources and
location types is frequently occurring, with no obvious clustering of E. coli cgMLST subtypes by
either of these factors, or by the presence of antimicrobial resistance. However, along with
previous findings of distinct AMR patterns and genes (determined by PCR) between swine farms
and conservation areas (Bondo et al., 2016a), our examination of resistance genes, plasmid Inc
146
types, internationally important sequence types, and pathogenic serovars, offers a more nuanced
picture of the distribution and movement of antimicrobial resistance within this study population.
A diversity of antimicrobial resistance genes and plasmid replicons were identified in our
study, some of which varied significantly depending on the source and location type. Although
most genes and Inc types could not be modeled due to their low overall prevalence (<5%),
several of the more prevalent resistant genes (e.g., tet(A), tet(B), blaTEM-1) and Inc types (e.g.,
IncFIB[AP001918]) occurred in at least 30% of samples in our study. Our models provide
evidence of the complexity of antimicrobial resistance, in that factors influencing the occurrence
of each resistance determinant are variable. Several overarching patterns emerged, however. For
instance, the presence of resistance genes sul2, and blaTEM-1 were consistently higher in water
samples compared to other sources. Conversely, for resistance determinants that were analyzed
by location type (i.e., wildlife and soil samples), the odds of identifying many of these genes and
plasmid Inc types were significantly greater in swine farm environments compared to
conservation areas, regardless of sampling source, which suggests that agriculture may be the
primary source of certain AMR genes and plasmids for wildlife populations in our study.
Combining findings from the overall dataset with the subset analysis of wildlife and soil samples
provides some clues about potential primary sources of certain resistance genes, albeit based on
cross-sectional data. Tetracycline genes tet(A) and tet(B), for instance, demonstrated contrasting
epidemiological patterns; the odds of an isolate carrying tet(A) were significantly greater on
swine farms than in conservation areas for wildlife and soil samples, but this gene was not
associated with any particular source overall. Meanwhile, the odds of an isolate carrying tet(B)
were not different between swine farms and conservation areas, but the odds of this gene being
147
present were significantly greater in water and swine manure samples compared to wildlife and
other environmental samples. Taken together, these findings suggest that tet(A) is widely shared
among different sources in this geographic region and those tet(A) genes found in wildlife in our
study may be originating from the swine farms where they were sampled. Conversely, tet(B) was
prevalent in swine manure pit samples included within this study, but it does not seem to be
widely disseminating into raccoons, soil or other wildlife based on proximity to farm
environments. Based on the high prevalence of tet(B) in water samples, it appears that water may
play an important role in moving this gene, which would account for its relatively even
distribution in wildlife and other environmental samples, regardless of proximity to swine farm
environments. Thus, for certain resistance determinants (e.g., tet(A)), it appears that the local
environment may act as an important predictor of their distribution. However, keeping in mind
the extensive nature of this particular watershed and the convergence of many different
waterways, the difference between farm environments and conservation areas in this region that
share broadly similar geographic characteristics may not, in fact, be sufficiently different to
influence the distribution of other genes and plasmids (e.g., tet(B)). The only resistance
determinant that was consistently associated with source type for both the overall dataset, and for
the wildlife subset, was blaTEM-1; moderate clustering by sampling site was also consistently
noted for this gene (for both datasets). With the odds of identifying blaTEM-1 being significantly
higher in water and wildlife, and with clustering by site regardless of location type, this suggests
an anthropogenic source besides the swine farms sampled in this study.
In addition to a handful of pathogenic serovars (e.g., Shiga-toxin producing O103), E.
coli isolates representing international sequence types that are commonly associated with human
148
illness were also rarely identified in raccoon, water, and soil samples (e.g., ST69, ST131, ST95;
Doumith et al., 2015), and many of these isolates also displayed MDR. Although these sequence
types could not be analyzed statistically due to their low overall prevalence, their appearance
exclusively in water sources and conservation areas suggests that anthropogenic sources are
important, and that human sources in particular may be the culprit, given the apparent association
with human recreational areas. Indirect exposure to agricultural sources via run-off cannot,
however, be ruled out, given the widespread presence of agriculture in this watershed, and the
numerous waterways which feed into water sources within these conservation areas. Our
findings of human-associated sequence types in conservation areas are consistent with previous
epidemiological findings by Bondo et al. (2016a), in that E. coli isolates resistant to >1
antimicrobial were significantly more likely to be detected on raccoon paw samples from
conservation areas compared to paw samples obtained on swine farms. Similarly, AmpC-
producing E. coli in raccoon and soil samples previously identified in this population (Bondo et
al., 2016a) were almost exclusively identified in conservation areas. Additional resistance genes
conferring resistance to quinolones and phenicols (qnrS1, catA1) identified in this study were
also only identified in raccoons sampled in conservation areas. Given the opportunistic foraging
nature of raccoons, numerous routes of transmission may plausibly contribute to the acquisition
of AMR determinants by these animals from human sources in conservation areas, including
human refuse, dog refuse, littered garbage, and/or dumpsters (Bondo et al., 2016a).
Limitations
The selection of isolates based on demonstrated phenotypic resistance contributes to a biased
representation of certain outcomes; for instance, the prevalence of pathogenic E. coli among the
149
samples included in our study should not necessarily be considered as representative of the
population of untyped E. coli obtained in the previous wildlife study (Bondo et al., 2016a). Any
outcomes presented in this work must therefore be interpreted as being within the population of
resistant E. coli from these sampling sources. Knowing that the prevalence of resistant E. coli
varied between different sources (57% of swine manure samples; 22% of water samples; 14% of
dumpster samples; 6-7% of raccoon paw and fecal samples; 10% of soil samples), the measures
of association reported within our study should not be interpreted literally as the precise odds of
identifying a certain gene (for example) in one source compared to another. Rather, our analyses
are targeted at addressing where certain genes and plasmid Inc types are most concentrated
within this population of resistant E. coli, in order to postulate primary sources, and to identify
potential transmission between sources. Using cross-sectional data to address these research
questions has inherent limitations, particularly for determining transmission, and, by extension,
the direction of transfer. Although the presence of the same genes and plasmid Inc types in
wildlife, livestock, and environmental sources is consistent with transmission events, further
characterization and examination of isolates in greater detail, particularly for plasmids, is
warranted in future work.
A number of methodological limitations also merit consideration. The prediction of
antimicrobial resistance using in silico tools generally appears to be highly sensitive and specific.
These assessments, however, did not include two major drug classes (quinolones,
aminoglycosides), since we identified only acquired resistance genes, and did not account for
chromosomal resistance mechanisms. Similarly, although all MDR isolates were positive for
plasmid Inc groups that have previously been associated with the resistance genes identified,
150
further confirmation that these genes were in fact plasmid-mediated is required. Finally, the
inclusion of location type as a predictor for our hypothesis-generating study may need to be
revisited in future work, since the characterization of agricultural sources in the ecosystem is
undoubtedly more complex than what we have suggested here, and the contribution of upstream
agricultural sources that were not directly sampled within our study were not accounted for in
our analyses. Consequently, a lack of association with location type in our study should therefore
not be taken to mean that agriculture is not an important risk factor for certain resistance genes
and plasmid Inc types in this region.
4.6 Conclusions
Recognizing that regional approaches are needed to better understand complex issues such as
antimicrobial resistance which affects all components of the ecosystem, our cross-sectional study
addresses knowledge gaps regarding the distribution and transmission of resistance genes and
plasmid Inc types in a southern Ontario ecosystem. By examining samples from wildlife,
livestock, and environmental sources, this work contributes to a more comprehensive
examination of the role of wildlife in the maintenance and transmission of antimicrobial
resistance within the Grand River watershed. While some resistance genes were associated with
certain sources (i.e., blaTEM-1 in water and wildlife), others were associated with certain location
types (i.e., aph(3’’)-Ib, tet(A), and aph(6)-Id were higher on swine farms than conservation
areas). Meanwhile, sequence types frequently implicated in human illness (i.e., ST69, ST131,
ST95) were found exclusively in water samples, and in raccoon and soil samples on conservation
areas, but not on swine farms. In combination with previous work on this raccoon population
151
demonstrating that AmpC-producing E. coli were almost exclusively identified in conservation
areas, these findings are suggestive of anthropogenic sources for these particular types of
resistance. Overall, the variability in distribution of different cgMLST subtypes, sequence types,
genes and plasmid Inc types by source and location type underscores the complex set of factors
and interactions which can influence the distribution of various determinants of resistance. Based
on our findings, it is clear that apparently healthy wildlife may act as sentinels of environmental
contamination with antimicrobial resistance, and that they may differ in their carriage of certain
sequence types, genes and plasmids according to their local environment. Future investigations
focused on intervention-based approaches, integration of antimicrobial use data on farms,
sampling of additional livestock sources, and use of whole-genome sequencing data to confirm
plasmid structure and associated genes will help to address remaining knowledge gaps.
152
4.7 References
Ahlstrom, C. A., Bonnedahl, J., Woksepp, H., Hernandez, J., Reed, J. A., Tibbitts, L., Olsen, B., Douglas, D. C., & Ramey, A. M. (2019). Satellite tracking of gulls and genomic characterization of faecal bacteria reveals environmentally mediated acquisition and dispersal of antimicrobial-resistant Escherichia coli on the Kenai Peninsula, Alaska. Molecular Ecology, 28(10), 2531–2545.
Ahlstrom, C. A., Bonnedahl, J., Woksepp, H., Hernandez, J., Olsen, B., & Ramey, A. M. (2018). Acquisition and dissemination of cephalosporin-resistant E. coli in migratory birds sampled at an Alaska landfill as inferred through genomic analysis. Scientific Reports, 8(1), 7361.
Allen, S. E., Boerlin, P., Janecko, N., Lumsden, J. S., Barker, I. K., Pearl, D. L., Reid-Smith, R. J., & Jardine, C. (2011). Antimicrobial resistance in generic Escherichia coli isolates from wild small mammals living in swine farm, residential, landfill, and natural environments in southern Ontario, Canada. Applied and Environmental Microbiology, 77(3), 882–888.
American Public Health Association (APHA), the American Water Works Association (AWWA), and the Water Environment Federation (WEF) (2012). Standard methods for the examination of water and wastewater(22nd ed.). Method 9222D. Washington, DC: APHA, AWWA & WEF.
Báez, J., Hernández-García, M., Guamparito, C., Díaz, S., Olave, A., Guerrero, K., Cantón, R., Baquero, F., Gahona, J., Valenzuela, N., Del Campo, R., & Silva, J. (2015). Molecular characterization and genetic diversity of ESBL-producing Escherichia coli colonizing the migratory Franklin's gulls (Leucophaeus pipixcan) in Antofagasta, North of Chile. Microbial drug resistance (Larchmont, N.Y.), 21(1), 111–116.
Baldi, M., Barquero Calvo, E., Hutter, S. E., & Walzer, C. (2019). Salmonellosis detection and evidence of antibiotic resistance in an urban raccoon population in a highly populated area, Costa Rica. Zoonoses and Public Health, 66(7), 852–860.
Bankevich, A., Nurk, S., Antipov, D., Gurevich, A. A., Dvorkin, M., Kulikov, A. S., Lesin, V. M., Nikolenko, S. I., Pham, S., Prjibelski, A. D., Pyshkin, A. V., Sirotkin, A. V., Vyahhi, N., Tesler, G., Alekseyev, M. A., & Pevzner, P. A. (2012). SPAdes: A new genome assembly algorithm and its applications to single-cell sequencing. Journal of Computational Biology, 19(5), 455–477.
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016a). Epidemiology of antimicrobial resistance in Escherichia coli isolates from raccoons (Procyon lotor) and the environment on swine farms and conservation areas in Southern Ontario. PLoS One, 11(11).
153
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016b). Impact of season, demographic and environmental factors on Salmonella occurrence in raccoons (Procyon lotor) from swine farms and conservation areas in southern Ontario. PLoS One, 11(9).
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016c). Epidemiology of Salmonella on the paws and in the faeces of free-ranging raccoons (Procyon Lotor) in Southern Ontario, Canada. Zoonoses and Public Health, 63(4), 303–310.
Bondo, K. J., Pearl, D. L., Janecko, N., Reid-Smith, R. J., Parmley, E. J., Weese, J. S., Rousseau, J., Taboada, E., Mutschall, S., & Jardine, C. M. (2019). Salmonella, Campylobacter, Clostridium difficile, and anti-microbial resistant Escherichia coli in the faeces of sympatric meso-mammals in southern Ontario, Canada. Zoonoses and Public Health, 66(4), 406–416.
Bonnedahl, J., & Järhult, J. D. (2014). Antibiotic resistance in wild birds. Upsala Journal of Medical Sciences, 119(2), 113–116.
Boonkusol, D., Thongyuan, S., Jangsuwan, N., & Sanyathitiseree, P. (2020). Antimicrobial resistance profiles in bacterial species isolated from fecal samples of free-ranging long-tailed macaques (Macaca fascicularis) living in Lopburi Old Town, Thailand. Veterinary World, 13(7), 1397–1403.
Cassini, A., Högberg, L. D., Plachouras, D., Quattrocchi, A., Hoxha, A., Simonsen, G. S., Colomb-Cotinat, M., Kretzschmar, M. E., Devleesschauwer, B., Cecchini, M., Ouakrim, D. A., Oliveira, T. C., Struelens, M. J., Suetens, C., Monnet, D. L., & Burden of AMR Collaborative Group (2019). Attributable deaths and disability-adjusted life-years caused by infections with antibiotic-resistant bacteria in the EU and the European Economic Area in 2015: A population-level modelling analysis. The Lancet. Infectious Diseases, 19(1), 56–66.
Chen, S., Cui, S., McDermott, P. F., Zhao, S., White, D. G., Paulsen, I., & Meng, J. (2007). Contribution of target gene mutations and efflux to decreased susceptibility of Salmonella enterica serovar Typhimurium to fluoroquinolones and other antimicrobials. Antimicrobial Agents and Chemotherapy, 51(2), 535–542.
Ciebin, B. W., Brodsky, M. H., Eddington, R., Horsnell, G., Choney, A., Palmateer, G., ... Shears, G. (1995). Comparative evaluation of mod-ified m-FC and m-TEC media for membrane filter enumeration of Escherichia coli in water. Applied and Environmental Microbiology, 61, 3940–3942.
Cole, D., Drum, D. J., Stalknecht, D. E., White, D. G., Lee, M. D., Ayers, S., Sobsey, M., & Maurer, J. J. (2005). Free-living Canada geese and antimicrobial resistance. Emerging Infectious Diseases, 11(6), 935–938.
154
Croucher, N. J., Harris, S. R., Grad, Y. H., & Hanage, W. P. (2013). Bacterial genomes in epidemiology--present and future. Philosophical Transactions of the Royal Society of London. Series B, Biological sciences, 368(1614), 20120202.
Cummins, M. L., Sanderson-Smith, M., Newton, P., Carlile, N., Phalen, D. N., Maute, K., Monahan, L. G., Hoye, B. J., & Djordjevic, S. P. (2020). Whole-genome sequence analysis of an extensively drug-resistant Salmonella enterica serovar Agona isolate from an Australian silver gull (Chroicocephalus novaehollandiae) reveals the acquisition of multidrug resistance plasmids. mSphere, 5(6), e00743-20.
De Oliveira, D., Forde, B. M., Kidd, T. J., Harris, P., Schembri, M. A., Beatson, S. A., Paterson, D. L., & Walker, M. J. (2020). Antimicrobial resistance in ESKAPE pathogens. Clinical Microbiology Reviews, 33(3), e00181-19.
Dohoo, I, Martin, W & Stryhn, H. (2014). Veterinary epidemiologic research, 3rd ed. Charlottetown, Prince Edward Island: VER Inc.
Dolejska, M., & Papagiannitsis, C. C. (2018). Plasmid-mediated resistance is going wild. Plasmid, 99, 99–111.
Doumith, M., Day, M., Ciesielczuk, H., Hope, R., Underwood, A., Reynolds, R., Wain, J., Livermore, D. M., & Woodford, N. (2015). Rapid identification of major Escherichia coli sequence types causing urinary tract and bloodstream infections. Journal of Clinical Microbiology, 53(1), 160–166.
Fischer, J., Schmoger, S., Jahn, S., Helmuth, R., & Guerra, B. (2013). NDM-1 carbapenemase-producing Salmonella enterica subsp. enterica serovar Corvallis isolated from a wild bird in Germany. The Journal of Antimicrobial Chemotherapy, 68(12), 2954–2956.
Government of Canada (2015). Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS) 2015 annual report. Guelph, ON: Public Health Agency of Canada.
Jamborova, I., Janecko, N., Halova, D., Sedmik, J., Mezerova, K., Papousek, I., Kutilova, I., Dolejska, M., Cizek, A., & Literak, I. (2018). Molecular characterization of plasmid-mediated AmpC beta-lactamase- and extended-spectrum beta-lactamase-producing Escherichia coli and Klebsiella pneumoniae among corvids (Corvus brachyrhynchos and Corvus corax) roosting in Canada. FEMS Microbiology Ecology, 94(11).
Kadykalo, S., Thomas, J., Parmley, E. J., Pintar, K., & Fleury, M. (2020). Antimicrobial resistance of Salmonella and generic Escherichia coli isolated from surface water samples used for recreation and a source of drinking water in southwestern Ontario, Canada. Zoonoses and Public Health, 67(5), 566–575.
Lee, D.-Y., Lee, H., Trevors, J. T., Weir, S. C., Thomas, J. L., & Habash, M. (2014). Characterization of sources and loadings of fecal pollutants using microbial source tracking
155
assays in urban and rural areas of the Grand River Watershed, Southwestern Ontario. Water Research, 53, 123–131.
Liakopoulos, A., Mevius, D. J., Olsen, B., & Bonnedahl, J. (2016). The colistin resistance mcr-1 gene is going wild. The Journal of Antimicrobial Chemotherapy, 71(8), 2335–2336.
Kozak, G. K., Boerlin, P., Janecko, N., Reid-Smith, R. J., & Jardine, C. (2009). Antimicrobial resistance in Escherichia coli isolates from swine and wild small mammals in the proximity of swine farms and in natural environments in Ontario, Canada. Applied and Environmental Microbiology, 75(3), 559–566.
MacKinnon, M. C., Sargeant, J. M., Pearl, D. L., Reid-Smith, R. J., Carson, C. A., Parmley, E. J., & McEwen, S. A. (2020). Evaluation of the health and healthcare system burden due to antimicrobial-resistant Escherichia coli infections in humans: A systematic review and meta-analysis. Antimicrobial Resistance and Infection Control, 9(1), 200.
Marotta, F., Janowicz, A., Di Marcantonio, L., Ercole, C., Di Donato, G., Garofolo, G., & Di Giannatale, E. (2020). Molecular characterization and antimicrobial susceptibility of C. jejuni isolates from Italian wild bird populations. Pathogens (Basel, Switzerland), 9(4), 304.
Oracová, V., Peixe, L., Coque, T. M., Novais, C., Francia, M. V., Literák, I., & Freitas, A. R. (2018). Wild corvid birds colonized with vancomycin-resistant Enterococcus faecium of human origin harbor epidemic vanA plasmids. Environment International, 118, 125–133.
Poirel, L., Potron, A., De La Cuesta, C., Cleary, T., Nordmann, P., & Munoz-Price, L. S. (2012). Wild coastline birds as reservoirs of broad-spectrum-β-lactamase-producing Enterobacteriaceae in Miami Beach, Florida. Antimicrobial Agents and Chemotherapy, 56(5), 2756–2758.
Tausova, D., Dolejska, M., Cizek, A., Hanusova, L., Hrusakova, J., Svoboda, O., Camlik, G., & Literak, I. (2012). Escherichia coli with extended-spectrum β-lactamase and plasmid-mediated quinolone resistance genes in great cormorants and mallards in Central Europe. The Journal of Antimicrobial Chemotherapy, 67(5), 1103–1107.
Thaller, M. C., Migliore, L., Marquez, C., Tapia, W., Cedeño, V., Rossolini, G. M., & Gentile, G. (2010). Tracking acquired antibiotic resistance in commensal bacteria of Galápagos land iguanas: No man, no resistance. PloS One, 5(2), e8989.
Veldman, K., van Tulden, P., Kant, A., Testerink, J., & Mevius, D. (2013). Characteristics of cefotaxime-resistant Escherichia coli from wild birds in the Netherlands. Applied and Environmental Microbiology, 79(24), 7556–7561.
Vetting, M., Roderick, S. L., Hegde, S., Magnet, S., & Blanchard, J. S. (2003). What can structure tell us about in vivo function? The case of aminoglycoside-resistance genes. Biochemical Society Transactions, 31(Pt 3), 520–522.
156
Wang, J., Ma, Z. B., Zeng, Z. L., Yang, X. W., Huang, Y., & Liu, J. H. (2017). The role of wildlife (wild birds) in the global transmission of antimicrobial resistance genes. Zoological Research, 38(2), 55–80.
Zhou, Z., Alikhan, N. F., Sergeant, M. J., Luhmann, N., Vaz, C., Francisco, A. P., Carriço, J. A., & Achtman, M. (2018). GrapeTree: Visualization of core genomic relationships among 100,000 bacterial pathogens. Genome Research, 28(9), 1395–1404.
157
4.8 Tables
Table 4.1: Distribution of sequence types of international importance identified in antimicrobial resistant Escherichia coli isolates collected from
wildlife, and environmental sourcesa in southern Ontario, Canada 2011-2013 (n=200)
Sequence type Serotype Source type Phenotypic resistance
patternb Resistance genes Location type Location id
ST69 O77/O17/O44/O106:H19 Water NAL None identified Grand River, drinking water intake
13
O77/O17/O44/O106/O73:H18
Water AMP-AZM-CIP-NAL-STR-SOX-TCY-SXT
aph(3”)-Ib, aph(6)-Id, dfra14, mphA, sul2,
tet(B), blaTEM-1
Grand River, drinking water intake
13
O25:H4 Water KAN-STR-SOX-TCY tet(B), blaTEM-1 Grand River, drinking water intake
13
O25:H4 Water AMP-NAL-TCY tet(B), blaTEM-1 Grand River, drinking water intake
13
O25:H4 Water AMP-NAL-TCY tet(B), blaTEM-1 Grand River, drinking water intake
13
O77/O17/O44/O106/O73:H18
Raccoon fecal
AMP-TCY tet(A), blaTEM-1 Conservation area 4
O77/O17/O44/O106:H18 Raccoon fecal
AMP-KAN-STR-SOX-TCY
aph(3”)-Ib, aph(3’)-Ia, aph(6)-Id, sul2, tet(B),
blaTEM-1
Conservation area 1
O77/O17/O44/O106:H18 Raccoon paw
AMP-TCY tet(A), blaTEM-1 Conservation area 5
ST95 O25:H4 Raccoon paw
CHL-SOX-SXT aadA2, ant(3”)-Ia, cmlA1, dfrA12, mefB, sul3
Conservation area 1
ST131 O25:H4 Raccoon fecal
CIP-NAL-TCY tet(A), blaTEM-1 Conservation area 1
O25:H4 Soil AMP-TCY tet(A), blaTEM-1 Conservation area 1
a No sequence types of international importance were identified in swine manure pit samples, or from wildlife or soil samples collected on swine farms. b AMP = ampicillin; CHL = chloramphenicol; CRO = ceftriaxone; FOX = cefoxitin; SOX = sulfisoxazole; STR = streptomycin; SXT = trimethoprim
sulfamethoxazole; TCY = tetracycline; NAL = nalidixic acid; CIP = ciprofloxacin.; KAN = kanamycin.
158
Table 4.2: Multi-locus sequence types identified using whole-genome sequencing data by source type for
antimicrobial resistant Escherichia coli isolates from wildlife, livestock, and environmental sources in
southern Ontario, Canada 2011-2013 (n=200)
Sequence type Count (%) 58 15 (7.5%) 10 14 (7.0%) 101 12 (6.0%) 69 8 (4.0%) 155 7 (3.5%) 542 5 (2.5%) 345 5 (2.5%) 648 4 (2.0%) 1633 4 (2.0%) 898 3 (1.5%) 3714 3 (1.5%) 1727 3 (1.5%) 34 3 (1.5%) 43 3 (1.5%) 12 3 (1.5%)
1844 3 (1.5%) 48 3 (1.5%)
*Sequence types identified in fewer than 3 samples were: 131, 295, 963, 1304, 847, 162, 354, 1721, 127, 515, 7324,
196, 641, 349, 93, 4085, 388, 547, 683, 5614, 646, 362, 602, 372, 906, 2307, 4574, 68, 2526, 2171, 6796, 6706,
2562, 2035, 457, 297, 3076, 6975, 1406, 9962, 164, 3856, 2077, 38, 4398, 88, 8097, 871, 227, 2354, 2178, 23, 9982,
3531, 1079, 117, 398, 1152, 1324, 106, 218, 154, 1585, 962, 4429, 1086, 212, 1670, 357, 1112, 716, 5259, 95, 973,
6777, 206.
159
Table 4.3: Serovars identified using whole-genome sequencing data for antimicrobial resistant Escherichia
coli isolates from wildlife, livestock, and environmental sources in southern Ontario, Canada 2011-2013
(n=200)
Serovar Count (%) O82:H8 7 (3.5%) O24:H4 6 (3.0%)
O153:H42 4 (2.0%) O46/O8/O134:H21 3 (1.5%)
O42/O28:H21 3 (1.5%) O8:H49 3 (1.5%) O83:H14 3 (1.5%) O4:H5 3 (1.5%) O6:H10 3 (1.5%)
O109:H10 3 (1.5%) O77/O17/O44/O106:H18 3 (1.5%)
O2:H18 2 (1.0%) O166:H15 2 (1.0%) O108:H2 2 (1.0%) O91:H7 2 (1.0%) O1:H34 2 (1.0%)
*Additional serovars not listed here were identified in fewer than two samples.
160
Table 4.4: Frequencies of antimicrobial resistance genes identified from whole-genome sequencing data of antimicrobial resistant Escherichia coli
isolates from wildlife, livestock, and environmental sources in southern Ontario, Canada 2011-2013 (n=200)
Antimicrobial group Resistance gene Accession no.†
Wildlifea (n=72)
Swine manure pit (n=31)
Water (n=20)
Other environmentalb (n=76) Total (%)
Aminoglycoside aac(3)-IId EU022314 3 0 1 1 5 (2.5%) aac(3)-IVa NC_009838 4 0 1 3 8 (4.0%) aadA2 JQ364967 4 1 1 3 9 (4.5%) aadA5 AF137361 6 0 1 1 8 (4.0%) ant(3'')-Ia X02340 11 7 3 14 35 (17.5%)
aph(3')-Ia V00359/
EF015636 5 0 3 4 12 (6.0%)
aph(3')-IIa V00618 0 1 0 0 1 (0.5%)
aph(3")-Ib AF321551/
AF024602 24 13 11 29 77 (38.5%)
aph(4)-Ia V01499 0 0 1 0 1 (0.5%) aph(6)-Ic X01702 0 1 0 0 1 (0.5%) aph(6)-Id M28829 24 13 11 29 77 (38.5%) Beta-lactam blaCMY-2 X91840 7 0 1 2 10 (5.0%) blaTEM-1 AY458016/
HM749966/FJ560503
28 7 11 17 63 (31.5%)
blaTEM-35 KP860986 1 0 0 0 1 (0.5%) blaCARB-2 M69058 0 0 0 1 1 (0.5%) blaOXA-1 HQ170510 0 0 1 0 1 (0.5%) Lincosamide lnuC AY928180 0 0 0 1 1 (0.5%) lnuF EU118119 0 0 0 1 1 (0.5%) Macrolide mphA U36578 3 0 1 0 4 (2.0%) mphB D85892 0 0 1 0 1 (0.5%) mef(B) FJ196385 1 0 0 1 2 (1.0%) ereA DQ157752 0 0 0 1 1 (0.5%)
161
Folate pathway inhibitors
dfrA1 AF203818/ X00926
2 1 1 5 9 (4.5%)
dfrA5 X12868 1 1 3 4 9 (4.5%) dfrA8 U10186 1 0 0 0 1 (0.5%) drfA12 AM040708 3 0 0 1 4 (2.0%) dfrA14 DQ388123 3 0 1 5 9 (4.5%) dfrA16 AF174129 0 0 0 1 1 (0.5%) dfrA17 FJ460238 6 0 1 1 8 (4.0%) dfrA23 AJ746361 1 0 0 0 1 (0.5%) sul1 EU780013 14 2 1 8 25 (12.5%) sul2 HQ840942/
AY034138 18 1 10 16 45 (22.5%)
sul3 AJ459418 3 1 1 3 8 (4.0%) Phenicol floR AF118107 7 1 2 4 14 (7.0%) catA1 V00622 2 0 2 0 4 (2.0%) cmlA1 M64556 2 1 1 2 6 (3.0%) Quinolone qnrB19 EU432277 0 0 1 0 1 (0.5%) qnrS1 AB187515 1 0 0 0 1 (0.5%) Fosfomycin fosA7 LAPJ01000014 1 0 0 3 4 (2.0%) Tetracycline tet(A) AF534183 30 11 5 37 83 (41.5%) tet(B) AF326777/
AP000342 19 16 10 19 64 (32.0%)
tet(C) AY046276/AF055345
0 0 0 3 3 (1.5%)
tet(M) X04388 0 0 0 1 1 (0.5%) † Values from Resfinder database. a Includes fecal samples from raccoons (n=51), skunks (n=4), opossums (n=2), and paw swab samples from raccoons (n=14), and one skunk. b Includes soil samples (n=73) and dumpster samples (n=3).
162
Table 4.5: Distribution of plasmid incompatibility (Inc) types identified using whole-genome sequencing data
by source type for antimicrobial resistant Escherichia coli isolates from wildlife, livestock, and environmental
sources in southern Ontario, Canada 2011-2013 (n=200)
Inc-type Source type Totalc (%)
Wildlifea (n=72)
Swine manure pit (n=31)
Water (n=20)
Other environmentalb
(n=76)
IncFIB(AP001918) 30 7 9 33 79 (39.5)
IncI1(1-alpha) 14 4 1 13 32 (16.0)
IncFII 8 7 4 10 29 (14.5)
IncFIA 7 4 2 6 19 (9.5)
p0111 7 7 2 3 19 (9.5)
IncY 3 5 1 8 17 (8.5)
IncQ1 3 1 4 6 14 (7.0)
IncX1-1 6 2 1 3 12 (6.0)
Col156 5 0 4 1 10 (5.0)
a Includes fecal samples from raccoons (n=51), skunks (n=4), opossums (n=2), and paw swab samples from raccoons
(n=15), and one skunk.
b Includes soil samples (n=73) and dumpster samples (n=3).
c Plasmid Inc types identified in fewer than 10 samples included: IncR (n=9), IncFIA(HI1) (n=6), IncFIC(FII) (n=5),
IncFIB(K) (n=4), IncFII(29), (n=4), IncHI2A (n=3), IncHI2 (n=3), IncC (n=3), IncB/O/K/Z (n=2), IncFII(pHN7A8)
(n=2), ColBS512 (n=2), ColE10 (n=2), ColpVC (n=2), IncFIB(pB171) (n=1), ColIMGS31 (n=1), IncFII(pRSB107)
(n=1), IncHI1A(CIT) (n=1), IncHI1B(CIT) (n=1), IncX1-4 (n=1).
163
Table 4.6: Test sensitivity and specificitya for in silico identification of antimicrobial resistance genes in
antimicrobial resistant Escherichia coli isolates from wildlife, livestock, and environmental sources in
southern Ontario, Canada 2011-2013 (n=200)
Antimicrobial class Test sensitivity (95%CI) Test specificity (95%CI)
Beta-lactam 90.2% (81.7-95.7%) 98.3% (94.0-99.8%)
Macrolide 100% (15.8-100.0%)* 97.0% (93.5-98.9%)
Sulfonamide 97.0% (89.8-99.6%) 97.0% (92.4-99.2%)
Phenicol 95.6% (78.0-99.9%) 99.4% (96.9-99.9%)
Tetracycline 98.6% (95.2-99.8%) 96.2% (87.0-99.5%)
Overallb 96.0% (93.2-97.8%) 97.6% (96.2-98.6%) a Phenotypic antimicrobial resistance test results were considered the gold standard. Detection of 15 antimicrobials
performed using the CMV3AGNF panel from National Antimicrobial Resistance Monitoring System (Sensititre,
Thermo Scientific). In silico acquired resistance genes detected using Abricate and the Resfinder database.
b Raw counts for all samples and antimicrobials were pooled together.
* One-sided, 97.5% confidence interval.
164
Table 4.7: Logistic regression modelsa,b,c assessing the association between source type and the occurrence of select plasmid incompatibility types and
antimicrobial resistance genes in antimicrobial resistant Escherichia coli isolates collected from wildlife, livestock and environmental sources in
southern Ontario, 2011-2013 (n=200, dataset A)
tet(A)a tet(B)b,c blaTEM-1b sul1b
Source type OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Swine manure pit REF 0.216
(global) REF 0.021 (global) REF 0.029
(global) REF 0.161 (global)
Water 0.61
(0.17-2.19) 0.433
0.92 (0.26-3.19)
0.892 4.03
(1.01-16.18) 0.050
0.72 (0.05-9.63)
0.805
Wildlife 1.27
(0.53-3.03) 0.593
0.28 (0.11-0.74)
0.010 2.11
(0.74-6.01) 0.161
3.76 (0.75-18.84)
0.107
Other environmentald 1.72
(0.73-4.09) 0.215
0.32 (0.13-0.79)
0.013 0.88
(0.31-2.54) 0.815
1.75 (0.34-9.05)
0.504
165
sul2a ant(3’’)-Iaa aph(3’’)-Ibb aph(6)-Id b
Source type OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Swine manure pit REF <0.001 (global) REF
0.810 (global) REF
0.379 (global) REF 0.379
(global)
Water 30.00
(3.40-264.50) 0.002
0.60 (0.14-2.68)
0.508 1.85
(0.51-6.66) 0.347
1.85 (0.51-6.66)
0.347
Wildlife 9.82
(1.25-77.20) 0.030
0.61 (0.21-1.75)
0.357 0.73
(0.28-1.91) 0.518
0.73 (0.28-1.91)
0.518
Other environmentald 8.00
(1.01-63.23) 0.049
0.77 (0.28-2.15)
0.624 0.91
(0.36-2.32) 0.847
0.91 (0.36-2.32)
0.847
166
IncFIB(AP001918)a IncI1(1-alpha)b,e IncFIIa
Source type OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Swine manure pit REF 0.186
(global) REF 0.594
(global) REF 0.425
(global)
Water 2.80
(0.83-9.49) 0.097
0.23 (0.01-4.40)
0.331 0.86
(0.21-3.41) 0.827
Wildlife 2.39
(0.91-6.26) 0.076
1.39 (0.28-7.00)
0.689 0.42
(0.14-1.29) 0.130
Other environmentald 2.63
(1.01-6.85) 0.047
1.09 (0.22-5.39)
0.917 0.52
(0.18-1.52) 0.232
a The random intercept to account for clustering by site or animal was not retained in the model, thus ordinary logistic regression was used.
b Included a random intercept for clustering by site. Variance components were: tet(B) 0.10 (95%CI: 0.00-4.12); blaTEM-1 0.24 (95%CI: 0.03-1.80); sul1 0.27
(95%CI: 0.02-3.12); aph(3'')-Ib 0.05 (095%CI: 0.00-28.92); aph(6)-Id 0.05 (95%CI: 0.00-28.92); IncI1(1-alpha) 0.93 (95%CI: 0.13-6.49).
c Adjusted for confounding by year of sampling
d Includes soil and dumpster samples.
e A random intercept to account for clustering of samples obtained from the same animal/dumpster/manure pit was retained in this model (variance components
0.54, 95%CI: 0.00-301.89).
167
Table 4.8: Contrasts from logistic regression modelsa,b,c assessing the association between source type and the occurrence of select antimicrobial
resistance genes in antimicrobial resistant Escherichia coli isolates collected from wildlife, livestock and environmental sources in southern Ontario,
2011-2013 (n=200, dataset A)
a Adjusted for confounding by year of sampling.
b Site of sampling was retained as a random intercept.
c The random intercept to account for clustering by site or animal was not retained in the model, thus ordinary logistic regression was used.
d Includes soil and dumpster samples.
tet(B)a,b blaTEM-1b sul2c
Contrast OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Water vs. wildlife 3.23 (0.99-10.52) 0.051 1.91 (0.59-6.16) 0.280 3.05 (1.09-8.52) 0.033
Other environmentald vs. wildlife
1.12 (0.51-2.48) 0.772 0.42 (0.19-0.90) 0.025 0.81 (0.38-1.75) 0.600
Other environmentald vs. water
0.35 (0.11-1.09) 0.071 0.22 (0.06-0.73) 0.014 0.27 (0.09-0.75) 0.012
168
Table 4.9: Logistic regression modelsa,b,c assessing the association between source type, location type, and the occurrence of select plasmid
incompatibility types and antimicrobial resistance genes in antimicrobial resistant Escherichia coli isolates from wildlife and soil samples collected on
swine farms and conservation areas in southern Ontario, 2011-2013 (n=146, dataset B)
Independent variable Sub-category
tet(A) tet(B) blaTEM-1
OR (95% CI) p-value OR
(95% CI) p-value OR (95% CI) p-value
Location type Conservation area REF
0.031a (global) REF
0.331b (global) REF
0.437b (global)
Swine farm 2.06
(1.06-4.01) 0.033
1.64 (0.60-4.44)
0.331 0.69
(0.26-1.77) 0.438
Source type Soil REF 0.698b
(global) REF 0.938b (global) REF 0.029b,c
(global)
Wildlife 0.88
(0.44-1.76) 0.716
0.97 (0.45-2.08)
0.938 3.13
(1.35-7.22) 0.008
sul1 sul2 ant(3’’)-Ia
OR (95% CI) p-value
OR (95% CI) p-value
OR (95% CI) p-value
Location type Conservation area REF
0.998a (global) REF
0.914a (global) REF
0.333a (global)
Swine farm 1.0
(0.29-3.50) 0.998
0.96 (0.44-2.08)
0.914 1.54
(0.64-3.750 0.336
Source type Soil REF
0.151a (global) REF
0.553a (global) REF
0.655a (global)
Wildlife 2.07
(0.77-5.57) 0.151
1.27 (0.58-2.76)
0.553 0.82
(0.34-1.97) 0.655
169
aph(3’’)-Ib aph(6)-Id IncFIB(AP001918)
OR (95% CI) p-value
OR (95% CI) p-value
OR (95% CI) p-value
Location type Conservation area REF
0.011a (global) REF
0.011a (global) REF
0.031a (global)
Swine farm 2.45
(1.22-4.92) 0.012
2.45 (1.22-4.92)
0.012 2.07
(1.06-4.04) 0.032
Source type Soil REF
0.492b (global) REF
0.492b (global) REF
0.738a (global)
Wildlife 0.78
(0.38-1.60) 0.492
0.78 (0.38-1.60)
0.492 0.89
(0.46-1.72) 0.738
IncI1(1-alpha) IncFII
OR (95% CI) p-value
OR (95% CI) p-value
Location type Conservation area REF
0.203b (global) REF
0.571b (global)
Swine farm 0.39
(0.09-1.66) 0.203
1.33 (0.49-3.59)
0.573
Source type Soil REF
0.516b (global) REF
0.614a (global)
Wildlife 1.36
(0.54-3.43) 0.516
0.77 (0.29-2.09)
0.615
a The random intercept for clustering by site or animal was not retained in the model, thus ordinary logistic regression was used.
b Included a random intercept for clustering at the site-level. Variance components were: tet(A) source type 0.22 (95%CI: 0.03-1.72); tet(B) source type 0.13
(95%CI: 0.00-4.63); tet(B) location type 0.16 (95%CI: 0.00-2.91); blaTEM-1 location type 0.22 (95%CI: 0.02-2.33); blaTEM-1 source type 0.24 (95% CI: 0.02-2.40);
sul1 location type 0.40 (95%CI: 0.05-3.39); sul1 source type 0.41 (95%CI: 0.05-3.36); aph(3’’)-Ib source type 0.32 (95%CI: 0.02-3.50); aph(6)-Id source type
0.32 (95%CI: 0.02-3.50); IncI1(1-alpha) source type 0.73 (95%CI: 0.11-4.69); IncI1(1-alpha) location type 0.63 (0.10-3.95).
c Adjusted for confounding by year of sampling.
170
4.9 Figures
Figure 4.1. Classification of source types for full dataset (A), and subset of data (B), with sample sizes and independent variables analyzed.
171
Figure 4.2. Causal diagram for analytical modeling. Solid lines show directionality of proposed relationships between dependent and independent
variables. Dashed grey lines show potential confounding relationships. Location type (swine farm vs. conservation area) only applies to samples
collected on swine farms and conservation areas as part of previous wildlife study.
172
173
174
175
Figure 4.3. Population structure of 200 antimicrobial resistant Escherichia coli isolates from wildlife, livestock, and environmental sources in southern
Ontario based on 2513-loci cgMLST scheme from Enterobase. Minimum spanning tree created using k=50 clustering threshold in GrapeTree. (A)
Distribution of source types. (B) Pathogenic serovars. (C) Sequence types of international importance. (D) Antimicrobial resistance by number of drug
classes. Frequency counts are in square brackets.
176
177
178
179
Figure 4.4. Population structure of 146 antimicrobial resistant Escherichia coli isolates from wildlife and soil
on swine farms and conservation areas in southern Ontario based on 2513-loci cgMLST scheme from
Enterobase. Minimum spanning tree created using k=50 clustering threshold in GrapeTree. (A) Distribution
of sources. (B) Distribution by location type. (C) Sequence types of international importance. (D)
Antimicrobial resistance by number of drug classes. Frequency counts are in square brackets.
180
CHAPTER 5
5 Genomic epidemiology of non-typhoidal Salmonella enterica and
associated antimicrobial resistance in raccoons (Procyon lotor),
humans, livestock, and environmental sources in southern
Ontario, and the potential for zoonotic transmission: A risk
factor analysis
5.1 Abstract
Non-typhoidal Salmonella infections represent a substantial burden of illness to human health,
and the increasing prevalence of antimicrobial resistance among these infections is a growing
concern. Using a combination of whole-genome sequencing data of bacterial isolates from
human, raccoon, livestock and environmental sources, and epidemiologically-informed statistical
modeling, our objective was to determine if there was evidence for potential transmission of
Salmonella and associated antimicrobial resistance determinants (determined in silico) between
these different sources in the Grand River watershed. Outputs from bioinformatics analyses were
used in logistic regression models to assess for potential associations between source type and
the occurrence of select resistance genes and plasmid replicons. A total of 608 isolates were
obtained from the following sources: human (n=58), raccoon (n=92), livestock (n=329), and
environment (n=129). Population structure based on 3002-loci cgMLST suggests that
transmission of Salmonella and associated plasmid types and resistance genes between humans,
181
livestock, and water samples may be occurring. In contrast, none of the Salmonella isolates from
raccoons clustered together with human isolates. Resistance genes of public health importance,
including AmpC producers (i.e., blaCMY-2) were identified in human, livestock, and
environmental sources, but not in raccoons. In addition, quinolone resistance proteins (e.g.,
qnrS1, qnrB20) were only identified in water and human samples, but not in livestock or raccoon
isolates. Some plasmid incompatibility groups were strongly associated with certain sources;
some were only isolated from raccoons and soil samples, whereas others were only isolated from
human and livestock sources. Collectively, our findings suggest that the rural populations of
raccoons on swine farms in the Grand River watershed are likely not major contributors to
antimicrobial resistant human Salmonella infections in this region.
Keywords
Antimicrobial resistance, Salmonella, source attribution, whole-genome sequencing, wildlife
5.2 Introduction
Non-typhoidal Salmonella infections represent a major threat to public health worldwide (World
Health Organization, 2015). It has been estimated that non-typhoidal Salmonella causes roughly
93.8 million infections every year, resulting in approximately 155,000 deaths worldwide
(Majowicz et al., 2010). The majority of infections result in self-limiting gastroenteritis, but
invasive bloodstream infections can also occur, with children, the elderly, and
immunocompromised individuals at highest risk of developing life-threatening illness (GBD
2017, 2019). The medical treatment of severe Salmonella infections can also become
182
complicated by the presence of antimicrobial resistance (AMR; Michael & Schwarz, 2016);
resistant infections contribute to an increased risk of complications and death, as well as
prolonged recovery times (Kern & Rieg, 2020; Tanwar et al., 2014). The widespread use of
antimicrobials in agriculture and in humans is thought to drive the selection of resistance and
virulence among pathogenic organisms such as Salmonella (Kumar et al., 2020; Bawn et al.,
2020), and although wildlife species are increasingly being examined for their potential role in
zoonotic diseases (Torres et al., 2020), their role in the epidemiology of Salmonella and AMR is
not entirely clear. To achieve a greater understanding of the drivers of zoonotic infections in
humans, and the transmission of AMR within the ecosystem, research approaches using multiple
sampling sources such as wildlife and environmental samples, are increasingly being employed
(Munck et al., 2020; Aung et al., 2019; Guyomard-Rabenirina et al., 2019; De Lucia et al., 2018;
Jokinen et al., 2015). In addition, the increasing availability and application of next-generation
sequencing technologies in public health surveillance and source attribution studies have
revolutionized our ability to address important research gaps concerning the primary sources and
drivers of transmission of zoonotic pathogens and AMR determinants (Wee et al., 2020;
Croucher et al., 2013).
Recent work using comparative genomics has provided some epidemiological evidence
that wild birds, as reservoirs, may contribute to the transmission of Salmonella between different
human, animal, and environmental sources (Pornsukarom et al., 2018; Bloomfield et al., 2017;
Toro et al., 2016; Mather et al., 2016). Work by Ahlstrom et al., (2018, 2019) examining Alaskan
gulls suggests that these birds can acquire resistance genes of public health importance (e.g.,
blaCTX-M, blaCMY-2) from anthropogenic sources such as landfills, and subsequently disseminate
183
them to other distant locations via foraging or migration. With the vast majority of zoonotic
disease investigations focused on wild birds, there is comparatively little work examining
terrestrial wild mammals such as raccoons (Procyon lotor). A number of cross-sectional surveys
have determined that raccoons can be asymptomatic carriers of Salmonella, with serovars that
overlap with those commonly responsible for causing disease in humans (Baldi et al., 2019;
Rainwater et al., 2017; Bondo et al., 2016; Lee et al., 2011). A couple of studies which included
genotypic assessments of Salmonella demonstrated the presence of identical pulsed-field gel
electrophoresis (PFGE) fingerprints for certain human and raccoon isolates (Very et al., 2016;
Compton et al., 2008). As a highly adaptable species that is prevalent in a variety of landscapes,
and that often lives and forages in close proximity to humans and domestic animals, raccoons
arguably merit further examination as a potential source of Salmonella in humans.
Using a combination of epidemiologically-informed statistical modeling and whole-
genome sequencing data of Salmonella enterica isolates collected from raccoons, humans,
livestock, and environmental sources in the Grand River Watershed in Ontario, our broad
research objective was to address current knowledge gaps in the role of raccoons in the
epidemiology of human Salmonella infections in this region, as well as the movement of
associated AMR determinants in this ecosystem. Our specific objective was to broadly
characterize Salmonella strains and the distribution of associated AMR determinants found in
raccoons, humans, livestock, and environmental sources, with a focus on potential raccoon-
human transmission. To address our study objective, the population structure of 608 isolates of
Salmonella was determined using Enterobase's 3002-loci cgMLST scheme
(https://enterobase.warwick.ac.uk), and resistance genes and plasmid incompatibility groups
184
were identified using in silico tools. Outputs from bioinformatics analyses were used in various
logistic regression models to assess for potential associations between source type and the
occurrence of select resistance genes and plasmid replicons.
5.3 Methods
5.3.1 Dataset
Isolates examined within this study were obtained from three different collections of Salmonella
cultures that were previously sampled through: 1) a previous wildlife research study (Bondo et
al., 2016) and public health surveillance programs led by 2) FoodNet Canada and 3) Public
Health Ontario. Salmonella isolates collected through surveillance from human, livestock, and
environmental sources were identified and included if the samples originated from the same
geographic region and time period as the wildlife study (Wellington-Dufferin and Region of
Waterloo Counties in southern Ontario, 2011-2013). The study region which includes Waterloo,
Kitchener, Guelph, Cambridge and surrounding areas, contains a population of ~1 million people
that is situated within a region of intensive agriculture, and also includes the Grand River
watershed (6800 km2), the largest watershed in southern Ontario. The intersection of humans,
agriculture, and wildlife in this particular region presents a unique opportunity to study the
potential transmission of bacteria and AMR determinants between different human, animal, and
environmental sources.
Raccoon, swine manure pit, and soil samples
185
Salmonella isolates for the wildlife component of this study were obtained from a previous three-
year repeated cross-sectional study of raccoons on swine farms and in conservation areas; the
subset of Salmonella isolates collected on swine farms were selected for sequencing (Bondo et
al., 2016). These samples included fecal swabs from live-trapped raccoons, soil samples
collected adjacent to trapped raccoons, and swine manure pit samples. The study site and
methods used for live-trapping and processing of raccoons have been described previously
(Bondo et al., 2016). Briefly, raccoons were live-trapped on five swine farms near Cambridge
and Guelph in southern Ontario from May through November, between 2011 and 2013. Trapping
occurred once every five weeks on each farm. Trapped animals were anesthetized and rectal
fecal swabs were obtained using Cary-Blair applicators (BBL CultureSwab, BD; Becton,
Dickinson and Company, Annapolis, Maryland, USA). Raccoons were ear-tagged and
microchipped for subsequent identification; individuals were sampled up to once monthly, but
additional samples were collected from the same animal if they were caught in subsequent
months. Soil samples were obtained within a 2-m radius of traps on the first day of each trapping
week at each site; approximately 10 g of soil was collected into sterile containers. For manure pit
samples, one pooled sample was collected on each swine farm at the beginning of each trapping
week. The pooled sample was obtained by sampling from two different depths (i.e., top 1/3, and
mid-depth) at three locations around the manure pit. Samples were kept on ice in the field and
refrigerated until further processing.
Livestock, water, and beach samples
The recruitment and selection of farms for FoodNet Canada has previously been described by
Flockhart et al. (2017). Livestock samples included samples from swine, chickens (broilers and
186
layers), and cattle (dairy and beef). Samples included both fresh fecal samples and manure pit
samples for all species except for layers, from which only fresh fecal samples were obtained.
Fresh fecal samples were obtained by pooling fresh samples from five individual animals.
Manure pit samples were obtained from either liquid manure pits or from dry piles. Liquid
manure pit samples were obtained in the same manner as described above for the wildlife study.
Dry manure piles were sampled by pooling samples from five different locations in the pile.
Water samples were obtained from five core water sites in the Grand River watershed
upstream of regional drinking water intake. Sampling of these water sites was performed twice
monthly, yielding approximately 120 samples per year. Samples were obtained from a fast-
flowing portion of the river using an extendable sampling pole and a sterile 1-litre bottle. Beach
water samples were obtained from local swimming sites in three different recreational areas in
southern Ontario (Shade's Mills, Laurel Creek, Elora Gorge). Beach sites were sampled twice
monthly from June to September between 2011 and 2013; these water samples were collected
15cm below the surface, using a 1-litre sterile sampling bottle.
All livestock, water, and beach samples were kept on ice in the field and refrigerated until
further processing. Traditional serotyping and phage typing were performed by the Public Health
Agency of Canada, Laboratory for Foodborne Zoonoses (LFZ), Office International des
Epizooties Reference Laboratory for Salmonellosis, Salmonella Typing Laboratory, Guelph,
Ontario. Sequencing of isolates by FoodNet Canada was performed for all samples, except for
samples with identical metadata (i.e., same submitter, collection date, commodity, serotype, and
phage type, if available); among samples with identical metadata, up to five isolates were
selected for sequencing. Contaminated samples, and samples without a serotype were not eligible
187
for sequencing. Sequenced water and livestock FoodNet samples were included in the present
study if they were collected within the Region of Waterloo sentinel site between 2011 and 2013.
Human samples
Salmonella isolates from human clinical cases collected by Public Health Ontario as part of the
reportable disease surveillance program were cultured from blood, stool, and urine specimens.
Selection of isolates for sequencing was based on criteria for specific research projects
examining select mechanisms of plasmid-mediated and multi-drug resistance (e.g., ciprofloxacin
resistance, extended-spectrum beta-lactamases, blaCMY-2 in S. Heidelberg), thus all human
isolates included in the present study demonstrated phenotypic resistance to at least one
antimicrobial. Previously sequenced human Salmonella isolates originating from the Wellington-
Dufferin-Guelph or Region of Waterloo Public Health Counties in Ontario between 2011 and
2013 were included in the present study.
5.3.2 Salmonella culture and phenotypic susceptibility testing
Samples were cultured within three days of collection. Isolation of Salmonella was performed
according to methods described in 'National Integrated Enteric Disease Surveillance Program:
Sample Collection, Preparation and Laboratory Methodologies, January 2010' (http://www.phac-
aspc.gc.ca/foodnetcanada.pdf/lab_sop-eng.pdf). Presumptive-positive colonies were confirmed
via biochemical reactions on triple sugar iron slant and urea slant agars (Becton, Dickinson).
Additional confirmation was performed using Salmonella O Poly A-I & Vi antiserum (Becton,
Dickinson). Susceptibility testing for human samples was performed using panels of 12, 26, or
27 antimicrobials. Wildlife, livestock, and environmental samples were tested for susceptibility
188
to the following 15 antimicrobials (CMV3AGNF panel) using the National Antimicrobial
Resistance Monitoring System (NARMS; Sensititre, Thermo Scientific): gentamicin (GEN),
kanamycin (KAN), streptomycin (STR), amoxicillin-clavulanic acid (AMC), cefoxitin (FOX),
ceftiofur (TIO), ceftriaxone (CRO), ampicillin (AMP), chloramphenicol (CHL), sulfisoxazole
(SOX), trimethoprim-sulfamethoxazole (SXT), tetracycline (TCY), nalidixic acid (NAL),
ciprofloxacin (CIP), and azithromycin (AZM). Susceptibility testing for livestock, wildlife, and
environmental samples was performed by the Antimicrobial Resistance Reference Laboratory
(National Microbiology Laboratory at Guelph, Public Health Agency of Canada, Guelph, ON,
Canada) in accordance with methods outlined by the Canadian Integrated Program for
Antimicrobial Resistance Surveillance (CIPARS; Government of Canada, 2015). Phage-typing
of the subset of isolates in the previous wildlife study was performed as described by Bondo et
al. (2016).
5.3.3 DNA extraction, whole-genome sequencing and genome assembly
Genomic DNA extractions were performed at the University of Guelph, or at the National
Microbiology Laboratory (NML) in Winnipeg, Manitoba. Briefly, cultures of 2 ml Salmonella
were grown on Mueller Hinton Agar and incubated at 35 °C overnight. Cultures were then
distributed to the NML in Winnipeg for DNA extraction and sequencing, or these steps were
performed on site, at the University of Guelph and the NML in Guelph, Ontario, respectively.
Cultures of 1 ml Salmonella were used as input to the Qiagen DNEasy plant and tissue 96 kit,
following manufacturer protocols (Qiagen, Hilden, Germany). Sequencing was performed using
the Nextera XT libraries and Illumina MiSeq version 3 (600-cycle kit) or NextSeq550 platforms,
according to the manufacturer’s protocols. Raw reads were assembled using Spades (Bankevich
189
et al., 2012), as part of the Shovill pipeline (version 1.0.1; https://github.com/tseeman/shovill)
with the following settings: "--minlen 200 --mincov 2; --assembler spades; --trim".
5.3.4 Analysis of whole-genome assemblies
MLST and core genome MLST analysis
Prediction of legacy multi-locus sequence types was performed using MLST (version 2.19.0;
https://github.com/tseemann/mlst), which uses the Achtman 7-loci scheme for Salmonella
enterica (https://pubmlst.org/mlst/). Core-genome multi-locus sequence typing (cgMLST) of
isolates was performed using the ‘fairly simple allele calling’ tool fsac (version 1.2.0;
https://github.com/dorbarker/fsac) with the 3002-loci Enterobase scheme
(https://enterobase.warwick.ac.uk/). Isolates with 20 or more missing loci were considered poor
quality and excluded from any further analyses.
Visualization of population structure
Population structure was visualized using minimum spanning trees generated by the standalone
Grapetree software package (version 1.5; Zhou et al., 2018), using the "MSTreeV2" algorithm. A
cut-off of 30 allelic differences was used for trees containing more than four serovars, and a cut-
off of 10 allelic differences was used for visualizations of up to four serovars.
In silico prediction of acquired AMR genes
Acquired resistance genes were identified using CARD-RGI (database version 3.0.8;
https://github.com/arpcard/rgi) which uses a curated database that includes optimized identity
and coverage settings for each gene (Alcock et al., 2020). The following setting was used:
"--exclude nudge" (i.e., loose hits were not nudged to strict). Only acquired resistance genes
190
identified with perfect and strict hits under the "protein homolog model" were reported;
regulatory genes, genes requiring mutation for expression, and genes representing
chromosomally encoded resistance genes were not reported. Test sensitivity and specificity of in
silico AMR prediction were calculated for each antimicrobial class, as well as overall (i.e., all
individual test results from all antimicrobials pooled together) for the NARMS panel of
antimicrobials (n=15), with phenotypic test results considered the gold standard, and resistance
was considered a positive test result. Since only acquired resistance genes were identified, we
elected not to assess test sensitivity or specificity of drug classes for which chromosomal
mutations are known to confer a considerable proportion of expressed resistance (e.g.,
quinolones, aminoglycosides; Chen et al., 2007; Vetting et al., 2003). Isolates with intermediate
resistance were categorized as susceptible. Isolates with greater than two missing genotypes for
resistant phenotypic test results (by antimicrobial class), and a sequencing depth of <30 using
Mash (Ondov et al., 2019) were excluded from further analyses.
Plasmid replicon prediction
In silico identification of plasmid incompatibility groups was performed using Abricate (version
1.0.1; https://github.com/tseemann/abricate) and the PlasmidFinder database (updated May-17
2020) which uses the database from the Center for Genomic Epidemiology, Technical University
of Denmark, DTU (https://cge.cbs.dtu.dk/services/PlasmidFinder/). Settings included nucleotide
coverage of 70% and percent identity of 98%.
In silico serotyping
Identification of serovars was performed using the SISTR command-line tool (version 1.1.1;
https://github.com/phac-nml/sistr_cmd), and default settings with the "centroid" allele database.
191
5.3.5 Statistical analyses
Multi-level logistic regression was used to model the odds of the presence of select plasmid
replicons and acquired resistance genes found in Salmonella from different animal, human, and
environmental sources. All statistical tests were performed using STATA (STATA Intercooled
14.2; StataCorp, College Station, Texas, USA). Only plasmid replicons or resistance genes with
an overall prevalence of at least 10% and at most 90% were analyzed using this modeling
approach. Plasmid types that were present only or predominantly (>95%) in one host-restricted
serovar were reported descriptively only, due to the potential for complete confounding by
serovar. Human and raccoon sources were retained as categories for analyses, and additional
sources were grouped together into the following categories: livestock (cattle, swine, chicken,
swine manure pit), environmental (beach, water, soil). A causal diagram illustrating the causal
relationships between different variables is provided in Figure 5.1. Logistic regression models
with source type as an independent variable were initially constructed using the ‘melogit’
command, with a random intercept to account for clustering of isolates obtained from the same
raccoon or swine manure pit (applicable to samples from the wildlife study). Site-level data for
isolates obtained through FoodNet were unavailable, thus, isolates that may have originated from
the same farm were treated as independent observations. For multi-level models that did not
converge using ‘melogit’, the model was fitted using QR decomposition of parameters, using the
command ‘meqrlogit’. All models were adjusted for the potential confounding effect of sampling
year, if deemed appropriate. Sampling year was retained in the model based on the following
criteria: it was statistically significant (p ≤ 0.05), or its removal resulted in a >20% change in the
coefficient of the source type variable (Dohoo et al., 2014). To examine the fit of multi-level
192
models, Pearson’s residuals were assessed for outliers, and the best linear unbiased predictions
(BLUPs) were assessed for normality and homoscedasticity. The random intercept was removed
if the variance components were very small (<1x10-3) and if model fit was not improved based
on changes to the Bayesian Information Criterion (BIC) value (Dohoo et al., 2014). If the
random intercept was not retained in the final model, the model was fit using ordinary logistic
regression. Exact logistic regression was used if low effective sample sizes posed estimation
issues or to obtain median unbiased estimates if certain categories had zero positive
observations; the score method was used to calculate p-values for these models. All tests were
two-tailed, and a significance level of a=0.05 was used.
5.4 Results
5.4.1 Description of dataset
A total of 628 isolates were identified using the study criteria. After exclusion of isolates with 20
or more missing loci based on the 3002-loci cgMLST scheme, 608 isolates were available for
subsequent analyses from the following sources: human (n=58), raccoon (n=92), chicken
(n=214), cattle (n=60), swine (n=34), soil (n=45), swine manure pit (n=21), water (n=76), and
beach (n=8). Among isolates derived from cattle samples, these were roughly equally
represented by beef (n=29/60) and dairy animals (n=31/60). The majority of chicken isolates
were obtained from broilers (n=204/214), with a handful of isolates from laying hens
(n=10/214). The subset of isolates obtained as part of the previous wildlife study (i.e., raccoon,
soil, and swine manure pit samples) totaled 158. Isolates from all sources were roughly equally
193
distributed across different sampling years: 2011 (36%), 2012 (34%), 2013 (30%). Since all
human isolates included in this study exhibited phenotypic resistance to at least one
antimicrobial, the prevalence of phenotypic multi-class resistance (to at least three drug classes)
was highest in this source (65.5%, 95%CI: 51.8-77.5%). The prevalence of multi-class resistant
and pan-susceptible isolates in the remaining source categories (which all included both
phenotypically resistant and susceptible isolates) was as follows: livestock (10.6% multi-class
resistant, 95%CI: 7.5-14.5%, 57.4% pan-susceptible, 95%CI: 51.9-62.8%), raccoon (2.2% multi-
class resistant, 95%CI: 0.2-7.6%, 96.7% pan-susceptible, 95%CI: 90.8-99.3%), and
environmental (3.1% multi-class resistant, 95%CI: 0.9-7.7%, 85.3% pan-susceptible, 95%CI:
77.9-90.9%).
5.4.2 Serovar distribution
A total of 45 serovars were identified, and the most commonly identified serovars among all
isolates were Kentucky (n=132), Heidelberg (n=99), Newport (n=55), Typhimurium (n=39),
Agona (n=38), Infantis (n=33) and Enteritidis (n=30; Figure 5.2A). The population structure
(based on cgMLST) and distribution of source types for all isolates is illustrated in Figure 5.2B.
Figure 5.3C illustrates the distribution of internationally recognized sequence types. Among
raccoon isolates, the most frequently identified serovars were Newport (n=33/92, 36%), Agona
(n=16/92, 17%), Infantis (n=9/92, 9.8%), and Paratyphi var. Java (n=9/92, 9.8%). A heatmap of
the distribution of human serovars in this study that were also identified in other sources is
provided in Figure 5.3. The most prevalent serovars in human isolates were Enteritidis (n=14/58,
24%), Heidelberg (n=14/58, 24%), and Kentucky (n=11/58, 19%); the latter two serovars were
infrequently represented among environmental and raccoon isolates, whereas they were
194
commonly identified in samples from chickens and cattle (Figure 5.3). Salmonella Enteritidis
was isolated from all sources except for swine and beach samples. Several serovars that were
infrequently identified in human samples (<5%) were not identified in any other sample in our
study (i.e., Litchfield, Dublin, Chester; Figure 5.3). Serovars common to both humans and
raccoons, in decreasing frequency relative to raccoons, included: Newport, Infantis,
Typhimurium, Heidelberg, and Enteritidis. The genetic relationships among isolates of Newport,
Infantis, and Typhimurium are provided in Figure 5.4; since only one isolate each of Heidelberg
and Enteritidis was identified among raccoon samples, these serovars were reported descriptively
only. For detailed information regarding the distribution of all serovars by source, see Table 5.1.
5.4.3 MLST results
A total of 54 sequence types were identified, and the distribution of the 11 most common
sequence types by source is provided in Table 5.2. Seven isolates were not typeable with MLST
(2 raccoon, 1 soil, 2 water, 1 swine, 1 chicken). The following internationally recognized
sequence types were identified in our study population: ST15 Heidelberg (n=95), ST19
Typhimurium (n=33), ST32 Infantis (n=32), ST198 Kentucky (n=15), ST96 Schwarzengrund
(n=7), ST10 Dublin (n=2), ST45 Newport (n=2), and ST65 Brandenburg (n=2; Figure 5.2C). As
chickens were the most prevalent source type in this study, sequence types from poultry adapted
serovars such as Kentucky (ST152), and Heidelberg (ST15) were most common (Table 5.2).
Salmonella Kentucky isolates identified in human cases (exclusively ST198) were infrequently
identified in chicken samples (n=4; Table 5.2). Among Salmonella Kentucky isolates from
chickens, the vast majority (n=109/114) of these were ST152, and the remaining samples were
ST198 (n=4/114; one sample could not be typed). Most Salmonella Newport isolated from
195
raccoon and environmental sources were ST350, whereas the two human Newport isolates
identified in our study were both ST45. Certain sequence types were identified in every source:
ST32 (Infantis) and ST19 (Typhimurium). Other sequence types were identified only or
predominantly in certain animal sources, in addition to soil or water samples: ST367 (Cerro) in
cattle, and ST592 (Worthington) in swine. For detailed information regarding the distribution of
sequence types by source, see Table 5.3.
5.4.4 Population structure based on cgMLST
Serovars identified in both human and raccoon samples were: Enteritidis, Heidelberg, Infantis,
Newport, and Typhimurium (Figure 5.4). Based on our previously specified cluster threshold of
30 loci, none of the human and raccoon isolates clustered together for any of Infantis, Newport,
or Typhimurium isolates, with a minimum of 73, 2589, and 76 different loci, respectively (Figure
5.4). Single antimicrobial susceptible isolates of Enteritidis (ST11) and Heidelberg (ST15) were
identified in raccoons and differed from human isolates by a minimum of 61 and 4 loci,
respectively. Certain serovars identified in all sources such as Salmonella Infantis were relatively
similar regardless of source; all 33 isolates were within 83 allelic differences of one another,
whereas all non-human sources containing Infantis could be grouped together at a lower
threshold of 58 allelic differences. Similarly, all 38 Salmonella Agona (isolated from all source
types except humans) clustered in a single group within 34 allelic differences of one another. A
similar pattern was seen with Salmonella Enteritidis that was isolated from all sources (but
predominantly humans and livestock), with the majority of isolates (>95%) clustering into three
major groups at a threshold of 28 allelic differences, except for the single raccoon Enteritidis
isolate, which differed from the nearest group of human isolates by 61 alleles. Salmonella
196
Typhimurium, also isolated from all sources, displayed a higher clustering threshold for all
isolates, with all but two isolates grouped together at a threshold of 400 allelic differences. In
contrast, serovars such as Heidelberg that were predominantly isolated from chickens displayed a
lower clustering threshold together with samples from humans, water, cattle, and the single
raccoon Heidelberg isolate (n < 15 allelic differences for 85% of isolates). Two distantly related
clusters of Salmonella Kentucky were identified (>2800 allelic differences), with one major
group that consisted mostly of chicken samples, with a small number of water (n=3) and cattle
(n=4) samples that could be grouped together with 118 allelic differences (no Salmonella
Kentucky were isolated from raccoons); the minor group of Salmonella Kentucky included 4
chicken isolates and 11 human isolates that were within 141 allelic differences of one another.
Finally, for serovars such as Salmonella Newport that were mainly identified in environmental
sources and raccoon samples, the majority (>90%) of isolates were within 62 allelic differences
of one another.
5.4.5 In silico determination of acquired AMR genes and plasmid replicons
A total of 65 acquired resistance genes and 43 plasmid incompatibility types were identified
among all isolates. Commonly identified resistance genes with an overall prevalence >10% were:
fosA7, tet(B), aph(6)-Id, aph(3'')-Ib, aac(6')-Iaa, and blaCMY-2 (Table 5.4). Human isolates
contained the greatest diversity of resistance genes, with 58/65 genes identified. Other sources
contained fewer resistance genes, in decreasing order: livestock (29/65), environmental samples
(27/65), and raccoons (9/65). A number of human and livestock isolates, along with a handful of
environmental isolates, contained class A and C beta-lactamase genes responsible for conferring
the extended-spectrum beta-lactamase (ESBL) phenotype (e.g., blaCTX-M-14, blaCTX-M-65) and the
197
cephamycinase-producing phenotype (i.e., blaCMY-2). One human isolate was identified with the
mcr-1 gene conferring resistance to colistin. Few resistance genes were identified among raccoon
isolates (Table 5.4), corresponding with the n=3/92 samples that demonstrated phenotypic
resistance (1 sample with STR-TCY, and 2 samples with SOX-STR-TCY). Most of the
resistance genes identified in raccoon isolates conferred resistance to aminoglycosides and
fosfomycin; no beta-lactam, phenicol, or quinolone resistance genes were identified in samples
from these animals (Table 5.4). In contrast, environmental isolates contained resistance genes to
several major classes of antimicrobials (including beta-lactamases, phenicols, and quinolones;
Table 5.4). The most commonly identified plasmid incompatibility types are presented in Table
5.5. Some plasmid replicons were restricted to certain sources (e.g., ColE10 in swine), or
appeared predominantly in certain sources (e.g., IncX1-3 in chicken isolates, IncFiip96a in
raccoon and environmental isolates). Other plasmid replicons were identified in all source types
(e.g., IncI1-Igamma, IncFIIS).
5.4.6 Sensitivity and specificity of in silico AMR identification
The overall sensitivity and specificity of in silico identification of AMR genes were 94.3%, and
99.1%, respectively (Table 5.6). Most test sensitivity and specificity values for drug classes were
96% or greater.
5.4.7 Associations between source type and carriage of genes and plasmid replicons
Logistic regression models for plasmid incompatibility types and resistance genes are presented
in Table 5.7, and contrasts from these models are available in Table 5.8. Except for IncI1-
Igamma, all plasmid replicons and resistance genes examined had a significant association with
198
source type. Certain plasmid replicons were significantly more likely to be identified in human
or livestock than in raccoon or environmental samples (i.e., IncX1-1, colRNAI, colpVC, IncI1-
Igamma), whereas others were significantly more likely to be identified in raccoon and
environmental samples compared to human and livestock samples (i.e., IncFIIS, IncFiip96a;
Tables 5.7, 5.8). Plasmid types IncX1-3 and IncFIB(AP001918) were not modeled since they
were exclusively, or almost exclusively (>95%), identified in one serovar (Salmonella
Kentucky), and the vast majority (>85%) of these isolates originated from broiler chickens. Most
resistance genes analyzed were significantly more likely to be identified in livestock and/or
human samples compared with environmental and/or raccoon samples (i.e., tetB, aac(6')-Iaa,
aph(6)-Id, aph(3'')-Ib, blaCMY-2). In particular, the odds of identifying tet(B) and aph(3’’)-Ib were
highest in livestock samples, and the odds of identifying blaCMY-2 were highest in human samples
(Tables 5.7, 5.8). The odds of identifying blaCMY-2 were also significantly greater in livestock
samples compared to raccoon or environmental samples (Table 5.8). Random intercepts were not
retained in several models since variance components were extremely small (i.e., ColpVC, tetB,
aac6’-Iaa, aph-6-Id, fosA7) or exact logistic regression was used (i.e., blaCMY-2, aph3’’-Ib,
IncFiip96a).
5.5 Discussion
A diversity of serovars and sequence types were isolated from human, livestock, wildlife and
environmental sources in our study region. Using cgMLST, we were able to examine in greater
detail the genetic similarities between Salmonella isolates from these different sources, than has
199
been historically possible using PFGE or legacy MLST typing methods (Alikhan et al., 2018; de
Been et al., 2015). The standardized 3002-loci cgMLST Enterobase scheme for Salmonella
offers a rapid, high-resolution typing method that demonstrates comparable sensitivity to SNP-
based methods for the investigation of outbreaks (Zhou et al., 2020; Alikhan et al., 2018). Our
epidemiological investigation of Salmonella from various sources in southern Ontario using this
cgMLST scheme has provided insights about population structure and potential transmission; the
identification of isolates with identical or near identical cgMLST types in different sources
suggests that either transmission may be occurring between these different sources, or,
alternatively, that these sources were exposed to a common source (not identified here).
Comparison of cgMLST subtypes revealed many instances where Salmonella isolates from
livestock (particularly chickens) and environmental sources were indistinguishable from human
Salmonella isolates, suggesting that transmission may be occurring between humans, livestock,
and the environment in our study region. Based on our 30 loci threshold, none of the raccoon
Salmonella isolates clustered together with isolates from humans, with the (near) exception of a
single Salmonella Heidelberg ST15 isolate from a raccoon that differed only at 4 loci; however,
since this serovar was most common to chickens (accounting for roughly one third of chicken
isolates), this finding most likely represents an infection/colonization in the raccoon that
originated from a livestock source.
The top three serovars responsible for human illness in our study, Enteritidis, Heidelberg,
and Kentucky, accounted for over 65% of all human isolates in this study. The presence of
Salmonella Kentucky among top serovars causing human illness is an unusual finding (Public
Health Agency of Canada, 2009-2013), however, due to biases in the selection of human isolates
200
for sequencing (based on resistance) and, therefore, inclusion in our study, the distribution of the
top human serovars in our study should not be taken as necessarily representative of the most
prevalent serovars causing human illness in this study region and time period. Of note,
Salmonella Kentucky accounted for only 2 of 465 human cases (<1%) in an earlier study of this
same FoodNet sentinel site directly preceding our study period (2006-2011; Flockhart et al.,
2017). Globally, the emergence of multi-drug resistant Salmonella Kentucky ST198 in human
salmonellosis is a growing concern (Hawkey et al., 2019). Previously identified and reported by
Mulvey et al., (2013), all human Salmonella Kentucky isolates in our study were exclusively
multilocus sequence type ST198, and all demonstrated phenotypic resistance to ciprofloxacin.
Apart from human samples, the only other source containing ST198 isolates were chicken
samples (n=4/214), which is consistent with previous reports (Hawkey et al., 2019).
Overall, the top five serovars of Salmonella in human isolates in our study (Enteritidis,
Heidelberg, Kentucky, Typhimurium, Infantis) largely overlap with the top five reported
serovars in humans previously reported by Flockhart et al. (2017) in this sentinel site (Enteritidis,
Typhimurium, Heidelberg, Newport, Thompson), and also at the provincial level (Enteritidis,
Typhimurium, Heidelberg, Typhi, Newport; Public Health Agency of Canada, 2006-2011).
Serovars Kentucky and Heidelberg were identified commonly among chicken isolates, but less
commonly in other sources; these findings are also consistent with the previous work in this
region by Flockhart et al. (2017), which demonstrated that poultry was the most likely source of
infection for human cases for certain serovars (e.g., Enteritidis, Heidelberg), based on phage-
typing and PFGE (Flockhart et al., 2017). No clear animal or environmental source of human
Enteritidis cases was identified in our study but based on their examination of phage types and
201
PFGE fingerprints, Flockhart et al. (2017) suggested that broiler chickens in this study region
likely are a major source of Salmonella Enteritidis for humans.
Analysis of population structure provides insight regarding potential human-raccoon
transmission in this region. Salmonella Newport was the sixth most common serovar causing
illness in humans and was also the most common serovar identified in our raccoon population;
however, the two human Newport isolates were a different sequence type from that in raccoons,
with a large number of allelic differences (>2500 loci) that suggest a different source for these
human cases. Unfortunately, other potential human exposures related to travelling abroad could
not be ruled out since these epidemiological data were unavailable. Other serovars that were
occasionally isolated from raccoons (e.g., Typhimurium, Infantis) were also commonly identified
in livestock and environmental sources, suggesting widespread, and frequent dissemination of
these generalist Salmonella serovars, making it difficult to propose a primary or predominant
source. Other serovars were restricted to certain livestock hosts (Cerro in cattle, Worthington in
swine), with a handful of environmental samples that suggests sporadic environmental
contamination with principally livestock-associated serovars (Cummings et al., 2010; Davies et
al., 1997).
The sampling bias related to inclusion of human isolates on the basis of demonstrated
phenotypic resistance was clearly reflected in the prevalence of phenotypic multi-class resistance
(3+ classes) in the various source types (i.e., 65% of human isolates, and <15% in other sources).
As a result of this sampling bias, human isolates also demonstrated a greater diversity of
acquired resistance genes compared to other sources. Among the remaining sources, which
included both phenotypically pan-susceptible and resistant isolates, the prevalence of multi-class
202
resistance and the diversity of resistance genes identified were highest among livestock isolates,
followed by environmental isolates, and, lastly, raccoons. Although raccoons on swine farms
were frequently infected with Salmonella (29% prevalence, as previously demonstrated by
Bondo et al., 2016), antimicrobial resistance was rare, with only 3/92 isolates demonstrating
phenotypic resistance. Our analyses revealed that the majority of acquired resistance genes and
plasmid replicons examined with logistic regression models were significantly associated with
source type. Antimicrobial resistance genes were generally highest in human and livestock
sources compared with raccoon and environmental samples, which is consistent with the pattern
observed in the distribution of phenotypic multi-class resistance in these different sources. The
odds of detecting certain resistance genes were significantly higher in livestock (e.g., tetB), or
humans (e.g., blaCMY-2) compared to raccoons, which is not surprising, given the low prevalence
of resistance in the raccoon population sampled. The predominant types of resistance identified
in the subset of samples from the previous wildlife study were to aminoglycosides, sulfa drugs,
and tetracyclines, the latter of which is plausibly related to the frequent use of tetracyclines in the
swine industry (Glass-Kaastra et al., 2013). The presence of resistance genes of public health
concern (e.g., blaCMY-2, blaCARB-3, qnrB20, qnrS1) in human, livestock, and environmental water
samples, and their absence in the subset of samples from the wildlife study suggests that
raccoons and the swine farms sampled within the Grand River watershed do not play a
substantial role to the dissemination of high-priority antimicrobial resistant strains of Salmonella
to humans in this region. In contrast, water samples from this watershed, similar to livestock,
contained resistance genes to numerous antimicrobial classes, which supports the theory that
environmental exposures in this region (e.g., the use of recreational areas for swimming) may be
important contributors to resistant human Salmonella cases, especially given the high number of
203
conservation areas located within this watershed (n=11). Salmonella has consistently been
isolated from water sources across Canada, including those within this watershed (Jokinen et al.,
2015; Flockhart et al., 2017; Thomas et al., 2013; Edge et al., 2012), further work is needed;
however, to clarify and assess the potential contribution of water sources to cases of human
salmonellosis.
The prevalence of plasmid incompatibility types in the dataset as a whole also varied by
source. For instance, colRNAI and colpVC were highest in Salmonella from livestock sources,
whereas IncX1-1 was highest in Salmonella from humans. Some plasmid types were restricted to
Salmonella Kentucky (e.g., IncX1-3) and almost exclusively found in broiler chicken samples.
IncFiip96a replicons were only identified in Salmonella from raccoon and soil samples and given
their high prevalence (48%) and the low prevalence of resistance in these sampling sources,
these particular replicons do not appear to be a major source of resistance genes. For this study,
we did not characterize plasmid mobility, so it is unknown whether these incompatibility types
representing plasmids have the potential to be an important factor in the circulation of
antimicrobial resistance genes. The appearance of most incompatibility types in multiple
serovars (including IncFiip96a) is consistent with widespread distribution and/or long-term
maintenance of incompatibility types representing plasmids; future work characterizing plasmid
mobility may help to shed light on this issue.
Limitations
As previously indicated, a major limitation of our study relates to systematic bias in the selection
of human isolates for sequencing as part of public health surveillance programs. Since
sequencing of human isolates in this study was performed primarily for other projects examining
204
AMR, our population of human samples is unlikely to be representative of all human Salmonella
infections in the region. Previous work documented the presence of Salmonella Thompson and
Newport within the top ten serovars causing human illness in Ontario and in our study region
between 2006-2011 (Flockhart et al., 2017; Public Health Agency of Canada 2006-2013).
Without a comprehensive examination of all human Salmonella cases in the region (including
those not reported to public health authorities), it remains to be determined whether milder, non-
resistant Salmonella infections in humans could be related to wildlife and environmental sources.
For instance, Thompson and Newport were both highly prevalent serovars in raccoon and soil
samples from swine farms, have previously been documented in water samples in this study
region (Thomas et al., 2013, Flockhart et al., 2017), and have also consistently been documented
as top serovars causing human illness in this region (Flockhart et al., 2017; Public Health
Agency of Canada 2006-2013). Our collection of resistant human isolates only contained two
Newport and no Thompson isolates; it remains to be seen whether mild human infections with
these serovars-not captured here-may be originating from wildlife or environmental sources.
Associations and prevalence measures involving livestock samples included in this study
may also be impacted by selection bias, since the inclusion and continuing enrollment of farms in
the FoodNet surveillance program was not necessarily random, as previously discussed by
Flockhart et al., (2017). Similarly, the population of raccoons in our study should not be
considered representative of all raccoons in the Grand River watershed, as a result of trapping
animals (Vogt et al., 2020), and focusing on a rural population of raccoons in agricultural areas.
Furthermore, raccoon samples were not plated onto selective media for ESBLs, thus the near
absence of these organisms in the wildlife study subset should be interpreted with some caution.
205
In silico determination of acquired AMR genes appears to be highly concordant with
phenotypic testing results for the antimicrobial drug classes assessed. Since the identification of
chromosomal resistance genes and expression was beyond the scope of our study, we were
unable to assess test sensitivity and specificity for drug classes for which these mechanisms
figure prominently (e.g., quinolones, aminoglycosides; Chen et al., 2007; Vetting et al., 2003).
Additional limitations relate to plasmid determination using incompatibility types, and lack of
characterization of plasmid mobility. Previous work has demonstrated that plasmid distribution
varies by serovar (Pornsukarom et al., 2018), however, the aim of our work was to characterize
patterns in the occurrence of genes and incompatibility types which reflect plasmids in different
sources to address transmission and propose primary sources. Due to the sheer number of
serovars with few isolates that prevented many models from converging when serovar was
included as a random intercept (data not shown), we were unable to control for the potential
confounding effect of serovar on the occurrence of different plasmid types or resistance genes in
different sources, which may be a particularly important factor for plasmids that are non-mobile
or that have limited mobility.
Finally, our use of “livestock” as a broad category in statistical models to capture
chicken, cattle, and swine samples could not identify species-specific differences. However, this
study provides a preliminary examination of transmission between broad groups of animals, and
the lumping of livestock into one category avoided potential misclassification of samples
associated with the surveillance-collected fecal samples from multi-species farms. Among these
surveillance-collected samples, fecal samples may have been obtained fresh, or from standing
pooled manure piles; in the future, efforts focused on the impact of these sampling methods will
206
be important for the utilization and interpretation of heterogeneous sampling methods and
sources.
5.6 Conclusions
This study offers new insights about the epidemiology of Salmonella and associated AMR from
human, animal, and environmental sources in a populous region of southern Ontario based on
whole-genome sequencing data. Overall, raccoons sampled on swine farms in this study were
rarely infected with internationally recognized Salmonella lineages, or with antimicrobial
resistant isolates (<5%). Conversely, resistance was more commonly identified in livestock and
water samples, and a number of resistance genes of public health importance (e.g., blaCMY-2,
blaCARB-3, qnrS1, qnrB20) were identified in some of these sources. Although other studies have
demonstrated strong links between Salmonella isolates from wildlife and humans (Toro et al.,
2016; Ford et al., 2019; Collins et al., 2019), our assessments of microbial population structure
suggest that raccoons are an unlikely source of Salmonella for resistant human infections in this
study region. However, on the basis of similar or identical cgMLST types, transmission of
resistant Salmonella from livestock and environmental sources to humans may be occurring. As
several researchers have recently highlighted, intervention-based, hypothesis-testing studies that
combine genomic data with epidemiological approaches are needed to confirm observational
data which purport livestock as a primary source of pathogens and AMR determinants for
humans, wildlife, and the environment (Wee et al., 2020, Ramey & Ahlstrom, 2020). Further
research, including a comprehensive examination of all human cases (including phenotypically
207
susceptible ones), and examination of additional raccoon populations (both urban and rural) may
help to provide additional insight on the role of raccoons in human Salmonella infections.
208
5.7 References
Ahlstrom, C. A., Bonnedahl, J., Woksepp, H., Hernandez, J., Reed, J. A., Tibbitts, L., Olsen, B., Douglas, D. C., & Ramey, A. M. (2019). Satellite tracking of gulls and genomic characterization of faecal bacteria reveals environmentally mediated acquisition and dispersal of antimicrobial-resistant Escherichia coli on the Kenai Peninsula, Alaska. Molecular Ecology, 28(10), 2531–2545.
Ahlstrom, C. A., Bonnedahl, J., Woksepp, H., Hernandez, J., Olsen, B., & Ramey, A. M. (2018). Acquisition and dissemination of cephalosporin-resistant E. coli in migratory birds sampled at an Alaska landfill as inferred through genomic analysis. Scientific Reports, 8(1), 7361.
Alcock, B. P., Raphenya, A. R., Lau, T., Tsang, K. K., Bouchard, M., Edalatmand, A., Huynh, W., Nguyen, A. V., Cheng, A. A., Liu, S., Min, S. Y., Miroshnichenko, A., Tran, H. K., Werfalli, R. E., Nasir, J. A., Oloni, M., Speicher, D. J., Florescu, A., Singh, B., Faltyn, M., … McArthur, A. G. (2020). CARD 2020: Antibiotic resistome surveillance with the comprehensive antibiotic resistance database. Nucleic Acids Research, 48(D1), D517–D525.
Alikhan, N. F., Zhou, Z., Sergeant, M. J., & Achtman, M. (2018). A genomic overview of the population structure of Salmonella. PLoS genetics, 14(4), e1007261.
Aung, K. T., Chen, H. J., Chau, M. L., Yap, G., Lim, X. F., Humaidi, M., Chua, C., Yeo, G., Yap, H. M., Oh, J. Q., Manogaran, V., Hapuarachchi, H. C., Maiwald, M., Tee, N., Barkham, T., Koh, T. H., Gutiérrez, R. A., Schlundt, J., & Ng, L. C. (2019). Salmonella in retail food and wild birds in Singapore-prevalence, antimicrobial resistance, and sequence types. International Journal of Environmental Research and Public Health, 16(21), 4235.
Baldi, M., Barquero Calvo, E., Hutter, S. E., & Walzer, C. (2019). Salmonellosis detection and evidence of antibiotic resistance in an urban raccoon population in a highly populated area, Costa Rica. Zoonoses and Public Health, 66(7), 852–860.
Bawn, M., Alikhan, N. F., Thilliez, G., Kirkwood, M., Wheeler, N. E., Petrovska, L., Dallman, T. J., Adriaenssens, E. M., Hall, N., & Kingsley, R. A. (2020). Evolution of Salmonella enterica serotype Typhimurium driven by anthropogenic selection and niche adaptation. PLoS Genetics, 16(6), e1008850.
Bloomfield, S. J., Benschop, J., Biggs, P. J., Marshall, J. C., Hayman, D., Carter, P. E., Midwinter, A. C., Mather, A. E., & French, N. P. (2017). Genomic analysis of Salmonella enterica serovar Typhimurium DT160 associated with a 14-Year outbreak, New Zealand, 1998-2012. Emerging Infectious Diseases, 23(6), 906–913.
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016). Impact of season, demographic and environmental factors on Salmonella
209
occurrence in raccoons (Procyon lotor) from swine farms and conservation areas in southern Ontario. PLoS ONE, 11(9), e0161497.
Carattoli, A., Zankari, E., García-Fernández, A., Voldby Larsen, M., Lund, O., Villa, L., Møller Aarestrup, F., & Hasman, H. (2014). In silico detection and typing of plasmids using PlasmidFinder and plasmid multilocus sequence typing. Antimicrobial Agents and Chemotherapy, 58(7), 3895–3903.
Chen, S., Cui, S., McDermott, P. F., Zhao, S., White, D. G., Paulsen, I., & Meng, J. (2007). Contribution of target gene mutations and efflux to decreased susceptibility of Salmonella enterica serovar Typhimurium to fluoroquinolones and other antimicrobials. Antimicrobial Agents and Chemotherapy, 51(2), 535–542.
Collins, J., Simpson, K. M. J., Bell, G., Durrheim, D. N., Hill-Cawthorne, G. A., Hope, K., … Flint, J. (2019). A one health investigation of Salmonella enterica serovar Wangata in north-eastern New South Wales, Australia, 2016–2017. Epidemiology and Infection, 147, e150.
Compton, J. A., Baney, J. A., Donaldson, S. C., Houser, B. A., San Julian, G. J., Yahner, R. H., Chmielecki, W., Reynolds, S., & Jayarao, B. M. (2008). Salmonella infections in the common raccoon (Procyon lotor) in western Pennsylvania. Journal of Clinical Microbiology, 46(9), 3084–3086.
Croucher, N. J., Harris, S. R., Grad, Y. H., & Hanage, W. P. (2013). Bacterial genomes in epidemiology--Present and future. Philosophical Transactions of the Royal Society of London. Series B, Biological sciences, 368(1614), 20120202.
Cummings, K. J., Warnick, L. D., Elton, M., Rodriguez-Rivera, L. D., Siler, J. D., Wright, E. M., Gröhn, Y. T., & Wiedmann, M. (2010). Salmonella enterica serotype Cerro among dairy cattle in New York: An emerging pathogen?. Foodborne Pathogens and Disease, 7(6), 659–665.
Davies, P. R., Morrow, W. E., Jones, F. T., Deen, J., Fedorka-Cray, P. J., & Harris, I. T. (1997). Prevalence of Salmonella in finishing swine raised in different production systems in North Carolina, USA. Epidemiology and Infection, 119(2), 237–244.
de Been, M., Pinholt, M., Top, J., Bletz, S., Mellmann, A., van Schaik, W., Brouwer, E., Rogers, M., Kraat, Y., Bonten, M., Corander, J., Westh, H., Harmsen, D., & Willems, R. J. (2015). Core genome multilocus sequence typing scheme for high-resolution typing of Enterococcus faecium. Journal of Clinical Microbiology, 53(12), 3788–3797.
De Lucia, A., Rabie, A., Smith, R. P., Davies, R., Ostanello, F., Ajayi, D., Petrovska, L., & Martelli, F. (2018). Role of wild birds and environmental contamination in the epidemiology of Salmonella infection in an outdoor pig farm. Veterinary Microbiology, 227, 148–154.
210
Edge, T. A., El-Shaarawi, A., Gannon, V., Jokinen, C., Kent, R., Khan, I. U., Koning, W., Lapen, D., Miller, J., Neumann, N., Phillips, R., Robertson, W., Schreier, H., Scott, A., Shtepani, I., Topp, E., Wilkes, G., & van Bochove, E. (2012). Investigation of an Escherichia coli environmental benchmark for waterborne pathogens in agricultural watersheds in Canada. Journal of Environmental Quality, 41(1), 21–30.
Flockhart, L., Pintar, K., Cook, A., McEwen, S., Friendship, R., Kelton, D., & Pollari, F. (2017). Distribution of Salmonella in humans, production animal operations and a watershed in a FoodNet Canada sentinel site. Zoonoses and Public Health, 64(1), 41–52.
Ford, L., Ingle, D., Glass, K., Veitch, M., Williamson, D. A., Harlock, M., … Kirk, M. D. (2019). Whole-genome sequencing of Salmonella Mississippi and Typhimurium Definitive Type 160, Australia and New Zealand. Emerging Infectious Diseases, 25(9), 1690–1697.
GBD 2017 Non-Typhoidal Salmonella Invasive Disease Collaborators (2019). The global burden of non-typhoidal Salmonella invasive disease: A systematic analysis for the Global Burden of Disease Study 2017. The Lancet Infectious Diseases, 19(12), 1312–1324.
Glass-Kaastra, S. K., Pearl, D. L., Reid-Smith, R. J., McEwen, B., McEwen, S. A., Amezcua, R., & Friendship, R. M. (2013). Describing antimicrobial use and reported treatment efficacy in Ontario swine using the Ontario Swine Veterinary-based Surveillance program. BMC Veterinary Research, 9, 238.
Government of Canada (2015). Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS) 2015 annual report. Guelph, ON: Public Health Agency of Canada.
Guyomard-Rabenirina, S., Reynaud, Y., Pot, M., Albina, E., Couvin, D., Ducat, C., Gruel, G., Ferdinand, S., Legreneur, P., Le Hello, S., Malpote, E., Sadikalay, S., Talarmin, A., & Breurec, S. (2020). Antimicrobial resistance in wildlife in Guadeloupe (French West Indies): Distribution of a single blaCTX-M-1/IncI1/ST3 plasmid among humans and wild animals. Frontiers in Microbiology, 11, 1524.
Guyomard-Rabenirina, S., Weill, F. X., Le Hello, S., Bastian, S., Berger, F., Ferdinand, S., Legreneur, P., Loraux, C., Malpote, E., Muanza, B., Richard, V., Talarmin, A., & Breurec, S. (2019). Reptiles in Guadeloupe (French West Indies) are a reservoir of major human Salmonella enterica serovars. PloS One, 14(7), e0220145.
Hawkey, J., Le Hello, S., Doublet, B., Granier, S. A., Hendriksen, R. S., Fricke, W. F., Ceyssens, P. J., Gomart, C., Billman-Jacobe, H., Holt, K. E., & Weill, F. X. (2019). Global phylogenomics of multidrug-resistant Salmonella enterica serotype Kentucky ST198. Microbial Genomics, 5(7), e000269.
Jia, B., Raphenya, A. R., Alcock, B., Waglechner, N., Guo, P., Tsang, K. K., Lago, B. A., Dave, B. M., Pereira, S., Sharma, A. N., Doshi, S., Courtot, M., Lo, R., Williams, L. E., Frye, J. G., Elsayegh, T., Sardar, D., Westman, E. L., Pawlowski, A. C., Johnson, T. A., …
211
McArthur, A. G. (2017). CARD 2017: Expansion and model-centric curation of the comprehensive antibiotic resistance database. Nucleic Acids Research, 45(D1), D566–D573.
Jokinen, C. C., Koot, J., Cole, L., Desruisseau, A., Edge, T. A., Khan, I. U., Koning, W., Lapen, D. R., Pintar, K. D., Reid-Smith, R., Thomas, J. L., Topp, E., Wang, L. Y., Wilkes, G., Ziebell, K., van Bochove, E., & Gannon, V. P. (2015). The distribution of Salmonella enterica serovars and subtypes in surface water from five agricultural regions across Canada. Water Research, 76, 120–131.
Kern, W. V., & Rieg, S. (2020). Burden of bacterial bloodstream infection-A brief update on epidemiology and significance of multidrug-resistant pathogens. Clinical Microbiology and Infection, 26(2), 151–157.
Kumar, S. B., Arnipalli, S. R., & Ziouzenkova, O. (2020). Antibiotics in food chain: The consequences for antibiotic resistance. Antibiotics (Basel, Switzerland), 9(10), E688.
Lee, K., Iwata, T., Nakadai, A., Kato, T., Hayama, S., Taniguchi, T., & Hayashidani, H. (2011). Prevalence of Salmonella, Yersinia and Campylobacter spp. in feral raccoons (Procyon lotor) and masked palm civets (Paguma larvata) in Japan. Zoonoses and Public Health, 58(6), 424–431.
Majowicz, S. E., Musto, J., Scallan, E., Angulo, F. J., Kirk, M., O'Brien, S. J., Jones, T. F., Fazil, A., Hoekstra, R. M., & International Collaboration on Enteric Disease 'Burden of Illness' Studies (2010). The global burden of nontyphoidal Salmonella gastroenteritis. Clinical Infectious Diseases, 50(6), 882–889.
Mather, A. E., Lawson, B., de Pinna, E., Wigley, P., Parkhill, J., Thomson, N. R., Page, A. J., Holmes, M. A., & Paterson, G. K. (2016). Genomic analysis of Salmonella enterica serovar Typhimurium from wild passerines in England and Wales. Applied and Environmental Microbiology, 82(22), 6728–6735.
Michael, G. B., & Schwarz, S. (2016). Antimicrobial resistance in zoonotic nontyphoidal Salmonella: An alarming trend?. Clinical Microbiology and Infection, 22(12), 968–974.
Mulvey, M. R., Boyd, D. A., Finley, R., Fakharuddin, K., Langner, S., Allen, V., Ang, L., Bekal, S., El Bailey, S., Haldane, D., Hoang, L., Horsman, G., Louis, M., Robberts, L., & Wylie, J. (2013). Ciprofloxacin-resistant Salmonella enterica serovar Kentucky in Canada. Emerging Infectious Diseases, 19(6), 999–1001.
Munck, N., Smith, J., Bates, J., Glass, K., Hald, T., & Kirk, M. D. (2020). Source attribution of Salmonella in macadamia nuts to animal and environmental reservoirs in Queensland, Australia. Foodborne Pathogens and Disease, 17(5), 357–364.
212
Ondov, B. D., Starrett, G. J., Sappington, A., Kostic, A., Koren, S., Buck, C. B., & Phillippy, A. M. (2019). Mash Screen: High-throughput sequence containment estimation for genome discovery. Genome Biology, 20(1), 232.
Pornsukarom, S., van Vliet, A., & Thakur, S. (2018). Whole genome sequencing analysis of multiple Salmonella serovars provides insights into phylogenetic relatedness, antimicrobial resistance, and virulence markers across humans, food animals and agriculture environmental sources. BMC Genomics, 19(1), 801.
Public Health Agency of Canada, 2006–2013: National Enteric Surveillance Program (NESP) Annual Summaries 2006-2013. Government of Canada.
Ramey, A. M., & Ahlstrom, C. A. (2020). Antibiotic resistant bacteria in wildlife: Perspectives on trends, acquisition and dissemination, data gaps, and future directions. Journal of Wildlife Diseases, 56(1):1-15.
Rainwater, K. L., Marchese, K., Slavinski, S., Humberg, L. A., Dubovi, E. J., Jarvis, J. A., McAloose, D., & Calle, P. P. (2017). Health survey of free-ranging raccoons (Procyon lotor) in Central Park, New York, USA: Implications for human and domestic animal health. Journal of Wildlife Diseases, 53(2), 272–284.
Tanwar, J., Das, S., Fatima, Z. & Hameed S. (2014). Multidrug resistance: An emerging crisis. Interdisciplinary Perspectives on Infectious Diseases, 541340, 1-7.
Thomas, J. L., Slawson, R. M., & Taylor, W. D. (2013). Salmonella serotype diversity and seasonality in urban and rural streams. Journal of Applied Microbiology, 114(3), 907–922.
Toro, M., Retamal, P., Ayers, S., Barreto, M., Allard, M., Brown, E. W., & Gonzalez-Escalona, N. (2016). Whole-genome sequencing analysis of Salmonella enterica serovar Enteritidis isolates in Chile provides insights into possible transmission between gulls, poultry, and humans. Applied and Environmental Microbiology, 82(20), 6223–6232.
Torres, R. T., Carvalho, J., Cunha, M. V., Serrano, E., Palmeira, J. D., & Fonseca, C. (2020). Temporal and geographical research trends of antimicrobial resistance in wildlife - A bibliometric analysis. One Health, 11, 100198.
Vetting, M., Roderick, S. L., Hegde, S., Magnet, S., & Blanchard, J. S. (2003). What can structure tell us about in vivo function? The case of aminoglycoside-resistance genes. Biochemical Society Transactions, 31(Pt 3), 520–522.
Very, K. J., Kirchner, M. K., Shariat, N., Cottrell, W., Sandt, C. H., Dudley, E. G., Kariyawasam, S., & Jayarao, B. M. (2016). Prevalence and spatial distribution of Salmonella infections in the Pennsylvania raccoon (Procyon lotor). Zoonoses and Public Health, 63(3), 223–233.
213
Wee, B. A., Muloi, D. M., & van Bunnik, B. (2020). Quantifying the transmission of antimicrobial resistance at the human and livestock interface with genomics. Clinical Microbiology and Infection, S1198-743X(20)30571-1. [Epub 22 Sept 2020].
World Health Organization. (2015). "WHO estimates of the global burden of foodborne diseases". Foodborne disease burden epidemiology reference group, 2007-2015. Available from: https://apps.who.int/iris/bitstream/handle/10665/199350/9789241565165_eng.pdf?sequence=1 (accessed on 10 August 2020).
Zhou, Z., Alikhan, N. F., Sergeant, M. J., Luhmann, N., Vaz, C., Francisco, A. P., Carriço, J. A., & Achtman, M. (2018). GrapeTree: Visualization of core genomic relationships among 100,000 bacterial pathogens. Genome Research, 28(9), 1395–1404.
Zhou, Z., Alikhan, N. F., Mohamed, K., Fan, Y., Agama Study Group, & Achtman, M. (2020). The EnteroBase user's guide, with case studies on Salmonella transmissions, Yersinia pestis phylogeny, and Escherichia core genomic diversity. Genome Research, 30(1), 138–152.
214
5.8 Tables
Table 5.1: Distribution of Salmonella enterica serovarsa from raccoons, humans, livestock, and environmental
sources in southern Ontario, 2011-2013 (n=608)
Serovar Source type Totald
(%) Raccoon (n=92)
Human (n=58)
Environ-mentalb
(n=129)
Cattle (n=60)
Swinec
(n=55) Chicken (n=214)
Kentucky 0 11 3 4 0 114 132 (21.7) Heidelberg 1 14 4 8 0 72 99 (16.3) Newport 33 2 20 0 0 0 55 (9.0) Typhimurium 7 6 16 4 5 1 39 (6.4) Agona 16 0 11 1 9 1 38 (6.2) Infantis 9 4 12 1 6 1 33 (5.4) Enteritidis 1 14 2 3 0 10 30 (4.9) Cerro 0 0 3 19 0 0 22 (3.6) Livingstone 0 0 1 0 10 8 19 (3.1) Thompson 6 0 10 0 0 0 16 (2.6) Worthington 0 0 0 1 12 0 13 (2.1) Give 0 0 8 4 0 0 12 (2.0) Paratyphi B var. Java 9 0 2 0 0 0 11 (1.8)
Schwarzengrund 1 0 4 0 2 0 7 (1.1) Derby 0 0 2 1 3 0 6 (0.9) Uganda 0 0 0 5 1 0 6 (0.9) Poona 1 0 2 0 1 1 5 (0.8) Anatum 0 2 1 0 0 2 5 (0.8) Oranienburg 1 0 1 3 0 0 5 (0.8) Mbandaka 0 0 3 0 2 0 5 (0.8) Hartford 2 0 3 0 0 0 5 (0.8) Hadar 1 0 3 1 0 0 5 (0.8)
a Determined in silico using SISTR. b Includes beach and water samples obtained through FoodNet Canada surveillance, as well as soil samples obtained
from a wildlife study (Bondo et al., 2016). c Includes swine fecal and manure samples collected through FoodNet Canada surveillance, as well as swine manure
pit samples obtained from a wildlife study (Bondo et al., 2016). d Other serovars identified with fewer than 5 samples were: Braenderup (n=4), London (n=3), Montevideo (n=3),
Ohio (n=3), Senftenberg (n=2), Kiambu (n=2), Dublin (n=2), Brandenburg (n=2), Litchfield (n=2), Orion (n=2), IIIb
11:k:z53 (n=2), Mikawasima (n=2), I 1,4,[5],12:i:- (n=1), Tennessee (n=1), Berta (n=1), I 1,4,[5],12:b:- (n=1),
Molade (n=1), Saintpaul (n=1), Rissen (n=1), Stanley (n=1), Holcomb (n=1), Bovismorbificans (n=1), Chester
(n=1).
215
Table 5.2: Counts of most frequently identified sequence types of Salmonella enterica isolates from raccoons, livestock, and environmental sources in
southern Ontario, 2011-2013 (n=608)
Counts of most common sequence types (all sources)
Source type Total no. of isolates collected (n)
ST15
2 (K
entu
cky)
ST15
(Hei
delb
erg)
ST35
0 (N
ewpo
rt)
ST13
(Ago
na)
ST19
(Typ
him
uriu
m)
ST32
(Inf
antis
)
ST11
(Ent
eriti
dis)
ST36
7 (C
erro
)
ST63
8 (L
ivin
gsto
ne)
ST26
(Tho
mps
on)
ST19
8 (K
entu
cky)
Human 58 0 13 0 0 1 3 14 0 0 0 11
Wildlife Raccoon 92 0 1 31 16 7 9 1 0 0 6 0
Livestock
Cattle
Swine
Chicken
60 4 8 0 1 4 1 3 19 0 0 0
55 0 0 0 9 5 6 0 0 10 0 0
214 109 69 0 1 1 1 10 0 8 0 4
Environmental
Soil
Water
Beach
45 0 0 11 8 6 5 1 0 1 4 0
76 3 4 5 3 7 5 1 3 0 5 0
8 0 0 1 0 2 2 0 0 0 1 0
Total (n) 608 116 95 48 38 33 32 30 22 19 16 15
216
Table 5.3: Distribution of Salmonella enterica legacy multi-locus sequence types for isolates obtained from
raccoons, humans, livestock, and environmental sources in southern Ontario, 2011-2013 (n=608)
Sequence typea (Serovar) Source type
Totald (%) Raccoon
(n=92) Human (n=58)
Environ-mentalb (n=129)
Cattle (n=60)
Swinec (n=55)
Chicken (n=214)
ST152 (Kentucky) 0 0 3 4 0 109 116 (19.1) ST15 (Heidelberg) 1 13 4 8 0 69 95 (15.6) ST350 (Newport) 31 0 17 0 0 0 48 (7.9) ST13 (Agona) 16 0 11 1 9 1 38 (6.3) ST19 (Typhimurium)
7 1 15 4 5 1 33 (5.4)
ST32 (Infantis) 9 3 12 1 6 1 32 (5.3) ST11(Enteritidis) 1 14 2 3 0 10 30 (4.9) ST367 (Cerro) 0 0 3 19 0 0 22 (3.6) ST638 (Livingstone)
0 0 1 0 10 8 19 (3.1)
ST26 (Thompson) 6 0 10 0 0 0 16 (2.6) ST198 (Kentucky) 0 11 0 0 0 4 15 (2.5) ST592 (Worthington)
0 0 0 1 12 0 13 (2.1)
ST654 (Give) 0 0 8 4 0 0 12 (2.0) ST404 (Paratyphi B var. Java)
9 0 2 0 0 0 11 (1.8)
ST96 (Schwarzengrund)
1 0 4 0 2 0 7 (1.1)
ST684 (Uganda) 0 0 0 5 1 0 6 (1.0) ST10 (Dublin) 0 2 0 0 0 0 2 (0.3) ST45 (Newport) 0 1 1 0 0 0 2 (0.3) ST65 (Brandenburg)
0 0 0 0 2 0 2 (0.3)
Bolded STs represent internationally recognized sequence types. a Sequence types (STs) determined using 7-loci Achtman scheme. b Includes beach and water samples obtained through FoodNet Canada surveillance, as well as soil samples obtained
from a wildlife study (Bondo et al., 2016). c Includes swine fecal and manure samples collected through FoodNet Canada surveillance, as well as swine manure
pit samples obtained from a wildlife study (Bondo et al., 2016). d Other STs identified within 5 or fewer samples were: ST11 (Enteritidis; n=2), ST14 (Seftenburg; n=2), ST118
(Newport; n=1), ST138 (Montevideo; n=3), ST150 (Bovismorbificans; n=1), ST155(London; n=3), ST1565
(Tennessee; n=1), ST163 (Newport; n=1), ST1675 (Oranienburg; n=1), ST1838 (I 1,4,[5],12:b:-, n=1), ST2076
(Typhimurium, n=2), ST214 (Litchfield, n=2), ST22 (Braenderup; n=4), ST23 (Oranienburg; n=4), ST2507
(Holcomb; n=1), ST27 (Saintpaul; n=1), ST2848 (IIIb 11:k:z53; n=2), ST29 (Stanley; n=1), ST2937 (Infantis; n=1),
ST308 (Poona; n=5), ST309 (Kiambu; n=2), ST329 (Ohio; n=3), ST33 (Hadar; n=5), ST34 (Typhimurium; n=4),
ST343 (Chester; n=1), ST36 (Typhimurium; n=1), ST3790 (Heidelberg; n=4), ST40 (Derby; n=4), ST405 (Hartford;
217
n=5), ST413 (Mbandaka; n=5), ST435 (Berta; n=1), ST469 (Rissen; n=1), ST544 (Molade; n=1), ST639 (Orion;
n=2), ST64 (Anatum; 5), ST72 (Derby; n=1). e Seven isolates (raccoon-, soil-, water-, swine-, chicken-) are not included as they were not typeable.
218
Table 5.4: Frequencies of antimicrobial resistance genes identified from whole-genome sequencing data of Salmonella enterica isolates from raccoons,
humans, livestock, and environmental sources in southern Ontario, Canada, 2011-2013 (n=608)
Antimicrobial group Resistance gene ARO†
Human (n=58)
Livestock (n=329)
Raccoon (n=92)
Environment (n=129)
Total (%)
Aminoglycoside aac(6')-Ib-cr* 3002547 16 0 0 0 16 (2.6%)
aac(6')-Ib8 3002579 1 0 0 0 1 (0.2%)
aac(6')-Iaa 3002571 17 35 8 22 82 (13.5%)
aac(3)-Id 3002529 10 0 0 0 10 (1.6%)
aac(3)-IId 3004623 3 1 0 1 5 (0.8%)
aac(3)-IV 3002539 5 1 0 0 6 (1.0%)
aac(3)-VIa 3002540 1 1 0 0 2 (0.3%)
aadA1 3002601 2 0 0 0 2 (0.33%)
aadA2 3002602 1 2 0 3 6 (1.0%)
aadA3 3002603 2 0 0 0 2 (0.3%)
aadA4 3002604 0 5 1 0 6 (1.0%)
aadA7 3002607 10 0 0 0 10 (1.6%)
aadA16 3002616 13 0 0 0 13 (2.1%)
ant(2'')-Ia 3000230 2 1 0 0 3 (0.5%)
ant(3'')-IIa 3004089 7 12 1 4 24 (3.9%)
aph(6)-Id 3002660 9 72 1 6 88 (14.5%)
aph(3')-Ia 3002641 4 6 0 1 11 (1.8%)
aph(3")-Ib 3002639 5 71 0 3 79 (13.0%)
aph(4)-Ia 3002655 5 1 0 0 6 (1.0%) Beta-lactam blaCMY-2 3002013 17 57 0 3 77 (12.7%) blaCTX-M-14 3001877 2 0 0 0 2 (0.3%) blaCTX-M-65 3001926 3 0 0 0 3 (0.5%) blaTEM-1 3000873 14 15 0 4 33 (5.4%) blaCARB-3 3002242 0 5 0 2 7 (1.2%) blaDHA-1 3002132 1 0 0 0 1 (0.2%) blaOXA-1 3001396 3 0 0 0 3 (0.5%) Lincosamide linG 3002879 1 0 0 0 1 (0.2%)
219
lnuG 3004085 0 3 0 0 3 (0.5%) Macrolide mphA 3000316 1 0 0 0 1 (0.2%) Nucleoside SAT-2 3002895 1 1 0 0 2 (0.3%) Folate pathway inhibitors dfrI 3004645 0 3 0 0 3 (0.5%) dfrA1 3002854 1 0 0 0 1 (0.2%) drfA12 3002858 1 0 0 1 2 (0.3%) dfrA14 3002859 5 0 0 0 5 (0.8%) dfrA23 3003019 1 0 0 0 1 (0.2%) dfrA25 3003020 0 0 0 2 2 (0.3%) dfrA27 3004550 13 0 0 0 13 (2.1%) sul1 3000410 31 11 2 4 48 (7.9%) sul2 3000412 10 2 0 1 13 (2.1%) sul3 3000413 2 3 0 1 6 (1.0%) Phenicol floR 3002705 13 7 0 3 23 (3.8%) catB3 3002676 3 0 0 0 3 (0.5%) cmlA1 3002693 2 0 0 1 3 (0.5%) cmlA5 3002695 1 0 0 1 2 (0.3%) catII from
Escherichia coli K12
3004656 1 0 0 0 1 (0.2%)
Polymyxin mcr1.1 3003689 1 0 0 0 1 (0.2%) Quinolone qnrA1 3002707 1 0 0 0 1 (0.2%) qnrB4 3002718 1 0 0 0 1 (0.2%) qnrB6 3002720 13 0 0 0 13 (2.1%) qnrB19 3002734 1 0 0 0 1 (0.2%) qnrB20 3002735 0 0 0 2 2 (0.3%) qnrS1 3002790 5 0 0 1 6 (1.0%) qacH 3003836 2 0 0 1 3 (0.5%) oqxA** 3003922 3 0 0 0 3 (0.5%) oqxB** 3003923 2 0 0 0 2 (0.3%) adeF** 3000777 1 0 0 0 1 (0.2%) Rifamycin arr-3 3002848 16 0 0 0 16 (2.6%)
220
Fosfomycin fosA3 3002872 2 0 0 0 2 (0.3%) fosA7 3004113 14 94 18 17 143 (23.5%) mdtG 3001329 1 2 1 4 8 (1.3%) Tetracycline tet(A) 3000165 34 12 2 9 57 (9.4%) tet(B) 3000166 4 75 1 2 82 (13.5%) tet(C) 3000167 0 10 0 0 10 (1.6%) tet(D) 3000168 2 5 0 2 9 (1.5%) tet(M) 3000186 2 1 0 1 4 (0.6%)
† Antibiotic Resistance Ontology Accession number as listed in the Comprehensive Antibiotic Resistance Database.
*Gene may also confer fluoroquinolone resistance.
**Gene may also confer resistance to tetracyclines.
221
Table 5.5: Distribution of plasmid replicon types identified using whole-genome sequencing data by source
type for Salmonella enterica isolates from raccoons, humans, livestock, and environmental sources in
southern Ontario, Canada, 2011-2013 (n=608)
Rep-type Source type Totalc (%)
Raccoon (n=92)
Human (n=58)
Environmentala (n=129)
Cattle (n=60)
Swineb (n=55)
Chicken (n=214)
IncX1-1 22 28 11 9 0 80 150 (24.7) IncI1-Igamma 2 14 11 28 3 75 133 (21.9) IncFIIS 46 15 35 8 4 11 119 (19.6) IncX1-3 0 0 2 2 0 101 105 (17.3) ColRNAI 4 10 10 19 5 47 95 (15.6) ColpVC 1 6 13 22 2 36 80 (13.2) IncFiip96a 44 0 26 2 0 0 72 (11.8) IncFIB(AP001918) 0 1 2 1 1 64 69 (11.3) ColpHAD28 4 5 10 11 15 15 60 (9.9) IncFIB(S) 2 14 9 6 4 10 45 (7.4) IncX3 21 1 6 1 0 8 37 (6.1) ColYe4449 15 0 9 1 9 0 34 (5.6) Col156 1 1 3 15 1 3 24 (3.9) IncN1 0 14 2 0 0 1 17 (2.8) Col440II 2 4 4 1 1 3 15 (2.5) Col8282 0 4 3 1 0 4 12 (2.0) Col440I 2 0 2 0 2 5 11 (1.8) Pkpccav1321 1 4 0 0 0 5 10 (1.6) ColE10 0 0 0 0 10 0 10 (1.6)
a Includes beach and water samples obtained through FoodNet Canada surveillance, as well as soil samples obtained
from a wildlife study (Bondo et al., 2016). b Includes swine fecal and manure samples collected through FoodNet Canada surveillance, as well as swine manure
pit samples obtained from a wildlife study (Bondo et al., 2016). c Plasmid replicon types identified in fewer than 10 samples included: IncH(1a) (n=2), IncHI2 (n=9), IncC (n=7),
IncI2(delta) (n=6), IncFIB (pN55391) (n=4), IncQ2 (n=4), IncFII(pHN7a8) (n=3), IncFII (n=3), IncFII(p14) (n=2),
IncY1 (n=3), Col(MP18) (n=2), IncI1a (n=2), IncHI1(bR27) (n=2), IncFIA(HI1) (n=2), IncFIB(pB171) (n=2),
IncX4 (n=1), IncN2 (n=1), Col(MG828) (n=1), IncFICFII (n=1), IncI(gamma) (n=1), IncFIA (n=1), pENTAS02
(n=1), Incx4(2) (n=1), IncFIB(pHCM2) (n=1).
222
Table 5.6: Test sensitivity and specificitya for in silico identification of antimicrobial resistance genes in
Salmonella enterica isolates from raccoons, humans, livestock, and environmental sources in southern
Ontario, 2011-2013 (n=608)
Antimicrobial class Test sensitivity (95%CI) Test specificity (95% CI)
Beta-lactam 97.4% (92.7-99.4%) 99.6% (98.5-99.9%)
Macrolide 100% (0.03-100%)* 100% (99.4-100%)*
Sulfonamide 89.1% (76.4-96.4%) 96.6% (94.7-97.9%)
Phenicol 100% (83.1-100%)* 99.6% (98.8-99.9%)
Tetracycline 98.0% (94.4-99.6%) 99.5% (98.4-99.9%)
Overallc 94.3% (94.2-98.4%) 99.1% (98.6-99.4%) a Phenotypic antimicrobial resistance test results were considered the gold standard. Detection of 15 antimicrobials
performed using the CMV3AGNF panel from National Antimicrobial Resistance Monitoring System (NARMS;
Sensititre, Thermo Scientific). In silico acquired resistance genes detected using CARD-Resistance Gene Identifier. c Raw counts for all samples and antimicrobials were pooled together.
*One-sided 97.5% confidence interval.
223
Table 5.7: Logistic regression modelsa,c,d assessing the association between source type and the occurrence of
select antimicrobial resistance genes and plasmid replicons in Salmonella enterica isolates from raccoons,
humans, livestock, and environmental sources in southern Ontario, 2011-2013 (n=608)
IncX1-1a IncI1-Igammaa IncFIISa
Source type OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Human REF 0.019 (global) REF 0.092
(global) REF <0.001 (global)
Livestock 0.21 (0.06-0.79) 0.021 1.93
(0.63-5.90) 0.249 0.20 (0.08-0.50) 0.001
Raccoon 0.15 (0.03-0.76) 0.022 0.03
(0.00-0.60) 0.021 3.14 (1.22-8.09) 0.018
Environment 0.02 (0.00-0.24) 0.002 0.18
(0.03-0.98) 0.047 1.07 (0.50-2.30) 0.856
ColRNAIa ColpVCb,d IncFiip96ac
Source type OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Human REF 0.010 (global) REF <0.001
(global) REF <0.001 (global)
Livestock 1.86 (0.44-7.80) 0.398 2.32
(0.94-5.75) 0.068 0.42* (0.03-∞) 0.999
Raccoon 0.10 (0.01-0.78) 0.029 0.10
(0.01-0.85) 0.035 73.63* (12.72-∞) <0.001
Environment 0.22 (0.04-1.23) 0.085 0.96
(0.34-2.70) 0.943 20.46* (3.50-∞) <0.001
tet(B)b,d aac(6')-Iaad aph(6)-Idb,d
Source type OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Human REF <0.001 (global) REF 0.001
(global) REF <0.001 (global)
Livestock 4.23 (1.47-12.16) 0.007 0.29
(0.15-0.56) <0.001 1.62 (0.75-3.47) 0.218
Raccoon 0.12 (0.01-1.15) 0.067 0.23
(0.09-0.58) 0.002 0.05 (0.01-0.45) 0.007
Environment 0.20 (0.04-1.15) 0.072 0.49
(0.24-1.03) 0.059 0.26 (0.09-0.77) 0.015
224
aph(3'')-Ibc blaCMY-2c fosA7d
Source type OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Human REF <0.001 (global) REF <0.001
(global) REF 0.003 (global)
Livestock 2.91 (1.11-9.68) 0.030 0.51
(0.26-1.02) 0.045 1.26 (0.66-2.40) 0.488
Raccoon 0.09* (0.00-0.66) 0.008 0.02*
(0.00-0.12) <0.001 0.76 (0.35-1.69) 0.506
Environment 0.25 (0.04-1.36) 0.062 0.06
(0.01-0.22) <0.001 0.48 (0.22-1.05) 0.066
a A random intercept was used to account for clustering of samples obtained from the same raccoon or swine manure
pit. Variance components were as follows: IncX1-1 5.12 (95%CI: 0.94-27.80); IncI1-Igamma 3.28 (95%CI: 0.15-
69.86); IncFIIS 0.44 (95%CI: 0.00-224.40); ColRNAI 7.68 (95%CI: 1.91-30.80). b Adjusted for year of sampling. c Exact logistic regression was used. d Ordinary logistic regression model.
* Median unbiased estimates obtained with exact logistic regression.
225
Table 5.8: Contrasts from logistic regression modelsa,b,c,d assessing the association between source type and
the occurrence of select antimicrobial resistance genes and plasmid replicons in Salmonella enterica isolates
collected from raccoons, humans, livestock, and environmental sources in southern Ontario, 2011-2013
(n=608)
IncX1-1a IncFIISa ColRNAIa
Contrast OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Livestock vs. Raccoon
1.44 (0.55-3.77) 0.461 0.06
(0.02-0.22) <0.001 19.07 (2.76-131.58) 0.003
Environment vs. Raccoon
0.15 (0.03-0.71) 0.017 0.34
(0.15-0.76) 0.008 2.23 (0.36-13.87) 0.388
Livestock vs. Environment
9.67 (2.06-45.42) 0.004 0.18
(0.08-0.42) <0.001 8.53 (1.85-39.29) 0.006
ColpVCb IncFiip96ac tet(B)b
OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Livestock vs. Raccoon
23.72 (3.21-175.21) 0.002 0.01
(0.00-0.03) <0.001 34.03 (4.61-251.02) 0.001
Environment vs. Raccoon
9.82 (1.25-77.23) 0.030 0.28
(0.14-0.52) <0.001 1.64 (0.15-18.64) 0.687
Livestock vs. Environment
2.41 (1.26-4.63) 0.008 0.02
(0.00-0.10) <0.001 20.68 (4.97-85.98) <0.001
aac(6')-Iaad aph(6)-Idb aph(3'')-Ibc
OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value
Livestock vs. Raccoon
1.25 (0.56-2.80) 0.587 29.73
(4.04-218.97) 0.001 36.03* (6.49-∞) <0.001
Environment vs. Raccoon
2.16 (0.91-5.09) 0.079 4.76
(0.56-40.51) 0.153 2.78* (0.29-∞) 0.268
Livestock vs. Environment
0.58 (0.33-1.03) 0.064 6.24
(2.63-14.84) <0.001 11.52 (3.67-58.30) <0.001
226
a Multi-level model with random intercept to account for clustering of samples obtained from the same animal or
swine manure pit. b Adjusted for year of sampling. c Exact logistic regression was used. d Ordinary logistic regression model.
* Median unbiased estimates were obtained using exact logistic regression.
blaCMY-2c
OR (95%CI) p-value
Livestock vs. Raccoon
27.41* (4.91-∞) <0.001
Environment vs. Raccoon
2.77* (0.29-∞) 0.268
Livestock vs. Environment
8.77 (2.77-44.64) <0.001
227
5.9 Figures
Figure 5.1. Causal diagram for analytical modeling. Solid lines show directionality of proposed relationships
between dependent and independent variables. Dashed grey lines show potential confounding relationships. A
random intercept was included to account for multiple samples originating from the same animal or swine
manure pit.
228
229
230
Figure 5.2. Population structure of 608 Salmonella enterica isolates from raccoons, livestock, humans, and environmental sources in southern Ontario
based on 3002-loci cgMLST scheme from Enterobase. Minimum spanning tree created using k=30 clustering threshold in GrapeTree. (A) Distribution of
45 serovars determined using SISTR. (B) Distribution of source types. (C) Distribution of internationally recognized multi-locus sequence types. Seven
isolates (2 raccoon, 1 soil, 2 water, 1 swine, 1 chicken) could not be typed. Frequency counts are in square brackets.
231
Human serovars Proportion of isolates from other source types
Source Type
Isolates (n)
Serovars (n) S.
Ent
eriti
dis
S. H
eide
lber
g
S. K
entu
cky
S. T
yphi
mur
ium
S. In
fant
is
S. N
ewpo
rt
S. A
natu
m
S. L
icht
field
S. D
ublin
S. C
hest
er
Proportion of all
isolates (%)
Human 58 10 24 24 19 10 7 3 3 3 3 2 100
Wild
life
Raccoon 92 17 1 1 0 8 10 36 0 0 0 0 56
Liv
esto
ck
Cattle 60 17 5 13 7 7 2 0 0 0 0 0 34
Swine 55 13 0 0 0 9 11 0 0 0 0 0 20
Chicken 214 11 5 34 53 0 0 0 1 0 0 0 93
Env
iron
men
tal Soil 45 13 2 0 0 13 11 27 0 0 0 0 53
Water 76 30 1 5 4 11 7 9 1 0 0 0 38
Beach 8 6 0 0 0 25 25 13 0 0 0 0 63
232
Figure 5.3. Heatmap of all serovars identified in humans and their proportional occurrence in Salmonella enterica isolates from raccoons, livestock, and
environmental sources in southern Ontario, 2011-2013 (n=608). Counts of samples and serovars are indicated on the left. Heat map indicates the
proportion of samples from each source type in common with human serovars, with darker shading representing a higher proportion. Except for
human sources, row totals of proportions do not add up to 100% since only serovars identified in human samples are presented here.
233
Figure 5.4. Population structure of 127 Salmonella enterica isolates (serovars Newport, Typhimurium,
Infantis) from raccoons, livestock, humans, and environmental sources in southern Ontario based on 3002-
loci cgMLST scheme from Enterobase. Serovars depicted here were those found to be overlapping in human
and raccoon sources; serovars Heidelberg and Enteritidis were not included here since only one raccoon
isolate was identified for each, these were reported descriptively in text. Minimum spanning tree created
using k=10 clustering threshold in GrapeTree. (A) Population structure with serovars determined using
SISTR. (B) Distribution of internationally recognized multi-locus sequence types. Three isolates (2 raccoon, 1
soil) could not be typed. (C) Distribution of source types. Frequency counts are in square brackets.
234
CHAPTER 6
6 Summary discussion and conclusions
6.1 Summary discussion
Foodborne illness and antimicrobial resistant infections continue to be major global causes of
illness and mortality in humans, despite ongoing research and surveillance efforts (World Health
Organization, 2015). Recognition of ecosystem health/One Health approaches has prompted
researchers to assess the role of wildlife and environmental sources in the epidemiology of
antimicrobial resistance (AMR) and foodborne illness (Karesh et al., 2012). The encroachment
of human developments into wildlife habitat, the rise of intensive farming practices, and the
dissemination of pathogens and antimicrobial resistance determinants via waterways, are some of
the ways in which shared environments may contribute to the transmission of potentially
pathogenic organisms between humans, domestic animals, wildlife, and the environment (Abebe
et al., 2020). Until recently, however, the contribution of wildlife to this dissemination has
received little attention, with very limited work specifically focusing on raccoons (Bondo et al.,
2016a; Bondo et al., 2016b, Baldi et al., 2019). Ultimately, the aim of this thesis was to
contribute to a greater understanding of the epidemiology of Campylobacter, Salmonella, and E.
coli, and associated antimicrobial resistance at the wildlife-human-livestock-environmental
interface, utilizing a combination of whole-genome sequencing data and epidemiologically
informed statistical modeling. Data acquired from a previous three-year repeated cross-sectional
235
study of raccoons on swine farms and conservation areas in southern Ontario (Bondo et al.,
2016a; Bondo et al., 2016b) served as the basis for the data used in this current thesis.
Our study location in southern Ontario represents an ideal setting to investigate the
interface between human, wildlife, livestock, and environmental sources of enteric pathogens
and AMR, since this region contains intensive agriculture, interspersed with populated cities (~1
million people), along with an abundance of conservation areas (~11) contained within the Grand
River watershed, the largest watershed in southern Ontario. Using previously collected data,
multi-level multivariable logistic regression was used to assess the impact of seasonal, climatic,
location, annual, and raccoon demographic factors on the occurrence of C. jejuni in raccoons on
swine farms and conservation areas (Chapter 2). The remaining chapters incorporated a genomic
epidemiological approach using whole-genome sequencing data from Salmonella and E. coli
isolates previously collected through the prior cross-sectional wildlife study (Bondo, 2015), to
assess for potential transmission between human, wildlife, livestock, and environmental sources
at a local scale (Chapters 3 & 4), and at a broad regional scale (Chapters 4 & 5). Potential
transmission was assessed by examining overlap between core-genome multi-locus sequence
types from different sources, assessments of population structure, the use of statistical modeling,
as well as direct comparisons of extensively multi-drug resistant isolates to assess for overlap in
AMR genes, plasmid types, and strain types (based on core-genome multi-locus sequence types [
cgMLST]). Local scale transmission of Salmonella, E. coli, and associated AMR on swine farms
was assessed in Chapter 3 for the following sources: raccoons, swine manure pits, and soil
samples. Transmission of E. coli and associated AMR was assessed both at the local scale (for
wildlife and soil samples on swine farms and conservation areas), and at a broad regional scale,
through a comparison of previous wildlife study samples to water samples previously obtained
236
from the Grand River watershed (Chapter 4). Finally, Chapter 5 examined the potential for
transmission of Salmonella and associated AMR at a broad scale in southern Ontario,
incorporating isolates from the previous wildlife study, along with isolates from previously
sequenced antimicrobial resistant human clinical Salmonella cases, and isolates from livestock
samples obtained through FoodNet Canada, the latter two of which were collected in the same
region and time period as the previous wildlife study.
6.1.1 Major findings
The carriage of Campylobacter jejuni by raccoons in our study population was determined to be
associated with various climatic factors (e.g., season, temperature, rainfall), but the effects were
not consistent, and varied by location and year. Findings from spatial scan analyses also
demonstrated inconsistent effects of temporal and spatial factors on C. jejuni prevalence. The
prevalence of C. jejuni in raccoons did not vary between swine farms and conservation areas,
suggesting that local scale transmission is not frequently occurring between farms and raccoons.
However, a single C. coli isolate was identified in a raccoon trapped on a swine farm, which was
also the predominant species of Campylobacter identified in swine manure pit samples from
these farms (Mutschall et al., 2020). This previous work, comparing Campylobacter genotypes
(based on comparative genomic fingerprinting and 40-gene locus; CGF-40) from this raccoon
population against a larger Canadian database of isolates obtained from animals, livestock, and
environmental sources, revealed that the majority of isolates in raccoons were strongly
associated with livestock sources and human infections (Mutschall et al., 2020). Together, these
findings suggest that transmission of Campylobacter between raccoons and livestock is
occurring on a broad scale, but there is currently little evidence confirming that direct
237
transmission is occurring on swine farms within our study region. However, the identification
and distribution of different Campylobacter species, in particular, C. coli (commonly associated
with pigs), provides some evidence of limited transmission from agricultural sources to wildlife
on farms.
Evidence of local transmission of Salmonella, E. coli and associated AMR determinants
was identified among wildlife, swine manure pit and soil samples on swine farms. Our findings
also provided some evidence of potential on- and between-farm transmission of Salmonella
between these particular sources. In addition, “clusters” of Salmonella serovars were identified
among raccoons, manure pits, and soil samples on swine farms, but restricted to certain years and
farms. Findings consistent with clonal dissemination were also apparent in these populations of
Salmonella and E. coli obtained on swine farms. The lack of association for many AMR
determinants identified in E. coli and Salmonella isolates with either source type or farm location
suggests widespread sharing among these sources and locations, but those genes associated with
source were often significantly more likely to be identified in raccoons than in samples from
swine manure pits. Overall, these findings suggest a lack of transmission between raccoons and
livestock sources on swine farms (i.e., manure pits) for certain AMR determinants. Meanwhile,
other resistance genes and plasmid replicons were not associated with any particular farm
location or source type, suggesting widespread dissemination of those AMR determinants in this
region, regardless of source.
Examination of the resistant E. coli population in the previous wildlife study, along with
water samples obtained from the Grand River watershed, revealed that some resistance genes
were significantly associated with source type, whereas other genes and plasmid types were
238
associated with location type, with a significantly greater odds of identification on swine farms
compared to conservation areas. International sequence types (e.g., ST131), extended-spectrum
beta-lactamase and cephamycinase-producing E. coli isolated in less than 10% of samples were
nearly exclusively identified in water samples, and in raccoon and soil samples from
conservation areas, but not on swine farms. Our findings from this chapter suggest that raccoons
may be acquiring AMR determinants from water, upstream agricultural sources, and
anthropogenic sources in conservation areas (e.g., refuse, dumpsters).
Our broad scale comparison of Salmonella from raccoons on swine farms with public
health surveillance samples obtained from livestock, water, and human clinical cases in the same
region and time period revealed that transmission of Salmonella and associated plasmids and
resistance genes between humans, livestock and water sources may be occurring. In contrast,
none of the Salmonella isolates from raccoons clustered together with human isolates, with the
(near) exception of a single Salmonella Heidelberg ST15 isolate from a raccoon that differed
only at 4 loci; however, since this serovar was most common to chickens, this finding most likely
represented an infection/colonization in the raccoon that originated from a livestock source.
Many plasmid types were strongly associated with certain sources; some plasmid types were
only isolated from human and livestock sources, whereas others were only isolated from raccoon
and soil samples. Antimicrobial resistance genes were generally highest in human and livestock
sources compared with raccoon and environmental samples; genes of public health importance
(e.g., blaCMY-2) common to human, livestock and environmental water samples, were nearly
absent in the subset of samples from the wildlife study. Overall, our findings suggest that the
rural populations of raccoons on swine farms in the Grand River watershed are likely not major
contributors to resistant human Salmonella infections in this region, but examination of
239
susceptible Salmonella isolates from human cases is warranted in future work, particularly for
serovars such as Thompson and Newport which were commonly identified in raccoons.
6.1.2 Limitations and future work
With the exception of the Campylobacter chapter, our sample sizes did not permit multivariable
modeling, or the ability to account for potential confounding by serovar (due to the sheer number
of serovars identified with few observations, models could not converge). Although our logistic
regression analyses do not adequately capture the complex relationships between AMR
determinants and various risk factors, they represent an important first step in this process.
Shortcomings in the modeling of Campylobacter in this thesis are related to the potentially
inaccurate characterization of predictors (e.g., location type), potential missing predictors, and
our lack of accounting for strain-related differences between different Campylobacter jejuni
isolates. Future work examining Campylobacter and other foodborne pathogens would benefit
from incorporating additional sampling sources on swine farms, sampling of water sources in the
surrounding region, and sampling of additional livestock species on farms in the near vicinity.
A fundamentally important source of bias in this thesis is related to the selection of
animals and farms for sampling, as well as the opportunistic inclusion of sequenced isolates from
humans and livestock, based on available data. The process of trapping wildlife to obtain
samples may have resulted in a study population that differs from the general population of
wildlife in the region, since some animals may have been reticent to enter traps, and others which
were re-sampled several times contributed to our study to a greater degree (Vogt et al., 2020).
Although this limitation is unavoidable, practically speaking, the representativeness of this
raccoon population could be improved in future work with the inclusion of raccoon populations
240
in urban environments-animals which are more likely to interact frequently with humans and
domestic animals such as dogs and cats.
The selection of E. coli isolates for sequencing in this thesis on the basis of demonstrated
phenotypic resistance represents a major bias in the selection of these isolates. As a result, our
findings of distinct populations of plasmid types in Salmonella and E. coli on swine farms may
be an artifact of this selection process, and assessments of phenotypically susceptible E. coli
isolates in future work will help to validate these findings. In addition, the opportunistic
inclusion of human Salmonella cases allowed us a preliminary assessment of possible
transmission between humans and raccoons in this region, but recognition of the biased
population of human isolates included here is critical to an accurate interpretation of our
findings. Not only are these human isolates a subset of phenotypically resistant Salmonella cases,
but they also represent cases reported to public health, and are subject to reporting bias. Thus,
our lack of findings connecting human Salmonella cases with Salmonella isolates from raccoons
should be interpreted with caution, since future assessments including more representative and
comprehensive populations of isolates from human and raccoon sources may demonstrate
otherwise. Unfortunately, we also could not rule out travel-a major potential confounder-as a
source of Salmonella for human cases, as these data were not available to our research team.
The characterization of plasmids by incompatibility (Inc) types also represents a major
limitation of this thesis. The identification of Inc types does not confirm its presence in a
plasmid, rather than a chromosome. Additionally, certain Inc families (e.g., IncF) have no core
genes, thus detection of IncF plasmids in two different samples does not necessarily mean they
are similar or related, as assumed here. Furthermore, Inc types which have not previously been
241
identified would not have been identified here. Our inability to conclusively determine whether
resistance genes were located on a plasmid could be amended in future work by reconstructing
plasmids using tools such as MOB-suite (Robertson et al., 2020).
Finally, assessing for transmission of Salmonella, E. coli and AMR determinants based
on overlap in cgMLST types and a cross-sectional study design has inherent limitations.
Identification of identical cgMLST subtypes in two different sources may represent transmission,
or it could represent exposure to a common source. Although the use of cgMLST appears to have
provided adequate discriminatory power for our population of isolates, future work may benefit
from the assessment of single-nucleotide polymorphisms (SNPs) to confirm instances of possible
clonal dissemination suggested by our work. In the future, intervention-based, hypothesis-testing
studies that combine genomic data with epidemiological approaches are needed to confirm
observational data which purport livestock as a primary source of pathogens and AMR
determinants for humans, wildlife, and the environment (Wee et al., 2020, Ramey & Ahlstrom,
2020).
6.2 Conclusions
This thesis contributes new insights about risk factors for the carriage of Campylobacter by
raccoons, as well as the genomic epidemiology of Salmonella, E. coli and associated AMR in
wildlife in southern Ontario, on both a local and a broad regional scale. To date, there is limited
research examining subtyping data and gene-level antimicrobial resistance data in raccoons
(Bondo et al., 2016a; Bondo et al., 2016b; Very et al., 2016; Compton et al., 2008); this study
contributes new data to this literature regarding common Salmonella and E. coli serovars,
242
sequence types, microbial population structure, resistance genes and plasmid types carried by a
rural raccoon population. Comparison of whole-genome sequencing data from Salmonella
isolates from a previous wildlife study with isolates from human, livestock, and water samples
from the Grand River watershed revealed that Salmonella obtained on swine farms were more
similar, and, on average, demonstrated fewer allelic differences than isolates obtained from the
broader geographic region. Based on similarities in cgMLST subtypes, on- and between-farm
transmission of Salmonella between raccoons, soil, and swine manure pits may be occurring. Our
examination of resistant E. coli in the previous wildlife study, along with water samples from the
Grand River watershed revealed that international sequence types (e.g., ST131), extended
spectrum beta-lactamase and cephamycinase-producing E. coli were nearly exclusively identified
in water samples, and in raccoon and soil samples from conservation areas, but not swine farms.
Statistical analyses revealed associations between certain genes and plasmid types, with different
sources and location types; in many cases, livestock and water samples were significantly more
likely to harbour resistance genes and plasmid types compared to wildlife. In conclusion, it is
clear that apparently healthy wildlife may act as sentinels of environmental contamination with
AMR, and that they may differ in the carriage of certain strains, genes and plasmid types
according to their local environment. Our findings of a lower prevalence of AMR genes in
wildlife (both descriptively and statistically) suggest that anthropogenic sources such as
conservation areas, swine farms, and agricultural run-off in water may represent sources of AMR
contamination for wild animals, but intervention-based studies that combine genomic data with
epidemiological approaches are needed to confirm these observational data (Wee et al., 2020,
Ramey & Ahlstrom, 2020).
243
6.3 References
Abebe, E., Gugsa, G., & Ahmed, M. (2020). Review on major food-borne zoonotic bacterial pathogens. Journal of Tropical Medicine, 2020, 4674235.
Baldi, M., Barquero Calvo, E., Hutter, S. E., & Walzer, C. (2019). Salmonellosis detection and evidence of antibiotic resistance in an urban raccoon population in a highly populated area, Costa Rica. Zoonoses and Public Health, 66(7), 852–860.
Bondo, K. J. (2015). Epidemiology of Salmonella and antimicrobial resistant enteric bacteria in raccoons and the environment on swine farms and conservation areas in southern Ontario [Unpublished Doctoral thesis]. University of Guelph. Available from: http://hdl.handle.net/10214/9454
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016a). Epidemiology of antimicrobial resistance in Escherichia coli isolates from raccoons (Procyon lotor) and the environment on swine farms and conservation areas in Southern Ontario. PLoS ONE, 11(11).
Bondo, K. J., Pearl, D. L., Janecko, N., Boerlin, P., Reid-Smith, R. J., Parmley, J., & Jardine, C. M. (2016b). Impact of season, demographic and environmental factors on Salmonella occurrence in raccoons (Procyon lotor) from swine farms and conservation areas in southern Ontario. PLoS ONE.
Compton, J. A., Baney, J. A., Donaldson, S. C., Houser, B. A., San Julian, G. J., Yahner, R. H., Chmielecki, W., Reynolds, S., & Jayarao, B. M. (2008). Salmonella infections in the common raccoon (Procyon lotor) in western Pennsylvania. Journal of Clinical Microbiology, 46(9), 3084–3086.
Karesh WB, Loh E, Machalaba C. Food safety: A view from the wild side. In: Institute of Medicine (US). Improving Food Safety Through a One Health Approach: Workshop Summary. Washington (DC): National Academies Press (US); 2012. A7. Available from: https://www.ncbi.nlm.nih.gov/books/NBK114500/
Mutschall, S. K., Hetman, B. M., Bondo, K. J., Gannon, V., Jardine, C. M., & Taboada, E. N. (2020). Campylobacter jejuni strain dynamics in a raccoon (Procyon lotor) population in southern Ontario, Canada: High prevalence and rapid subtype turnover. Frontiers in Veterinary Science, 7, 27.
Ramey, A. M., & Ahlstrom, C. A. (2020). Antibiotic resistant bacteria in wildlife: Perspectives on trends, acquisition and dissemination, data gaps, and future directions. Journal of Wildlife Diseases, 56(1):1-15.
244
Robertson, J., Bessonov, K., Schonfeld, J., & Nash, J. (2020). Universal whole-sequence-based plasmid typing and its utility to prediction of host range and epidemiological surveillance. Microbial Genomics, 6(10), mgen000435.
Very, K. J., Kirchner, M. K., Shariat, N., Cottrell, W., Sandt, C. H., Dudley, E. G., Kariyawasam, S., & Jayarao, B. M. (2016). Prevalence and spatial distribution of Salmonella infections in the Pennsylvania raccoon (Procyon lotor). Zoonoses and Public Health, 63(3), 223–233.
Vogt, N. A., Stevens, C., Pearl, D. L., Taboada, E. N., & Jardine, C. M. (2020). Generalizability and comparability of prevalence estimates in the wild bird literature: Methodological and epidemiological considerations. Animal Health Research Reviews, 21(1), 89–95.
Wee, B. A., Muloi, D. M., & van Bunnik, B. (2020). Quantifying the transmission of antimicrobial resistance at the human and livestock interface with genomics. Clinical Microbiology and Infection, S1198-743X(20)30571-1. [Epub 22 Sept 2020].
World Health Organization. (2015). "WHO estimates of the global burden of foodborne diseases". Foodborne disease burden epidemiology reference group, 2007-2015. Available from: https://apps.who.int/iris/bitstream/handle/10665/199350/9789241565165_eng.pdf?sequence=1 (accessed on 10 August 2020).