Development of Intranasal Bacterial Therapeutics to Mitigate the Bovine Respiratory Pathogen Mannheimia Haemolytica

University of Calgary PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2019-11 Development of Intranasal Bacterial Therapeutics to Mitigate the Bovine Respiratory Pathogen Mannheimia haemolytica

Amat, Samat

Amat, S. (2019). Development of Intranasal Bacterial Therapeutics to Mitigate the Bovine Respiratory Pathogen Mannheimia haemolytica (Unpublished doctoral thesis). University of Calgary, Calgary, AB. http://hdl.handle.net/1880/111258 doctoral thesis

University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca UNIVERSITY OF CALGARY

Development of Intranasal Bacterial Therapeutics to Mitigate the Bovine Respiratory Pathogen

Mannheimia haemolytica

Samat Amat

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN VETERINARY MEDICAL SCIENCES

CALGARY, ALBERTA

NOVEMBER, 2019

The emergence of multidrug-resistant pathogens associated with bovine respiratory

disease (BRD) presents a significant challenge to the beef industry, as antibiotic administration is

commonly used to prevent and control BRD in commercial feedlot cattle in North America. Thus,

developing antibiotic alternatives such as bacterial therapeutics (BTs) to mitigate BRD is needed.

Recent studies suggest that the nasopharyngeal (NP) microbiota, particularly lactic acid-producing bacteria (LAB), are important to bovine respiratory health and may be a source of BTs for the

inhibition of BRD pathogens. The research presented in this thesis aimed to develop intranasal

BTs to mitigate the BRD pathogen Mannheimia haemolytica and promote NP microbiota stability

in feedlot cattle. Results from Study 1 showed that commercial probiotic bacteria were able to

inhibit M. haemolytica growth and its adherence to epithelial cells. Study 2 revealed that the NP

microbial community structure and relative abundance of LAB families underwent significant

changes when cattle transported from the farm to an auction market, then to feedlot. Many of the

LAB families were inversely correlated with the BRD-associated Pasteurellaceae family, and

isolates from Lactobacillaceae, Streptococcaceae and Enterococcaceae families inhibited growth

of M. haemolytica in vitro. This study provided evidence of potential antagonistic competition taking place between LAB and BRD-associated pathogens within the respiratory tract. Following these studies, using a targeted approach based on criteria evaluating M. haemolytica inhibition,

adherence to turbinate cells, and immunomodulation, 6 Lactobacillus strains from an initial group of 178 bacterial isolates originating from nasopharynx of cattle were identified as the best BT candidates (Study 3). Intranasal inoculation of these BTs reduced colonization by M. haemolytica

and induced modulation of respiratory microbiota in dairy calves experimentally challenged with

M. haemolytica (Study 4). Finally, the longitudinal effects of intranasally administered BTs on 2

the NP microbiota and the prevalence of BRD pathogens including Mannheimia were evaluated

in post-weaned beef calves (Study 5). A single dose of intranasal BTs induced longitudinal modulation of the NP microbiota while showing no adverse effects on animal health and growth performance. With further characterization of inoculant dose and time of inoculation, the BTs

may have potential for application as an antimicrobial alternative for mitigation of M. haemolytica

in beef cattle.

Acknowledgements

The combined efforts of many great people and organizations have made the successful completion of the research work presented in this thesis. First and foremost, I would like to gratefully acknowledge my supervisor Dr. Trevor Alexander for his guidance, support and encouragement throughout my PhD program. I have been lucky to have a supervisor who always trusted me and gave me the freedom to expand my research experience and collaboration network, as well as to pursue some of my own research ideas which were not within the scope of my thesis project. Dr. Alexander has made significant influence on my research career particularly on my academic writing and organizing research-related thoughts. I truly appreciate all the effort and long hours that Dr. Alexander has put into finalizing my thesis. I will always remember him by his following words and advice:

“Samat, I have faith in you, you can do it” “Sky is the limit” “Please see the big picture, but also pay attention to the details on how to get there” and “Never use the same word twice in a sentence”. Next,

I would like to thank my supervisor Dr. Edouard Timsit for his support, guidance, and constructive criticism during the course of my program. Dr. Timsit has helped me to foster my logical thinking skills. I am thankful for his help during my candidacy exam preparation, and for sending me for additional training. I gratefully acknowledge my supervisory committee including Drs. Tim

McAllister, Andre Buret and Joroen De Buck. I truly appreciate all the time and constructive feedback they provided for my project. I am also very thankful to my supervisory committee for giving me the second chance to redo my Candidacy exam. Although it was a painful experience in failing the first attempt, this failure was a wakeup call for me and led me to realize how much I don’t know about my

“craft”. During the next 6 months of Candidacy exam preparation, I have increased knowledge on my subject which was very useful for my thesis project. This candidacy exam experience has also taught me an important life lesson and made me to realize the meaning of Steve Job’s quote “You can't connect the dots looking forward; you can only connect them looking backward”. 4

I also extend a special thank you to Dr. Frank van deer Meer who agreed to join my supervisor committee as co-supervisor at the late stage of my program, and provided guidance in thesis writing.

I would also like to thank Dr. Danica Baines, who provided special assistance during my cell culture work. My sincere thanks also go to Dr. Devin Holman for his excellent contribution to the bioinformatics analysis, manuscript preparation, and helping me to learn bioinformatics. I also sincerely thank Dr. Timothy Schwinghamer for his special help in statistical analysis particularly with path modeling analysis.

I wish to thank Dr. Matthew Workentine for his assistance in some bioinformatics analysis. I am also grateful to Jay Yanke, Grant Duke, and Darell Vedres for their technical support in biochemical tests, microscopic imaging, and GC analysis, respectively. I express my appreciation to the following lab team members including Dr. Long Jin, Pamela Caffyn and Leandra Schneider for their technical support during my animal trials and laboratory analysis. I also thank the feedlot crew at the Lethbridge

Research and Development Centre, and animal care staff at the Veterinary Science Research Station, and Pathologists at Diagnostic Services Unit, College of Veterinary Medicine, University of Calgary for their technical support during the animal trials.

I also thank my friends Dennis and Marie-Jeanne Will, Marvin Genno, and Bev Lanz for the support they provided me and my family during my PhD program. My special thanks also goes to my

MSc. supervisor Dr. Steve Hendrick, who kindly provided me several hours of lectures on

Epidemiology and feedlot practice which helped me to get ready for my candidacy exam.

I am sincerely grateful to my wife Mikrigul for her tremendous support throughout my study.

Without her support, I would not have been able to concentrate on my research work as much as I did and spend as much time in the lab and office as I spent. She made more time for my work by taking care of our kids in most of the evenings and weekends. I also thank my two beautifful daughters,

Subina and Nargiza for bringing me more joy and strength, and for being so forgiving to their dad 5

when he missed many nights to put them to sleep and missed attending some of their weekend activities.

I am extremely indebted to my parents, brothers, and sisters Wurgul and Nurgul Amat for their endless support during the pursuit of my education.

I also acknowledge the financial support provided by the Alberta and Agiculture Forestry, Beef

Cattle Research Council, and Agriculture and Agri-Food Canada (AAFC). I am grateful for the

Canadian Natural Science and Engineering Research Council (NSERC) for providing me with the

NSERC doctoral scholarship.

Last, but not the least, I would like to thank the support I received from the friendly and supportive research community at the AAFC Lethbridge Research and Development Centre. My thanks also goes to Department of Production Animal Health, and the Office of Graduate Study at the

Faculty of Veterinary Science, University of Calgary.

Dedication

I am dedicating this thesis to the few special people in my life. First, to my dad Amat

Niyaz and my mom Hansahan Yasin who have taught me the value of working hard, and to always

follow my dreams. The advice my dad gave me at the early stage of my life: “Hustle silently and your work will speak for itself” and “You are the only one who can create your own future” always regulate my life. Next, to my brothers Ahat and Mamat Amat who have genuinely given up their own higher education opportunities to support their brother and sisters’ education.

Table of Contents

Abstract ...... 2 Acknowledgements ...... 4 Table of Contents ...... 8 List of Tables ...... 13 List of Figures ...... 14 List of Abbreviations ...... 17

CHAPTER ONE: LITERATURE REVIEW ...... 20 1.1 Background ...... 20 1.2 Bovine respiratory disease ...... 21 1.2.1 Significance of the disease ...... 21 1.2.2 Predisposing factors ...... 22 1.2.3 Viral agents ...... 23 1.2.4 Bacterial agents ...... 23 1.3 BRD management strategies ...... 26 1.3.1 Preconditioning ...... 26 1.3.2 Vaccination ...... 27 1.3.3 Antimicrobial use ...... 28 1.4 Bovine respiratory defense mechanisms against bacterial pathogens ...... 30 1.4.1 Physical barriers: ...... 31 1.4.2 Biochemical barriers ...... 32 1.4.3 Cellular barriers ...... 33 1.4.4 The respiratory commensal bacteria ...... 33 1.5 Bovine respiratory microbiota ...... 35 1.5.1 Structure and composition of respiratory microbiota in beef cattle described using 16S rRNA high-throughput sequencing ...... 35 1.5.2 Management factors that influence the respiratory microbiota of beef cattle .37 1.5.3 Potential association between the respiratory microbiota and development of BRD ...... 39 1.6 Probiotics ...... 41 1.6.1 Mechanisms of probiotic action ...... 41 1.6.2 Selection criteria and requirements for probiotic strains ...... 42 1.6.3 Probiotic use in cattle ...... 44 1.6.4 Targeted development of bacterial therapeutics ...... 45 1.7 Conclusion ...... 46 1.8 Hypothesis ...... 47 1.9 Objectives ...... 47 1.10 Tables and Figures ...... 49

CHAPTER TWO: POBIOTIC BACTERIA INHIBIT THE BOVINE RESPIRATORY PATHOGEN MANNHEIMIA HAEMOLYTICA SEROTYPE 1 IN VITRO ...... 52 2.1 Introduction ...... 53 2.2 Results and discussion ...... 54

2.2.1 Inhibition of M. haemolytica serotype 1 (S1) by probiotic bacteria ...... 54 2.2.2 Bacterial adhesion to BBE cell monolayers ...... 55 2.2.3 Antagonistic activity of probiotic bacteria against M. haemolytica S1 ...... 56 2.3 Conclusion ...... 57 2.4 Materials and methods ...... 58 2.4.1 Bacterial strains and culture conditions ...... 58 2.4.2 Growth inhibitory effects of probiotic bacteria against M. haemolyica S1 .....58 2.4.3 Collection of BBE cells ...... 59 2.4.4 Assessment of bacterial adhesion to BBE cells ...... 60 2.4.5 Determination of antagonistic activity of probiotic strains against M. haemolytica S1 ...... 60 2.4.6 Statistical analysis ...... 61 2.5 Tables and Figures ...... 62

CHAPTER THREE: EVALUATION OF THE NASOPHARYNGEAL MICROBIOTA IN BEEF CATTLE TRANSPORTED TO A FEEDLOT, WITH A FOCUS ON LACTIC ACID-PRODUCING BACTERIA ...... 66 3.1 Introduction ...... 67 3.2 Materials and Methods ...... 69 3.2.1 Ethics statement ...... 69 3.2.2 Animal husbandry and experimental design ...... 69 3.2.3 Nasopharyngeal swab sampling and isolation of BRD-associated pathogens 70 3.2.4 16S rRNA gene sequencing and analysis ...... 70 3.2.5 In vitro growth inhibition of M. haemolytica by LAB isolates ...... 71 3.2.6 Statistical analysis ...... 72 3.3 Results ...... 73 3.3.1 Isolation and detection of BRD-associated pathogens ...... 73 3.3.2 Effect of transport to the auction market and feedlot on the structure of the nasopharyngeal microbiota ...... 74 3.3.3 Longitudinal changes among the Lactobacillales families ...... 74 3.3.4 Longitudinal changes among the archaeal and bacterial phyla and genera in the nasopharyngeal microbiota ...... 75 3.3.5 Associations between the 15 most relatively abundant genera ...... 76 3.3.6 Differentially abundant taxa in the nasopharyngeal microbiota following feedlot placement ...... 77 3.3.7 The relationship between LAB and BRD-associated Pasteurellaceae family members ...... 78 3.4 Discussion ...... 79 3.4.1 The prevalence of cultured BRD-associated pathogens ...... 80 3.4.2 Effect of transport and auction market commingling on the nasopharyngeal microbiota ...... 82 3.4.3 Relative abundance and antimicrobial activity of lactic acid-producing bacteria, and correlation with the BRD-associated Pasteurellaceae family ...... 85 3.5 Tables and Figures ...... 89

CHAPTER FOUR: DEVELOPMENT OF BACTERIAL THERAPEUTICS AGAINST THE BOVINE RESPIRATORY PATHOGEN MANNHEIMIA HAEMOLYTICA...... 101 4.1 Introduction ...... 102 4.2 Materials and Methods ...... 103 4.2.1 Isolation of commensal bacteria from the nasopharynx of feedlot cattle ...... 103 4.2.2 Identification of nasopharyngeal commensal bacterial isolates ...... 104 4.2.3 Growth inhibitory effects of nasopharyngeal commensal bacteria against M. haemolytica ...... 105 4.2.4 Adherence of commensal bacteria to bovine turbinate cell monolayers ...... 106 4.2.5 Antagonistic competition activity of commensals against M. haemolytica on bovine turbinate cells ...... 107 4.2.6 Fluorescent microscopy of bacteria adhering to bovine turbinate cells ...... 107 4.2.7 Evaluation of antibiotic susceptibility of selected isolates ...... 108 4.2.8 Effects of Lactobacillus spp. isolates on the expression of genes associated with adaptive and innate immune response in BT cells ...... 110 4.2.9 M. haemolytica inhibitory mechanisms of 6 Lactobacillus isolates, as candidate bacterial therapeutics ...... 111 4.2.10 Statistical analysis ...... 114 4.3 Results ...... 114 4.3.1 Isolation and identification of nasopharyngeal commensal bacterial isolates114 4.3.2 Growth inhibitory effects against M. haemolytica ...... 114 4.3.3 Adherence of selected isolates to bovine turbinate cells ...... 116 4.3.4 Antagonistic competition activity of selected isolates against M. haemolytica116 4.3.5 Antimicrobial susceptibility of selected isolates ...... 117 4.3.6 Stimulation of innate and adaptive immune responses in bovine turbinate cell monolayers ...... 118 4.3.7 M. haemolytica inhibitory mechanisms of 6 Lactobacillus isolates, as candidate bacterial therapeutics ...... 119 4.4 Discussion ...... 121 4.5 Tables and Figures ...... 130

CHAPTER FIVE: INTRANASAL BACTERIAL THERAPEUTICS REDUCE COLONIZATION BY THE RESPIRATORY PATHOGEN MANNHEIMIA HAEMOLYTICA IN DAIRY CALVES ...... 148 5.1 Introduction ...... 149 5.2 Materials and Methods ...... 151 5.2.1 Ethics statement ...... 151 5.2.2 Animals and husbandry ...... 151 5.2.3 Study design ...... 151 5.2.4 Preparation of the nasal inoculum ...... 153 5.2.5 Sampling and processing of NS ...... 154 5.2.6 Isolation and enumeration of M. haemolytica from NS ...... 154 5.2.7 PCR identification and pulsed field gel electrophoresis typing of M. haemolytica ...... 155 5.2.8 Isolation of LAB from NS ...... 155 10

5.2.9 PCR identification and rep-PCR typing of Lactobacillus isolates ...... 155 5.2.10 Blood sample collection and processing ...... 156 5.2.11 Quantification of cytokines in serum using enzyme-linked immunosorbent assay (ELISA)...... 157 5.2.12 Trans-tracheal aspiration sampling and processing ...... 157 5.2.13 Post-mortem examination ...... 157 5.2.14 DNA extraction from NS and TTA samples and 16S rRNA sequencing ...158 5.2.15 Estimation of Lactobacillus spp. abundance from NS using qPCR ...... 159 5.2.16 Statistical analysis ...... 159 5.3 Results ...... 162 5.3.1 Animal health ...... 162 5.3.2 Isolation and enumeration of M. haemolytica from NS ...... 162 5.3.3 PFGE typing of M. haemolytica isolates ...... 163 5.3.4 Lactobacillus abundance in NS determined by qPCR ...... 163 5.3.5 Rep-PCR typing of Lactobacillus isolates ...... 163 5.3.6 Effects of BT on the composition and diversity of the nasal microbiota ...... 164 5.3.7 Effects of BT on the recursive structure of causal relationships among the 10 most relatively abundant genera in the nasal microbiota ...... 165 5.3.8 Effects of BT on the composition and diversity of the tracheal microbiota ..167 5.3.9 Effects of bacterial therapeutics on serum cytokine concentrations ...... 168 5.4 Discussion ...... 168 5.5 Tables and Figures ...... 176

CHAPTER SIX: INTRANASAL ADMINISTRATION OF BACTERIAL THERAPEUTICS INDUCES LONGITUDINAL MODULATION OF THE NASOPHARYNGEAL MICROBIOTA IN POST-WEANED BEEF CALVES ...... 189 6.1 Introduction ...... 190 6.2 Materials and Methods ...... 192 6.2.1 Animals and experimental design ...... 192 6.2.2 Preparation of BT inoculum ...... 193 6.2.3 Administration of BT cocktail and tulathromycin ...... 193 6.2.4 Nasopharyngeal swab sampling and processing ...... 194 6.2.5 Genomic DNA extraction, 16S rRNA gene sequencing and analysis ...... 195 6.2.6 Statistical analysis ...... 197 6.3 Results ...... 200 6.3.1 Calf health and weight gain ...... 200 6.3.2 Prevalence of BRD-associated pathogens determined by NP swab culturing200 6.3.3 Total bacteria and Lactobacillus in NP swabs determined by qPCR ...... 201 6.3.4 Structure and composition of the NP microbiota ...... 201 6.3.5 Changes in microbial composition following BT and tulathromycin treatment203 6.3.6 Microbial interactions and dynamics of the NP microbiota ...... 205 6.3.7 Antimicrobial resistance determinants in the NP microbiota ...... 210 6.4 Discussion ...... 210 6.4.1 Colonization by BTs ...... 211 6.4.2 Longitudinal effects of BTs and tulathromycin on the respiratory microbiota211 11

6.4.3 Longitudinal effects of BTs and tulathromycin on BRD-associated pathogens216 6.4.4 Longitudinal effects of BTs and tulathromycin on antimicrobial resistant determinants ...... 218 6.5 Tables and Figures ...... 220

CHAPTER SEVEN: GENERAL DISCUSSION, CONCLUSIONS AND FUTURE DIRECTIONS ...... 233 7.1 General discussion ...... 233 7.2 Future work ...... 241 7.3 Concluding remarks ...... 242

REFERENCES ...... 244

APPENDIX A: SUPPLEMENTARY TABLES ...... 283

APPENDIX B: SUPPLEMENTARY FIGURES ...... 299

APPENDIX C: COPYRIGHT PERMISSIONS ...... 310

List of Tables

Table 2.1 Bacterial strains used in the present study...... 62

Table 2.2 Growth inhibitory effects of probiotic bacteria against Mannheimia haemolytica S1...... 63

Table 2.3 Antagonistic effects of probiotic bacteria against Mannheimia haemolytica S1 adherence to bovine bronchial epithelial cells*...... 64

Table 3.1 Calves (n = 13) positive for BRD-associate bacterial pathogens by culturing of nasopharyngeal swabs...... 89

Table 3.2 Associations among the 15 most relatively abundant genera in the nasopharyngeal microbiota of cattle (n =13) across time...... 90

Table 3.3 Differentially abundant OTUs in the nasopharnyngeal microbiota of feedlot cattle between the d 0 and d 14 (n = 13)...... 91

Table 3.4 Differentially abundant OTUs in the nasopharyngeal microbiota of feedlot cattl between the d 0 and d 2 (n = 13)...... 93

Table 3.5 Correlations between families within the order Lactobacillales, and the BRD- associated Pasteurellaceae family in the nasopharyngeal microbiota of cattle (n =13) across sampling times...... 94

Table 4.1 List of bacteria identified from the nasopharynx of healthy feedlot cattle and those selected for initial inhibition of M. haemolytica using agar slabsa...... 130

Table 4.2 Minimum inhibitory concentrations (µg/ml) of antibiotics against 15 bacterial strains isolated from the nasopharynx of feedlot cattlea...... 131

Table 4.3 Selected genes that had expression altered in bovine turbinate cells after incubation with bacteria isolated from the nasopharynx of cattle...... 133

Table 4.4 Antimicrobial properties of selected bacterial therapeutic strains (n = 6) evaluated by the measurement of their lactic acid and H2O2 production and bacteriocin-encoding genes...... 137

Table 5.1 Comparison of M. haemolytica counts determined by nasal swab culturing between dairy calves that received intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT + Mh) or only M. haemolytica (Mh)...... 176

Table 5.2 Path models fitted to data corresponding to statistics that are Schwarz’s (Bayesian) information criteria of three path models fitting subsets of experimental data...... 177

Table 6.1 Manifest variables in the modified path models...... 220 13

List of Figures

Figure 1.1 Defense mechanisms of bovine respiratory tract against pathogens...... 49

Figure 1.2 Mucosal clearance mechanism of respiratory tract...... 50

Figure 1.3 Probiotic mechanisms of action...... 51

Figure 2.1 Adhesion of probiotic bacteria and the bovine respiratory pathogen M. haemolytica S1 to bovine bronchial epithelial cells...... 65

Figure 3.1 Beta-diversity of nasopharyngeal microbiota...... 95

Figure 3.2 Relative abundance of lactic acid-producing bacteria (LAB)...... 96

Figure 3.3 Box and whisker plots of the 15 most abundant genera in the nasopharyngeal microbiota of cattle (n = 13) by sampling time...... 98

Figure 3.4 Box and whisker plots of the A) number of OTUs and B) Shannon diversity index of the nasopharyngeal microbiota of cattle (n = 13) by sampling time...... 99

Figure 3.5 Growth inhibitory effects of lactic acid-producing bacteria (LAB) isolated from the nasopharynx of healthy feedlot cattle against a bovine respiratory pathogen Mannheimia haemolytica serotype 1 strain, as determined by the agar slab method. The results are presented as mean zones of inhibition (plus standard deviations [SD]) from three replicates...... 100

Figure 4.1 The schematic workflow chart...... 138

Figure 4.2 Growth-inhibitory effects of bovine respiratory bacteria against M. haemolytica. ... 140

Figure 4.3 Adherence of bovine respiratory bacterial isolates to bovine turbinate cell monolayers...... 142

Figure 4.4 Antagonistic competition of bovine respiratory bacteria (n = 15) against M. haemolytica...... 144

Figure 4.5 Growth inhibition effects of lactic acid on the M. haemolytica...... 145

Figure 4.6 Scanning electron microscopy images of Mannheimia haemolytica after incubated with cell-free culture supernatants of selected bacterial therapeutic strains...... 146

Figure 5.1 Study design and sampling regimen...... 178

Figure 5.2 Bacterial counts of M. haemolytica and abundance of Lactobacillus spp...... 179

Figure 5.3 Beta diversiry of nasal microbiota...... 180

Figure 5.4 Relative abundance of the 5 most relatively abundant phyla and 10 most relatively abundant genera in the nasal microbiota of calves that received an intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT + Mh; n = 12) or only M. haemolytica (Mh; n = 12)...... 182

Figure 5.5 The A) number of OTUs and B) Shannon diversity index values) of the nasal microbiota of calves that received an intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT + Mh; n = 12) or only M. haemolytica (Mh; n = 12)...... 183

Figure 5.6 Path diagram of models showing the recursive structure of causal relationships among the 10 most abundant genera in the nasal microbiota of calves that received an intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT + Mh; n = 12) or only M. haemolytica (Mh; n = 12)...... 185

Figure 5.7 Description of the tracheal microbiota of calves that received an intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT + Mh) or only M. haemolytica (Mh)...... 186

Figure 5.8 Serum cytokine concentrations (IL-6, IL-8 and IL-10) by sampling day in dairy calves that received an intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT + Mh; n = 12), or only M. haemolytica (Mh; n = 12)...... 188

Figure 6.1 Prevalence of the BRD-associated pathogens in nasopharynx of cattle received either intranasal bacterial therapeutics (BT), PBS (CTRL) or subcutaneous tulathromycin (MP) (n = 20 per group) over the course of 28 days, determined by culturing nasopharyngeal swabs. *Significant difference between treatments (P < 0.05). 221

Figure 6.2 Abundance of total bacteria (A), and Lactobacillus (B) estimated by qPCR in nasopharyngeal swab samples obtained from calves received either intranasal bacterial therapeutics (BT), PBS (CTRL) or subcutaneous tulathromycin (MP) (n = 20 per group) over the course of 42 days...... 222

Figure 6.3 Beta diversity of nasopharyngeal microbiota...... 223

Figure 6.4 Alpha diversity of nasopharyngeal microbiota...... 224

Figure 6.5 Relative abundance of the 5 most abundant phyla in the nasopharyngeal microbiota of calves that received an intranasal inoculation of either PBS (CTRL) or bacterial therapeutics (BT), or subcutaneous metaphylaxis (MP) (n = 20 per group)...... 225

Figure 6.6 Taxa (n = 28) that showed a significant change from baseline (day-1) in BT and MP groups above and beyond any changes in the control (CTRL) group over the course of 42 days...... 226

Figure 6.7 The ecological network of the all observed genera in nasopharyngeal microbiota of calves received either intranasal bacterial therapeutics (BT), PBS (CTRL) or subcutaneous tulathromycin (MP) (n = 20 per group) over the course of 42 days...... 227

Figure 6.8 Path diagram of models showing the recursive structure of causal relationships among the 16 selected genera in the nasopharyngeal microbiota of calves that received an intranasal inoculation of either bacterial therapeutics (BT) (A) or PBS (CTRL) (B), or subcutaneous metaphylaxis (MP) (C) (n = 20 per group)...... 231

Figure 6.9 The proportion (%) of the resistance determinants msr(E) and tet(H) to 16S rRNA gene copies in nasopharyngeal samples obtained from cattle received either intranasal bacterial therapeutics (BT), PBS (CTRL) or subcutaneous metaphylaxis (MP) (n = 20 per group) over the course of 42 days...... 232

List of Abbreviations

Symbols Definition

AMPs Antimicrobial peptides

ANOVA Analysis of variance

ASL Air surface liquid

BAL Bronchoalveolar lavage

BBE cells Bovine bronchial epithelial

BHI Brain–heart infusion

BHV-1 Bovine herpesvirus type 1

BIC Bayesian information criterion

BRD Bovine respiratory disease

BRSV Bovine respiratory syncytial virus

BTs Bacterial therapeutics

BVDV Bovine viral diarrhea virus

DCA Detrended correspondence analysis

DMEM Dulbecco's Modified Eagle Medium

DNA Deoxyribonucleic acid

DNS Deep nasopharyngeal swab

DPBS Dulbecco’s phosphate-buffered saline

FDA Food and Drug Administration

GLIMMIX Generalized liner mixed model

IgA Immunoglobulin A

IgG Immunoglobulin G

IQR Interquartile range

LAB Lactic acid-producing bacteria

LPS Lipopolysaccharide

MIC Minimum inhibitory concentrations

MP Metaphylaxis

MRS De Man Rogosa and Sharpe agar

NGP Next generation probiotics

NK cells Natural killer cells

NP Nasopharyngeal

PAMPs Pathogen associated molecular patterns

PBS Phosphate buffered saline

PCA Principal coordinates analysis

PCR Polymerase chain reaction

PERMANOVA Permutational multivariate analysis of variance

PI3 Parainfluenza virus type 3

PRRs Pathogen recognition receptors

RMSEA Root mean square error of approximation rRNA Ribosomal ribonucleic acid

SBC Schwarz’s Bayesian criteria

SCFA Short chain fatty acid

TTA Transtracheal aspiration

URT Upper respiratory tract

VRE Vancomycin-resistant Enterococcus

QPS Qualified presumption of safety

Chapter One: Literature review1

1.1 Background

Multiple factors have been associated with the development of bovine respiratory disease

(BRD) but bacterial species, including Mannheimia haemolytica, Histophilus somni, Mycoplasma bovis, and Pasteurella multocida are frequently implicated (Confer, 2009). The upper respiratory tract is a reservoir of these opportunistic pathogens, which can proliferate and infect the lungs when immunity in cattle is compromised due to stress or primary viral infections (Hodgson et al.,

2005). High risk cattle populations entering feedlots are most susceptible to BRD, and as a result, are often administered metaphylactic antibiotics to prevent infections after feedlot placement.

However, there are public and scientific concerns regarding antimicrobial use in livestock production (Cameron and McAllister, 2016), and recent reports indicate high levels of resistance in important BRD pathogens from feedlot cattle, potentially limiting the effectiveness of the antimicrobials available for treatment (Portis et al., 2012; Anholt et al., 2017; Timsit et al., 2017).

Thus novel methods to reduce antimicrobial use and mitigate BRD-related pathogenic bacteria are

needed.

While pathogenic bacteria that can be cultured in the laboratory have been the main focus

of research on the bovine respiratory tract, advances and affordability of next generation

sequencing have led to an increased number of studies on the respiratory microbiome. Resulting

from these studies is an improved understanding of the importance of the mammalian microbiome

in relation to host health, and it is clear that the resident microbiota of the respiratory tract have a

1 A version of this literature review is being submitted to the Journal of Applied Microbiology as an invited review. 20

critical role in preventing the colonization of pathogens (Bogaert et al., 2004; Cho and Blaser,

2012). Establishment and stability of the respiratory microbiota is important for homeostasis while

disruption can promote pathogenesis (Man et al., 2017). For example, reductions in bacterial

community density and diversity following antibiotic treatment in humans have been associated

with an increased risk of overgrowth of bacterial pathogens and development of respiratory diseases (Pettigrew et al., 2012). There is evidence showing that the cattle respiratory microbiota are also susceptible to disturbances. In beef cattle, transportation to a feedlot (Holman et al., 2017), diet composition (Hall et al., 2017), and antimicrobial administration (Holman et al., 2018;

Holman et al., 2019) have previously been shown to affect the nasopharyngeal microbiota, highlighting that respiratory bacteria of cattle are perturbed by industry management practices. It has also been shown that bacterial diversity and richness were reduced in the nasopharynx of cattle that developed BRD early in the feeding period compared with cattle that remained healthy

(Holman et al., 2015a) and certain bacterial communities have been associated with bovine respiratory health (Timsit et al., 2018). These data suggest that microbiota-based interventions

(e.g. bacterial therapeutics) may provide new opportunities for managing BRD.

1.2 Bovine respiratory disease

1.2.1 Significance of the disease

Bovine respiratory disease still remains the most common endemic disease in North

American commercial feedlots, despite the advances in vaccination and increased use of metaphylactic antimicrobials upon feedlot arrival (Brooks et al., 2011; Cernicchiaro et al., 2013).

Approximately 21% of cattle are estimated to be affected by BRD during feedlot placement

(NAHMS, 2011). The disease accounts for 70-80% of total morbidity, and 40-50% of total

mortality in United States feedlots (Hilton, 2014). It has been estimated that the United States beef industry loses more than 4 billion annually due to the BRD-induced mortality, reduced feed efficiency and performance in BRD-affected cattle, as well as BRD-associated treatment costs

(Cernicchiaro et al., 2013). Given the similar nature of the beef industry between Canada and the

United States, economic losses are therefore expected to affect the Canadian beef cattle industry on a proportional basis.

1.2.2 Predisposing factors

Cattle are mostly affected by BRD within the first 45 days after feedlot placement

(Loneragan et al., 2005; Buhman et al., 2000). Due to the segmented nature of beef cattle production in North America, calves arriving at the feedlot are often exposed to various stresses associated with weaning, maternal-separation, transportation, comingling at the auction market and handling during transition from the farm to the feedlot. These stresses and exposure to bacterial and viral pathogens due to the mixing, often result in compromised host defense against respiratory infection. This ultimately predisposes the animals to primary viral infection and subsequent bacterial infection that are characteristic of BRD (Taylor et al., 2010).

Multi-etiological factors are involved in the development of BRD and these factors include host, environment, management, and viral and bacterial infectious agents (Mosier, 2014). Host factors such as age, body weight, immune status and genetics can influence BRD susceptibility in cattle (Taylor et al., 2010). Younger and lighter-weight calves entering the feedlot are more prone to develop BRD during the feeding phase compared to older and heavier calves (Townsend et al.,

1989; Sanderson et al., 2008). Shipping is one of the leading environmental risk factors for BRD because almost all cattle placed in the feedlot are transported from elsewhere. Studies have shown that cattle transported from longer distance become less resilient to respiratory infections at feedlot 22

placement (Baruch et al., 2019; Taylor et al., 2010). The majority of feedlot cattle are sourced from

auction markets, and comminged with other calves at the auction market. Auction market-derived

calves are at greater risk for BRD compared to the ones purchased directly from the farm (Step et

al., 2008) due to sale barn calves having greater exposure to pathogens and stress as a result of

mixing with calves from different sources (Step et al., 2008).

1.2.3 Viral agents

The main viral agents associated with BRD are bovine herpesvirus type 1 (BHV-1),

parainfluenza-3 virus (PI3), bovine viral diarrhea virus (BVDV), and bovine respiratory syncytial

virus (BRSV) (Grissett et al., 2015). These viral pathogens often colonize the upper respiratory

tract of cattle and can induce respiratory infections in cattle on their own. In post weaned beef

calves, however, they are involved in predisposing the respiratory tract to bacterial infections

(Gershwin et al., 2015; Grissett et al., 2015). Bacterial pathogen colonization of the upper

respiratory tract is enhanced by the viral agents that induce upregulation of mucosal binding sites

for the bacterial agents. In addition, viral agents promote the translocation of bacterial pathogens

from the upper respiratory tract to the lung by interfering with host respiratory defence

mechanisms including mucosal and immune cell clearance (Caswell, 2014; Murray et al., 2016).

The collaboration between viral and bacterial agents in causing BRD have been experimentally

demonstrated. Calves challenged with BHV-1, followed by M. haemolytica harbored greater abundance and more persistant colonization in their nasopharynx with M. haemolytica compared to those that were solely infected with M. haemolytica (Jericho et al., 1986; Word et al., 2019).

1.2.4 Bacterial agents

The predominant BRD bacterial agents are Mannheimia haemolytica, Pasteurella multocida, Histophilus somni and Mycoplasma bovis (Griffin et al., 2010) and are commonly

present as part of the normal nasopharyngeal microbiota of cattle. However, they become opportunistic when host immunity is compromised due to stress and/or viral infections, as well as when local microbial homestasis is disrupted.

Of the BRD bacterial pathogens, M. haemolytica is the principal agent recovered from

BRD cases. This pathogen induces acute lung infections characteristic to fibrinous pneumonia, and has a considerable economic impact on the North American feedlot industry (Rice et al., 2007).

M. haemolytica taxonomically belongs to genus Mannheimia, family Pasteurellaceae, and phylum

Proteobacteria. Mannheimia spp. are gram-negative, non-motile, facultative anaerobic coccobacilli or rods, encompassing five species: M. haemolytica, M. glucosida, M. ruminalis, M. varigena, and M. granulomatis (Angen et al., 1999). In cattle, M. haemolytica is the most abundant

Mannheimia species present in the respiratory tract and a recent metagenomic analysis of bronchoalveolar lavage samples from BRD mortalities did not detect an abundance of sequences from non-M. haemolytica species (Klima et al., 2019; < 0.05%). M. haemolytica has more than a dozen serotypes, with serotype 1, 2 and 6 being most frequently isolated from feedlot cattle. The pathogenicity of M. haemolytica depends on its capsular serotypes and the host species. Serotypes

1 and 6 are pathogenic to the cattle, while serotype 2 is less pathogenic to the cattle but more pathogenic in sheep (Rice et al., 2007; Klima et al., 2014a and 2016). M. haemolytica serotype 1 bacterium is more frequently isolated from the respiratory tract of cattle with BRD than serotype

6. Klima et al (2014a) observed that the majority of M. haemolytica isolates from the nasopharynx of feedlot cattle with BRD were primarily serotype 1 (70.7%), with serotype 6 accounting for

19.5% of total M. haemolytica isolates. Although M. haemolytica serotype 2 is less pathogenic to cattle, this serotype is frequently cultured from the nasopharynx of healthy feedlot cattle, with prevalence reaching up to 75.5% (Klima et al. 2014a). 24

In healthy cattle, M. haemolytica in the upper respiratory tract are part of the normal flora, and remain mostly confined to this site until respiratory homeostasis becomes disturbed due to stress and viral infection (Singh et al., 2011). Following stress and viral infection, M. haemolytica

can proliferate in the upper respiratory tract, which subsequently leads to inhalation and

colonization of M. haemolytica in the lung. Virulence factors including adhesion, capsular

polysaccharide, and fimbriae enable M. haemolytica to attach and colonize the respiratory tract. In

addition, M. haemolytica uses sialoglycoprotease, neuraminidase, transferrin-binding proteins,

leukotoxins, lipopolysaccharide (LPS) and lipoproteins to survive in the lung and avoid host

mucosal and immune defenses (Confer, 2009). The M. haemolytica-derived leukotoxins and host

defense interactions in the lung results in necrosis of apoptotic cells. As a result of necrosis,

antimicrobial agents and reactive oxygen species and lysozyme enzymes are released from the

necrotic cells to the surrounding tissue and ultimately cause lung tissue damage characteristic to

pneumonia (Rice 2007; Singh et al., 2011). M. haemolytica is the most commonly used BRD

pathogen to experimentally induce bronchopneumonia in cattle (Hanzlicek et al., 2007; Theurer et

al., 2013).

Like M. haemolytica, P. multocida and H. somni are also opportunistic bacterial pathogens

involved in bronchopneumonia in cattle with clinical signs similar to M. haemolytica-associated

pneumonia (Taylor et al., 2010). The isolation rate of P. multocida and H. somni from clinically

healthy feedlot cattle ranges from 15% up to 60%, but these pathogens are also isolated with

greater frequency from the lower respiratory tract of BRD-affected feedlot cattle compared to

healthy pen mates (Timsit et al., 2017). The pathogenesis of P. multocida in the development of pneumonia is accomplished by its virulence factors of LPS, cytotoxic, and iron acquisition proteins

(Griffin et al., 2010). H. somni can induce expression of immunoglobulin binding proteins and 25

endothelial cell apoptosis, and has the ability to escape killing by host immune cells using its phase

and antigen variation, and endotoxin activity of LPS (Griffin et al., 2010). These virulence factors

enable H. somi to induce lung infection in cattle. The prevalence of H. somni is more frequent in

cattle at the later stage of feedlot placement (Timsit et al., 2017).

Although Mycoplasma has been identified as one of the most predominant genera residing

in both nasopharynx and lower airways of cattle (Caswell et al., 2010), the pathogenesis of M.

bovis is less characterized compared to the other three BRD bacterial pathogens. M. bovis is

fastidious and it can only be grown with a special medium and culturing conditions, which makes

it difficult to characterize the pathogenic mechanisms of it employs in the development chronic

pneumonia (Panciera and Confer, 2010).

1.3 BRD management strategies

1.3.1 Preconditioning

Preconditioning has been used in some cow-calf operations in North America to reduce the

incidence of BRD (Thrift and Thrift, 2011; Taylor et al., 2010). Preconditioning programs vary

but typically involve the following processs:1) vaccination against Clostridial organisms shortly

after birth; 2) vaccination against respiratory viral and bacterial pathogenic agents 21-30 days prior to weaning; 3) weaning the calves 5-7 weeks in advance of sale, and training the calves to eat from the feed bunk and drink from a trough at the cow-calf operations; and 4) deworming (Bailey and

Stenquist, 1996). These preconditioning processes are aimed at increasing the calf’s disease resistance to respiratory infections at weaning, and to reduce stress associated with weaning and maternal separation, as well as to boost host defenses against respiratory viral and bacterial infections during the transition from cow-calf operations to commercial feedlots. All of which

ultimately reduce the incidence of BRD in cattle during feedlot placement. A study reported that

weaning 45 days before shipped to the feedlot reduced the risk of BRD in feedlot cattle as

compared to the calves shipped immediately after weaning (Step et al., 2008).

Although preconditioning programs are effective in mitigating BRD in feedlot cattle, the

costs associated with preconditioning are typically relegated to cow-calf producers who may not benefit and have little incentive to precondition. In addition, filling up larger pens in commercial feedlots requires mixing of calves from many different cow-calf farm sources. Furthermore, feedlots currently rely on metaphylactic antimicrobials to prevent BRD, which is another contributing factor for limiting the preconditioning practices.

1.3.2 Vaccination

Vaccination has long been shown to be a cost-effective strategy for controlling infectious diseases in livestock, when the vaccines are efficacious (Meeusen et al., 2007). Vaccination against both viral and bacterial agents associated with BRD in feedlot cattle is common practice in North

America. There are commercial vaccines available against M. haemolytica, P. multocida, H. somni and M. bovis (Larson et al., 2012). These vaccines are made from bacterins or killed whole bacterium. However, these conventional bacterial vaccines have shown limited efficacy with only partial protection against M. haemolytica and P. multocida and no benefit against H. somni (Larson et al., 2012). Part of the reason for this may be timing of vaccine administration. While larger feedlots may employ vaccination protocols for calves upon entry, immune-protection may not be achieved until some time afterwards, and may also be less effective given the stress the calves incur at placement. It is also possible that new vaccines are needed to achieve effective BRD protection, with older vaccines being less effective against pathogenic strains currently present in feedlots (Klima et al., 2018). 27

1.3.3 Antimicrobial use

Currently, prevention and control of BRD in feedlot cattle in North America are mainly targeted towards respiratory bacterial pathogens through the use of antimicrobials. Cattle considered at high-risk for the development of BRD are given metaphylactic antimicrobials at feedlot arrival (Ives and Richeson, 2015). Metaphylactic treatment reduces BRD-associated morbidity and mortality through eradicating existing bacterial infections, and preventing colonization and proliferation of pathogens in immune-suppressed and vulnerable animals.

Although effective in reducing BRD incidence in feedlots (Nickell and White, 2010; Ives and Richeson, 2015), indiscriminate use of antimicrobials is facing increased public and scientific scrutiny due to the selection of resistant BRD pathogens in Canadian and the American feedlots

(Timsit et al., 2017; Snyder et al., 2017). For example, a study evaluating commercial feedlots in

Alberta revealed that more than 60% of the M. haemolytica and P. multocida isolates from the lower respiratory tract of feedlot cattle with (n = 210) and without (n =107) BRD, exhibited multidrug resistance against antimicrobials used for metaphylaxis (i.e. tulathromycin and

oxytetracycline) (Timsit et al., 2017). In addition, a study conducted in the United States reported

a dramatic increase (from 4% to 99%) in the prevalence of M. haemolytica resistant to

tulathromycin in newly-received feedlot cattle (n = 169), within 2 weeks after metaphylactic

administration of tulathromycin (Snyder et al., 2017). Recent studies have also shown the impact

of metaphylactic treatment at feedlot entry on antimicrobial resistance determinants in the

nasopharyngeal (NP) microbiome. Holman et al. (2018) observed a significant increase in the

abundance of the tetracycline resistance gene tet(H) in genomic DNA extracted from the nasal

swabs of commercial feedlot cattle administrated oxytetracycline at feedlot entry. The same

authors also evaluated the longitidunal effects of oxytetracycline and tulathromycin on the 28

antibiotic resistance determinants in the NP microbiome of beef cattle entering the feedlot (Holman et al., 2019). In that study, oxytetracycline injection at feedlot entry resulted in an increased abundance of resistance genes erm(X) (macrolide resistance), sul2 (sulphonamide resistance),

tet(H), tet(M), and tet(W) in NP samples collected after 12 days. The cattle that received

tulathromycin had a significantly greater abundance of resistance genes erm(X), sul2, and tet(M)

compared to control cattle on day 34. These data indicate the association between metaphylactic

antimicrobial use and the development of antimicrobial resistance in BRD-associated pathogens

in feedlot cattle.

In addition to the link between the metaphylactic antimicrobial use and resistance in BRD

pathogens, metaphylaxis may also have negative impact on the homeostasis of the respiratory

microbiota. Administration of a single oxytetracycline or tulathromycin injection at feedlot entry

resulted in a noticeable change in both community structure and composition of NP microbiota

during the first 60 d following feedlot placement (Holman et al., 2018). Likewise, a significant

alteration of NP microbial community structure was observed in beef cattle in response to a

injection of oxytetracycline and tulathromycin within the first 5 days of treatment (Holman et al.,

2019). In that study, the authors noted an increase in Mycoplasma in oxytetracycline-treated

calves, however no studies have yet evaluated the impact of antibiotic-perturbed NP microbiota

on the respiratory health of feedlot cattle. In children, perturbations in the respiratory microbiota

have been linked to increased susceptibility to otitis media (Camelo-Castillo et al., 2019). Given

the increasing data showing the contribution of the mucosal microbiota to bovine respiratory

health, antibiotic-related perturbations of NP microbiota in feedlot cattle may compromise the

resilience and resistance of the respiratory tract against infection at later stages of feedlot

placement (Zeineldin et al., 2019). 29

Several studies have now identified multidrug resistant BRD pathogens from feedlot calves

(Klima et al., 2014; Noyes, et al. 2015; Timsit et al., 2017). In addition, multiple resistance genes

have been identified within integrative and conjugative elements of BRD pathogens, with potential

to be mobilized within and across species (Klima et al., 2014b; Bhatt et al., 2018; Klima et al.,

2019). Thus antimicrobial-resistant BRD pathogens pose a significant threat to effectively

mitigating and treating BRD in feedlots, should those pathogens carry resistance genes that provide

protection against the currently used classes of antimicrobials indicated for BRD.

1.4 Bovine respiratory defense mechanisms against bacterial pathogens

The bovine respiratory system consists of the upper (including nostrils, nasal cavity, and

pharynx) and lower (containing the larynx, trachea, bronchi and lungs) respiratory tracts. Based on the structure and function, the system is arbitrarily further divided into three continuous systems including the conductive, transitional, and gas exchange systems (López and Martinson, 2017).

The conductive system is made up of the nasal cavity, sinuses, larynx, trachea and bronchi. The main cells lining the conductive system are ciliated pseudostratified columnar epithelium cells, goblet cells and serous cells. The transitional system includes the bronchioles and is composed of mainly Clara cells and neuroendocrine cells, and sparsely distributed ciliated cells presenting at the proximal bronchiolar region. The gas exchange systems contains the alveoli which are lined by type I and type II pneumocytes on the surface (López and Martinson, 2017). Multiple layers of defense mechanisms including physical, biochemical, and cellular barriers have evolved in the respiratory tract in order to counter act pathogen attachment rapidly and efficiently (Figure 1.1).

1.4.1 Physical barriers:

The first defense mechanism that pathogens encounter in the respiratory tract is physical

barriers of the nasal cavity that are comprised of unique external and internal anatomical structures.

Hairs present in the external nares provide a physical barrier to inhalation of large particulate matter that may carry bacterial and viral pathogens (Srikumaran et al., 2008; Ackermann et al.,

2010). The conchae (turbinate bones) of the nasal mucosa prevent particles and pathogens from the nasal cavity to the lower respiratory tract by causing a vortex of inhaled air and enhancing the contact between infiltrating air and the nasal mucosa (Ackermann et al., 2010). In addition, mucosal clearance is important for limiting bacterial pathogen adherence and colonization (Lillie and Thomson, 1972). Mucosal clearance is one of the most important functions of the airway epithelium. It is accomplished by the effective collaboration of the airway surface liquid layer, ciliated epithelial cells, goblet cells, and cilia on the surface of ciliated cells (Figure 1.2)

(Bustamante-Marin and Ostrowski, 2017; López and Martinson, 2017). The air surface liquid

(ASL) is the thin layer of liquid solution lining between the airway epithelium and the gas in the lumen. It consists of two layers: a periciliary sol layer and a mucus gel layer. The sol layer is close to the apical cell surface and separates the mucus gel layer from the underlying epithelia

(Knowles and Boucher, 2002). The mucus gel layer which is mainly composed of goblet cells produces mucus traps of inhaled particulate matter including pathogens. The cohesive beating of all ciliated epithelial cells expel the trapped pathogens from the nasal passage to the throat or from the lower airway to the throat from where it is swallowed into the digestive tract (Ackermann et al., 2010; Tam et al., 2011; Bansil and Turner, 2018).

1.4.2 Biochemical barriers

Pathogens encounter numerous molecules presenting in the ASL that can mediate antimicrobial activity and immune response against invading pathogens. Mucin is the principal component of the mucus and transmembrane mucins can bind bacterial pathogens and facilitate the elimination of pathogens from the epithelial surface by mucin shedding (van Putten et al.,

2017). The ASL contains lactoferrin and lysozymes, both of which have known direct bactericidal activity (Travis et al., 1999). An array of antimicrobial peptides such as defensins and cathelicidins are also present in the respiratory tract that target pathogens and regulate host defenses (Bartlett et al., 2008). In addition, antimicrobial proteins play an important role in protecting the respiratory tract from pathogen invasion. Of the antimicrobial proteins, surfactant proteins A and D are found in the lung epithelial surfaces and can coat the surface of the pathogens and enhance their killing by phagocytic immune cells (Pastva et al., 2007). In addition, pathogen recognition receptors (PRRs) presenting in the airway epithelial surface mediate the airway epithelial cell recognition of inhaled pathogens and facilitate the immune cell response to pathogen associated molecular patterns (PAMPs) (Li et al., 2012). In response to the PRRs, pro- inflammatory cytokines and chemokines are secreted from the airway epithelial cells and initiate the killing and elimination of pathogenic agents by adaptive and innate immune cells (Kato and

Schleimer, 2007; Ackermann et al., 2010). Furthermore, immunoglobins A (IgA) and G (IgG) are present in bovine nasal secretions, and they also provide the airway system with a first line of immune defense against pathogens (Butlor, 1969; Nelson and Frank 1989). Secretary IgA, being the most predominant immunoglobin in nasal secretion, is involved in immune exclusion, and promotes the clearance of pathogenic bacterial and viral agents from the respiratory tract by entrapping pathogens in mucus, blocking their access to the epithelial receptors, and by modulating 32

the proinflammatory process (Pilette et al., 2004; Brandtzaeg et al., 2007). Bovine respiratory bacterial and viral pathogen-specific IgG also contribute to the protection of the upper and lower

respiratory tract against invasion (Singh et al., 2010; Ellis et al., 2018).

1.4.3 Cellular barriers

The respiratory mucosal epithelium contains a number of specialized epithelial and

immune cells that collectively protect the respiratory system against infection through interactions

with physical and biochemical barriers (discussed above), and the immunological responses

targeted at invasive pathogens (López and Martinson, 2017). Within the respiratory tract, an

array of immune cells such as dendritic cells, T and B cells, vascular endothelium, alveolar and

intravascular macrophages, NK cells, eosinophils, neutrophils, NK T cells, and Mast cells mediate

immune response to pathogen attachment (Ackermann et al., 2010). Of these immune cells,

macrophages and neutrophils play a more important role in the clearance of pathogenic agents

from the lower respiratory tract of cattle through phagocytosis. Alveolar macrophages, the first

phagocytic cells to encounter pathogens, initiate the killing of the pathogens by phagocytosis, and

the release of pro-inflammatory cytokines including IL-8 (Gordon and Read, 2002; Srikumaran et

al., 2008). As a result of neutrophil-attractant IL-8 release, the neutrophils are recruited to the

site of infection where they engage in killing pathogens by phagocytosis and by releasing a myriad

of antimicrobial molecules (Srikumaran et al., 2008; Bassel and Caswell, 2018). The macrophage

and neutrophils may also be involved in modulating adaptive immune responses (Gordon and

Read, 2002; Bassel and Caswell, 2018).

1.4.4 The respiratory commensal bacteria

Overall balance between host response, commensal microbiota, and pathogen prevalence

is the major determinant of heath and disease (Libertucci and Young, 2019). The commensal

microbiota plays an important role in maintaining this balance by restricting colonization of

potential pathogens and by being involved in the maturation and maintenance of homeostasis of

physiology and immunity (Man et al., 2017; de Steenhuijsen Piters et al., 2015).

The commensal microbiota can directly affect pathogen colonization through competition

for nutrients and production of bacteriocins (Hand, 2016). Bacteriocins can have broad or narrow

spectrum and function by altering nucleic acids and protein metabolism in Gram-negative bacteria, or the cell envelope of Gram-positive bacteria (Cotter et al., 2013). In humans, the abundance of bacteriocins were greatest in vaginal, airway, and oral metagenomes, with a lower prevalence in the digestive tract (Zheng et al., 2015). Thus bacteriocin production is important in establishing bacterial dominance within respiratory communities and therefore pathogen resistance (Dobson et al., 2012; Kommineni et al., 2015).

Commensal bacteria can also mediate pathogen resistance indirectly through modulation of the host innate and adaptive responses. Antimicrobial peptides (AMPs), secreted from mucosal surfaces are important component of innate immune defense against pathogens (Price et al., 2019).

The production of AMPs in epithelial cells has been reported to be stimulated by short chain fatty acids (SCFAs) produced by commensal microbiota (Zhao et al., 2018). Microbiota-derived

SCFAs are also known for enhancing the epithelia tight junctions that are often exploited by the pathogenic bacteria (Libertucci and Young, 2019). Other mechanisms though which commensal microbiota mediate innate immunity are regulating mucin and IgA secretion (Kubinak et al., 2015) and the development of antigen presenting cells (Dendritic cells and macrophages) and phagocytic activity of neutrophils and other innate immune cells (Wu and Wu, 2012). Commensals also mediate the induction of effector T and B cells against pathogens by modulating the functions of dendroid cells or other innate immune cells (Belkaid and Hand, 2014). In addition, T cell fate 34

(Furusawa et al., 2015) and host antibody response (Kim et al., 2016) can be modulated by

commensal microbiota.

Some commensal strains can suppress host inflammatory responses and thereby decrease

inflammation associated with pathogen infection (Buffie and Pamer, 2013). Metabolites,

particularly immune-modulins, produced by commensal bacteria affect the function of epithelial

and mucosal immune cells and can stimulate anti-inflammatory processes (Hemarajata and

Versalovic, 2012). Microbiota-derived immune-modulins have shown to stimulate the production

of transforming growth factor-β in epithelial cells while down regulating the expression of pro-

inflammatory cytokines (TNF, IL-6 and IL-12) in macrophages and dendritic cells, resulting in

suppression of the pro-inflammatory responses.

1.5 Bovine respiratory microbiota

1.5.1 Structure and composition of respiratory microbiota in beef cattle described using 16S rRNA high-throughput sequencing

NP microbiota: given its accessibility and by being the primary niche for opportunistic

BRD pathogens to colonize and proliferate, the nasopharynx of feedlot cattle has been the primary

target for characterizing the microbial community in the respiratory tract. The NP microbiota of

feedlot cattle contains a rich and diverse bacterial community, harboring approximately 29

different phyla and 300 different genera (Timsit et al., 2016a). Proteobacteria, Firmicutes,

Actinobacteria, Bacteriodetes and Tenericutes are the predominant phyla and constitute over 90%

of total NP microbiota (Holman et al., 2015a, 2017 and 2019; Zeineldin et al., 2017a; Timsit et al.,

2018;). The most common genera include Corynebacteria, Moraxella, Mycoplasma, Pasteurella,

Mannheimia, Psychrobactor and Staphylococcus (Holman et al., 2015a and 2017; Timsit et al.,

2018; Zeineldin et al., 2017a). Although the nasopharynx is colonized predominately by aerobic

(aerotolerant) bacteria populations, some obligate anaerobic species such as Prevotella,

Clostridium, Bacteriodes, Bifidobacterium and Ruminococcus commonly present in the rumen of

cattle have also been detected from NP swabs, likely as a result of rumination. The proportions of

different bacterial communities in the NP varies between individual animals over time.

Lower airway Microbiota: compared to the NP microbiota, the microbial community

residing within the lower respiratory tract of feedlot cattle has been less characterized due to the

invasiveness and difficulty of sample collection. Trans tracheal aspiration (TTA) and

bronchoalveolar lavage (BAL) fluid samples have mainly been used to characterize the lower

airway microbiota of feedlot cattle. The tracheal microbiota has been evaluated in healthy and

BRD-diagnosed steers from commercial Albertan feedlots (Timsit et al., 2018) and weaned calves

from smaller farms in Italy (Nicola et al., 2017). In addition, tracheal samples from healthy

Albertan calves that were sampled longitudinally before and after shipment to an auction market,

and then again after feedlot placement, have been analyzed (Streobel et al., 2018). According to

the 16s rRNA sequencing of TTA samples obtained from these studies, a diverse microbial

community is present in the trachea of cattle, with colonization by 9-21 different bacterial phyla

and 91-182 different genera. Tenericutes (> 50% relative abundance), Firmicutes, Proteobacteria

and Actinobacteria are the most predominant phyla in the trachea. The most relatively abundant bacterial genera include Mycoplasma (> 50%), Moraxella, Pasteurella, Lactococcus, Histophilus

and Bacteroides. Similar to the tracheal microbiota, bronchoalveolar microbial communities from

healthy feedlot cattle contained mostly Proteobacteria, Bacteroidetes, Actinobateria, and

Tenericutes (Zeineldin et al., 2017b).

Overall, most of the predominant phyla and genera identified from the tracheal (Nicola et al., 2017; Timsit et al., 2018) and bronchoalveolar (Zeineldin et al., 2017b) samples have been shown to also be present in NP samples from the same cattle. However, the structure of bacteria from the lower respiratory tract differed from those in the nasopharynx and also had lower bacterial richness and evenness compared to NP microbiota (Nicola et al., 2017; Timsit et al., 2018). The lung microbiota has a distinct microbial community structure compared to the NP microbiota, but is influenced by the NP microbiota through inhalation of bacteria originating from the upper respiratory tract (Nicola et al., 2017). In addition, the microbiota is influenced by physiological gradients along the respiratory tract. These physiological gradient factors include the pH, relative humidity, temperature, and partial pressure of oxygen and carbon dioxide, which vary across the respiratory tract (Man et al., 2017).

1.5.2 Management factors that influence the respiratory microbiota of beef cattle

Several studies have shown that the NP microbiota of beef calves changes after transportation to a feedlot. In one study that sampled the nasopharynx of calves at weaning, upon feedlot arrival, and 40 days after feedlot placement, it was shown that the NP microbiota underwent a profound evolution, with the abundance of 92 operational taxonomic units changing over time

(Timsit et al., 2016b). In a subsequent study that focussed on shorter time points after feedlot placement, the structure and composition of the NP microbiota was observed to change within two days of feedlot placement, increasing in both phylogenetic diversity and richness (Holman et al.,

2017). In both those studies by Timsit (Timsit et al., 2016b) and Holman (Holman et al., 2017), cattle were not administered antimicrobials or implants which could have biased the results.

Interestingly, transportation to and commingling at an auction market for 24 h did not significantly influence NP or tracheal bacterial communities in recently weaned beef calves (Stroebel et al., 37

2018), indicating that feedlot introduction, rather than an auction market, has a strong influence in shaping respiratory microbiota throughout the beef continuum. It should be noted however that both farm of origin and feedlot practice can influence the types of microbiota that colonize the respiratory tract (McMullen et al., 2018; Stroebel et al., 2018), thus the changes in microbiota that occur upon feedlot arrival is not necessarily the same in all cattle. The instability in respiratory microbiota observed after feedlot placement might explain why cattle are most likely to be affected with BRD during the first weeks after weaning and arrival at a feedlot. Factors that may lead to changes in the respiratory microbiota may include a reduction in calf immunity due to stress from weaning and transportation, and colonization by bacteria originating from the feedlot environment or new pen mates (Timsit et al., 2016b). In addition, diet transition before or after feedlot placement may influence the respiratory microbiota (Hall et al., 2017).

Administration of antimicrobials can also affect the microbiota by inhibiting growth of certain bacteria, and potentially promoting the growth of others that have intrinsic or acquired resistance to the antimicrobial. In children, antibiotic use has been linked to an altered microbial community structure in the upper respiratory tract for up to six months after administration (Santee et al., 2016), indicating that this practice has a prolonged impact. Recently, it was observed that alterations in the NP microbiota of commercial cattle were apparent 60 days after injection with either oxytetracycline or tulathromycin (Holman et al., 2018). In a controlled study analyzing the effects of these same two antimicrobials on the NP microbiota across 34 days, perturbation of the

NP microbiota was greatest two and five days after administration (Holman et al., 2019). In the study by Holman (Holman et al., 2019), it took 12 days for the NP microbiota to recover after tulathromycin injection, whereas recovery was not apparent after 34 days for oxytetracycline- treated cattle. Interestingly, shortly after administration (2-5 days), both antimicrobials reduced 38

the abundance of Pasteurella spp.. However, one Mycoplasma sequence variant was enriched in

oxytetracycline-treated cattle at the end of the study (d 34). Oxytetracycline has previously been

associated with higher rates of chronic pneumonia and polyarthritis syndrome in feedlot calves,

which are M. bovis-associated syndromes (Hendrick et al., 2013). While indirect effects of

antimicrobials on respiratory microbiota and disease development have not been researched in

cattle, these studies show that administration of antimicrobials to cattle can have short- and long-

term impacts on the NP microbiota.

1.5.3 Potential association between the respiratory microbiota and development of BRD

Evidence of bacterial competition within the bovine respiratory tract was first shown by

Corbeil and colleagues (1985), when they observed that bacteria from the nasopharynx either enhanced or limited in vitro growth of the BRD pathogens M. haemolytica, P. multocida, and H. somni. Enhancing bacteria included Micrococcus, Staphylococcus, Corynebacterium,

Rhodococcus, Moraxella, and Actinobacter isolates, while isolates of Bacillus were the strongest

inhibitors. Subsequently, studies have shown associations between the bovine respiratory

microbiota and BRD. Holman et al., (2015a) characterized the NP microbiota of feedlot cattle that

remained healthy to those that developed BRD after feedlot placement. The cattle were sampled

by NP swab when they entered the feedlot. The NP bacteria of cattle that developed BRD had

significantly lower diversity and richness compared to those that remained healthy. Although the

NP microbiota of both healthy and BRD-affected cattle shared the predominant phyla

Proteobacteria and Firmicutes, the microbial communities were different among these calves at

the family level, with healthy having a significantly greater abundance of Micrococcaceae,

Lactobacillaceae, and Bacillaceae upon feedlot entry. Another study evaluated the NP microbiota

of feedlot cattle at entry into a commercial feedlot, during initial handling, and processing at the

feedlot, and upon diagnosis of BRD (Zeineldin et al., 2017a). In this study, cattle diagnosed with

BRD harbored a distinct NP microbial community from healthy cattle at feedlot entry, and when

diagnosed with BRD. At the genus level, BRD-diagnosed cattle had a greater relative abundance of Acinetobacter, Solibacillus and Pasteurella than healthy calves. Thus the studies by Holman et al. (2015) and Zeineldin et al. (2017a) indicate that the NP microbiota at feedlot entry are important for the respiratory health of feedlot cattle after placement. In a case control study, Timist et al.

(2018) compared the NP and tracheal microbiota between feedlot cattle diagnosed with BRD and matching pen mates that were healthy. The BRD-affected cattle exhibited lower bacterial diversity in both the upper and lower respiratory tracts compared to healthy pen mates. Additionally, distinct bacterial metacommunities were observed to be associated with upper or lower respiratory tract, as well as BRD status. One metacommunity included Lactococcus lactis and Lactobacillus casei, and was mostly associated with the trachea of healthy calves. Similarly, analysis of post- mortem lung samples collected from fatal cases of BRD in dairy calves revealed a relatively distinct microbial community among the lung tissues of BRD-affected and clinically healthy animals (Johnston et al., 2017). In support of this, Lima et al. (2016) also observed that dairy calves that developed pneumonia harbored a significantly higher relative abundance of Mannheimia,

Moraxella, and Pasteurella in the upper respiratory tract compared to healthy calves. The authors also showed that the calves that developed pneumonia had increased NP total bacterial loads compared to the healthy animals at 3 days of age. Overall, these studies highlight that differences in respiratory microbiota have been associated with BRD status in cattle. Given these observed associations, microbiota-based applications could potentially be used for both diagnosis and prevention of BRD. 40

1.6 Probiotics

Probiotics are defined as “live microorganisms administered in adequate amounts that

confer a beneficial health effect on the host” (FOA/WHO, 2002). While typically associated with

delivery in food or feed, and administration to enhance gastrointestinal health, by definition

probiotics can also be targeted for extra-intestinal application including the oral cavity, and

respiratory and urogenital tracts (Caramia and Silvi, 2011). The use of probiotics in livestock dates

back decades (Vanbelle et al., 1989) however increased interest in their application resulted from

countries banning the use of in-feed subtherapeutic antimicrobials for growth promotion of animals. The bacterial species most commonly used as probiotics are within the genera

Lactobacillus and Bifidobacterium due to their history of safe use and ease of large-scale production (Bogovič-Matijašić and Rogelj, 2011).

1.6.1 Mechanisms of probiotic action

Probiotics can deliver their beneficial effects to the host through multiple mechanisms, and in some instances, a combination of methods (Figure 1.3). They may directly inhibit pathogens by producing lactic acid and hydrogen peroxide, which inhibit pathogenic bacteria by lowering the pH and inactivating biomolecules, respectively (Surendran Nair et al., 2017). Bacteriocin production is also an important mechanism by which probiotics inhibit pathogen growth (Gaspar et al., 2018) and impede the pathogenesis. For example, the bacteriocin-producing Lactobacillus salivarius UCC118 strain was able to protect mice against infection by Listeria monocytogenes, whereas a bacteriocin-mutant L. salivarius strain did not offer confer L. monocytogenes protection

(Corr et al., 2007). Probiotics can prevent adherence of pathogenic bacteria to the host mucosa cells through competitive exclusion by competing against pathogens for common binding sites and the utilization of nutrients (Monteagudo-Mera et al., 2019). Probiotic strains that are strongly

adhesive to the target mucosa have greater antagonistic competition effects against pathogenic

bacteria (Wall, 2008; Oelschlaeger, 2010).

Probiotic Lactobacillus strains have been shown to be involved in preservation of epithelial

barrier function by stimulating mucin secretion, strengthening tight junction and preventing

epithelial cell death, thereby inhibiting pathogen translocation (Lebeer et al., 2008). In addition,

probiotics are involved in modulation of host innate and adaptive immune response towards

pathogens, enhancing the detection and elimination of the pathogens by immune cells (Kemgang

et al., 2014; Malago et al., 2011). Some probiotic strains have also been shown to regulate host

inflammatory responses by suppressing the production of pro-inflammatory cytokines while

stimulating anti-inflammatory cytokines (Li et al., 2014; Vanderpool et al., 2008). Of note, the

immune modulation properties of probiotics are strain-specific, and therefore not common to all

strains within a species (Lebeer et al., 2008). Probiotics may also re-establish the composition of

the gut microbiota and confer beneficial effects on gut microbial communities (Sanchez et al.,

2010; Hemarajata and Versalovic, 2013).

1.6.2 Selection criteria and requirements for probiotic strains

Bacteria that are intended to be used as probiotics must meet the selection criteria for functionality, safety, and technological usability (Markowiak and Śliżewska, 2018). When searching for new probiotic candidates, the origin of the strains, and the host and target ecological site for which they are intended for must be taken into consideration. Probiotic properties of certain bacteria are typically host-specific, therefore, the candidate strains should be isolated from the same host species to ensure better probiotic efficacy (Markowiak and Śliżewska, 2018). The successful colonization and adaptation of a probiotic bacterium to the target niche relates to its origin. Allochthonous probiotics, which are bacteria non-indigenous to the host microflora, are 42

less adapted for successful colonization than autochthonous probiotics originating from the host target microbial community (Shewale et al., 2014). In addition, probiotics developed from healthy individuals of the same species to which they will be applied are less likely to cause health and safety risks.

The ability of probiotic bacteria to adhere to target mucosal sites, and their ability to establish within the target microbial community, is a prerequisite for colonization. Therefore, screening for the adherence of the candidate strains using cell models and ex vivo models is an important probiotic selection criterion (Salminen et al., 1996; Tuomola et al., 2001). If the application of probiotics is intended to ameliorate certain infections, demonstrating the antimicrobial activities of probiotics against infectious agents in vitro is also required. These antimicrobial properties include direct inhibition against pathogen growth, and competing for binding sites (Dunne et al., 2001; de Melo Pereira et al., 2018). Modulation of host immunity is an additional mechanism by which therapeutic bacteria enhance host resistance to pathogens (Yan and Polk, 2011), and it is often included as part of the selection criteria (de Melo Pereira et al.,

2018).

Another important criterion for probiotic selection and regulatory approval is screening candidate bacteria for acquired resistance elements which could be mobilized to host microbiota

(Bajagai et al., 2016; Gueimonde et al., 2013). Probiotic strains must also meet regulatory safety requirements by having no history of association with infectious disease and adverse effects due to excessive immno-stimulation (Doron and Snydman, 2015; Markowiak and Śliżewska, 2018).

In addition to the safety and functional criteria, the ability to survive and maintain metabolic activity, resistance to suboptimal physiological conditions, viability during large scale production and storage, genetic stability, as well resistance to bacteriophages should also be considered as 43

part of the probiotic selection criteria (Markowiak and Śliżewska, 2018)

1.6.3 Probiotic use in cattle

The primary expected beneficial outcome of probiotic use in animal nutrition is to increase

average daily gain in animals including ruminants, monogastric animals and poultry

(Chaucheyras-Durand and Durand, 2010). In ruminants (cattle, sheep and goat), probiotics have

mostly been used in the diet to target the rumen compartment and promote growth. The proposed

mode of probiotic action in the rumen is to modulate the rumen microbial ecosystem and thereby

enhance food digestion and metabolism, ultimately providing more energy for net growth

(Chaucheyras-Durand and Durand, 2010). Probiotics have also been used to reduce the risk of rumen acidosis by stabilizing the ruminal pH (Marden et al., 2008) or to decrease the fecal shedding of human pathogens, such as Escherichia coli O157 (Tabe et al., 2008) or Salmonella

(Stephens et al., 2007). Mitigation of methane excretion by ruminants has also been the target for probiotic development, which in itself would result in a net increase of digestible energy for growth

(Newbold and Rode, 2006). Live yeasts are most frequently used as probiotics in ruminants and

are commercially available (Chaucheyras-Durand and Durand, 2010). A meta-analysis of the

effects of yeast culture in dairy cattle found that milk and fat-corrected milk yields are improved

by yeast (Poppy et al., 2012). There are also data to show that yeast can improve feed efficiency

and alter rumen metabolism (Alugongo et al., 2017). Apart from yeast, the majority of probiotics

evaluated for cattle have shown variable responses (Cameron and McAllister, 2019). This likely

results from the fact that most probiotic strains originated from the gut or fermented foods, and

were selected based on their ability to grow vigorously and survive production, rather than

confirmation of health benefits (O’Toole et al., 2017). Additionally, many strains used for

livestock may not be adapted for agricultural animals as they were first developed for human use. 44

1.6.4 Targeted development of bacterial therapeutics

Research on probiotics has significantly increased over the last 10 years, and has coincided

with improved methods to culture bacteria and reduced costs of massive parallel sequencing

technologies to analyze microbiota. These technologies have allowed in-depth study of host

metagenomics and are providing important information on bacteria associated with certain

phenotypes, such as diseases like BRD (Lynch and Hsiao, 2019; Zeineldin et al., 2019). There are

ongoing efforts to identify strains of bacteria within a host’s microbiota and developing them for

therapeutic application to potentially alter a phenotype after administration or prevent infections by pathogens. Strains of this nature however likely fall outside of the traditionally defined probiotics (e.g. Lactobacillus, Bifidobacterium) and would be used for therapeutic purposes. As such, they have been coined “next-generation probiotics” (Patel and DuPont, 2015), or bacterial

therapeutics (BTs) (Bentley and Sebaihia, 2007). Although BTs conform to the definition of a

traditional probiotic, the strains identified as BTs can be non-traditional microorganisms that have

not been used to promote health, and typically lack a history of safe use (O’Toole et al., 2017).

The BTs can also be developed from strains that taxonomically belong to the same species as

pathogens. Examples of BTs include strains that have originated from the human gut within the

genera Bacteroides (Deng et al., 2016), Clostridium (Woo et al., 2011) and Faecalibacterium

(Sokol et al., 2008). The BTs can be categorized into two types: 1) strains that are used to abrogate

the disease phenotype and improve health and 2) strains, previously well characterized as

traditional probiotics, which are used as delivery vehicles to deliver specific antimicrobial/health

promoting molecules to the host and thereby ameliorate disease (O’Toole et al., 2017). In cattle,

the therapeutic application of bacteria has been explored against bovine mastitis (Assis et al., 2015;

Beecher et al., 2019) and reproductive health (Genís et al., 2018). 45

The process of developing BTs involves several steps. The first step is to identify BT target candidates either by the absence or depleted abundance of certain strains in disease affected subjects, or by a known record of modulating microbiota composition or health promotion. Then, the targeted strains are screened for desired probiotic activities in vitro (cell models) and in vivo

(animal models). As part of the safety assessments, the candidate NGPs are subjected to whole genome sequencing, and genomes are screened for presence of transmissible antibitiotic-resistance genes, and potential virulence factors. In addition to the initial genome screening, safety and toxicity assessments in animal models are also required (O’Toole et al., 2017; O’Toole and Paoli,

2017). Developed BTs are regulated by the same bodies that regulate traditional probiotics and they include the European Food Safety Authority and FDA Center for Biological Evaluation and

Research (O’Toole et al., 2017).

1.7 Conclusion

Despite continued attempts to mitigate BRD, the prevalence of this disease in feedlot cattle remains a significant economic and welfare problem. Given the emergence of multidrug-resistant bacteria associated with BRD and growing public demand for limiting the use of antimicrobials in food animals, there is an impetus to develop alternatives to metaphylactic antimicrobials. As a major pathogen, M. haemolytica is of interest for mitigation. This bacterium is an opportunistic pathogen and can proliferate in the upper respiratory tract of stressed calves, consequently causing bronchopneumonia after translocation to the lung. While the respiratory microbiota is now established as important to cattle health, this microbiota undergoes significant change throughout the beef cattle life cycle, potentially leading to increased susceptibility to BRD after feedlot placement. However, the identification of bovine respiratory bacteria associated with inhibition

of BRD pathogens highlights that the microbiota of cattle may include candidates for the development of BTs, to mitigate M. haemolytica through intranasal inoculation.

The intent of this project was to identify and characterize BT candidates originating from the respiratory tract of healthy feedlot cattle, which could be used for inhibition of M. haemolytica.

Lactic acid-producing bacteria were targeted for BT development and a serotype 1 strain of M. heamolytica was used as a model pathogen. For this, multiple objectives were defined, as outlined below.

1.8 Hypothesis

Intranasal BTs developed from the NP bacteria of healthy feedlot cattle can be used to

prevent colonization and proliferation of the bovine respiratory pathogen, M. haemolytica and

improve stability of the respiratory microbiota in beef cattle.

1.9 Objectives

1) To evaluate the potential of commercially-sourced probiotic bacteria to inhibit M.

haemoltyica growth and adhesion to bronchial epithelial cells (Chapter 2).

2) To characterize longitudinal changes in the NP microbiota, particularly LAB, of steers that

were transported to an auction before feedlot placement (Chapter 3).

3) To develop BTs originating from the respiratory tract of healthy cattle, for mitigation of

M. haemolytica, using a step-wise approach based on pathogen inhibition, cell adherence,

immunomodulation, and antimicrobial susceptibility criteria (Chapter 4).

4) To test the in vivo effectiveness of the best BT candidates for reducing colonization by M.

haemolytica in experimentally challenged calves (Chapter 5). 47

5) To investigate the longitudinal effects of intranasally administered BTs on BRD-associated

and commensal bacteria in post-weaned beef calves (Chapter 6).

1.10 Tables and Figures

Figure 1.1 Defense mechanisms of bovine respiratory tract against pathogens.

Figure 1.2 Mucosal clearance mechanism of respiratory tract.

The respiratory tract mucosa is consisted of basement membrane, one or more layers of ciliated

epithelial cells, and goblet cells which produces mucin, and mucus. Mucus consists of a superficial

gel layer and a liquid or periciliary fluid layer (sol layer) that bathes the epithelial cilia. These 2

layers are separated by a thin layer of surfactant. Dust particles and bacterial pathogens are trapped

by the mucus gel layer, and are removed from the respiratory tract by constant movement by cilia beating of ciliated epithelial cells (Lopez and Martinson, 2017).

Figure 1.3 Probiotic mechanisms of action.

Probiotic bacteria can confer their antimicrobial actibity against pathogenic bacteria by competing for binding sites and nutrients, competitive exclusion, enhancing the epithelial barrier, as as well as immune stimulation.

Chapter Two: Pobiotic bacteria inhibit the bovine respiratory pathogen Mannheimia

haemolytica serotype 1 in vitro

Chapter 2 has been published in Letters in Applied Microbiology.

Amat S, Subramanian S, Timsit E, Alexander TW. 2017. Probiotic bacteria inhibit the bovine respiratory pathogen Mannheimia haemolytica serotype 1 in vitro. Lett Appl Microbiol, 64: 343-349.

2.1 Introduction

Bovine respiratory disease (BRD) accounts for economic losses that surpass those incurred

by all other diseases of beef cattle combined, contributing to 65–80% of morbidity and 45–75%

of mortality cases in feedlots (Duff and Galyean, 2007). Treatment and control of BRD in the beef

sector are aimed mainly at bacterial pathogens, through use of antimicrobials. However, studies

have shown that antimicrobial resistance in BRD pathogens has increased (Portis et al., 2012) and multiresistance has been identified on mobile elements (Klima et al., 2014; Clawson et al., 2016) which may reduce the efficacy of current antimicrobials used to treat BRD. Novel methods to mitigate BRD-related pathogenic bacteria are therefore needed. Mannheimia haemolytica is an opportunistic pathogen commonly associated with BRD. Suppression of the host’s immune system due to stress or viral infection predisposes to rapid growth of M. haemolytica in the upper respiratory tract and reduced clearance in the lower respiratory system, initiating pathogenesis

(Griffin et al. 2010). While host immunity is important in controlling pathogens, there is evidence showing a bacterial component to regulating microbial populations in the bovine respiratory tract.

For example, aerobic bacteria isolated from the bovine upper respiratory tract have been shown to both inhibit and enhance growth of M. haemolytica in vitro (Corbeil et al., 1985) and

Bibersteinia trehalosi (Dassanayake et al., 2010) and Pasteurella multocida (Bavananthasivam et al., 2012) have previously been described to inhibit M. haemolytica by proximity-dependent mechanisms. Thus, bacterial competition has an important role in the respiratory tract and perhaps commensals could be used as probiotics to inhibit pathogenic bacteria.

Although typically associated with ingestion and treating gastrointestinal disorders, probiotics have also been investigated for beneficial effects at other sites. For example, in situ application of Lactobacillaceae strains has been shown to inhibit bacterial pathogens in the human 53

respiratory tract (Burton et al., 2006). More recently, a probiotic was administered to humans by

nasal spray and shown to colonize the rhinopharynx in 95% of subjects (Santagati et al., 2015),

highlighting the upper respiratory tract as a potential target for probiotics. The purpose of this

study was to evaluate probiotic bacteria for their in vitro ability to inhibit growth and adhesion of

M. haemolytica to bovine bronchial epithelial (BBE) cells.

2.2 Results and discussion

2.2.1 Inhibition of M. haemolytica serotype 1 (S1) by probiotic bacteria

The bacteria used in this study are described in Table 2.1. Except for S. thermophilus, both

cell culture and cell-free culture supernatant from each probiotic strain inhibited M. haemolytica

S1, with inhibition zones ranging from 12 to 19 mm (Table 2.2). The highest inhibition of M.

haemolytica was observed with cell cultures of the P. polymyxa strains, Lactobacillus rhamnosus,

Lactobacillus planterum, Lactobacillus casei and Lactobacillus acidophilus (inhibition zones ≥15 mm). Probiotic bacteria can exert growth inhibitory effects on target pathogens through secretion

of antimicrobial substances such as organic acids, bacteriocins and H2O2 (Pridmore et al. 2008).

They can also inhibit through physical interactions such as auto-aggregation, steric hindrance and

competition for binding sites (Popova et al., 2012). In the present study, the cell-free supernatants

of all antagonistic strains showed similar inhibition patterns against M. haemolytica S1 as their

corresponding cell culture, suggesting that these bacteria can inhibit M. haemolytica S1 through

secretion of antimicrobial substances. Inhibitory effects of probiotic bacteria against human

respiratory pathogens have been reported previously (Santagati et al., 2012). However, there is limited information on bacteria that inhibit BRD pathogens, including M. haemolytica. Corbeil et al. (1985) observed that strains of Streptococcus and Bacillus isolated from the bovine respiratory

and reproductive tracts were capable of inhibiting M. haemolytica in vitro. Although the authors

did not test any Lactobacillaceae, we have observed a reduced relative abundance of nasopharyngeal Lactobacillus in cattle that develop BRD (Holman et al., 2015a), suggesting that these bacteria are important in maintaining a healthy respiratory microbiota. Species of

Lactobacillus have previously been reported to have inhibitory effects against multiple bacterial

pathogens. For example, Perea Velez et al. (2007) reported that L. rhamnosus GG and L. casei

isolated from yogurt inhibited Salmonella typhimurium (zones of inhibition ≥12 mm), when analysed using a similar spot-on-lawn method to that used in the current study. In our study, P.

polymyxa displayed the strongest inhibitory effect against M. haemolytica. The P. polymyxa JB-

0501 strain has also been shown to inhibit Escherichia coli, and has been proposed as a potential direct fed microbial in livestock feed (Naghmouchi et al., 2013).

2.2.2 Bacterial adhesion to BBE cell monolayers

The mean adhesion of M. haemolytica S1 to BBE cell monolayers was 8.3% (Figure 2.1).

This was within the reported range observed for M. haemolytica isolated from ovine (7.8% adhesion) and bovine (16% adhesion) sources, when adhered to ovine bronchial epithelial cells

(Haig, 2011). The Lactobacillus strains used in this study, as well as Lactococcus lactis, displayed

greater adherence to BBE cell monolayers (10.5–17.8%; P < 0.05), compared with M. haemolytica

S1 (8.3%), S. thermophiles (2.2%) and the two P. polymyxa strains (1.6 and 0.25%). Similar to our study, Koo et al. (2012) reported greater adhesion rates for L. rhamnosus GG (24%), L. plantarum

(17%), L. acidophilus (16%) and L. casei (12%) to human intestinal CaCo-2 cells, in comparison with other LAB species. The differences in adherence between each species may be related to their cell surface hydrophobicity. It has previously been shown that Lactobacillus strains with greater

cell surface hydrophobicity displayed stronger adhesion to human intestinal cells (Duary et al.,

2011). Adhesion and colonization by probiotic bacteria at the target site of the host are important

features that affect their ability to inhibit pathogens and modulate the host immune system

(Collado et al., 2006; Wall et al., 2008). In feedlot production, probiotics that colonize the

respiratory tract long term would be optimal because cattle are generally processed two times, at

entry and at reimplantation (i.e. 80-100 days after entry), limiting the ability for recurring

intranasal administration. Of the probiotics tested, the Lactobacillus strains appeared to have the

greatest potential to colonize the bovine respiratory tract. However, because BBE cells were used

for adhesion assays, further in vitro studies with cells representative of the upper respiratory tract,

or in vivo models, are needed to verify colonization by the probiotics described.

2.2.3 Antagonistic activity of probiotic bacteria against M. haemolytica S1

Both P. polymyxa strains and L. acidophilus caused the greatest displacement of M.

haemolytica S1 from the BBE cell monolayers (reduction of 1.8 - 2.0 log10 CFU per ml; Table 2.

3). The only other probiotic to displace M. haemolytica S1 was L. helveticus. Similarly, these four

strains reduced M. haemolytica adhesion to BBE cells through competition (1.7 - 2.4 log10 CFU

per ml), as did L. casei, but to a lesser extent (0.4 log10 CFU per ml). In contrast, S. thermophilus

caused a slight increase in adhesion of M. haemolytica to BBE cells that was significant in the

competition assay (increase of 0.3 log10 CFU per ml), suggesting that this strain might enhance growth or adhesion of M. haemolytica. Interestingly, Streptococcus viridans has also been shown

to enhance the growth of M. haemolytica in vitro (Corbeil et al., 1985).

One of the important aspects of probiotic bacteria is to protect the host tissue from invading

pathogens (Reid et al., 1990). From our results, the P. polymxya strains and L. acidophilus showed

the greatest ability to reduce colonization of BBE cells by M. haemolytica S1 through displacement and competition. Although L. acidophilus had a lower growth inhibitory potential compared with other probiotics (Table 2.2), it did have one of the higher adhesion rates to BBE cells (Figure 2.1), perhaps enhancing its ability to antagonize M. haemolytica S1. Guglielmetti et al. (2010) reported that strongly adhesive probiotic bacteria had a greater antagonistic effect towards Streptococcus pyogenes, compared with lower adhering probiotics. However surprisingly, in our study, the P. polymyxa strains had the lowest BBE cell adherence yet displayed the strongest antagonistic exclusion against M. haemolytica. The reason for this remains unclear, although it may be related to the antimicrobial factors produced by P. polymyxa, as both these strains also showed strong growth inhibition (Table 2.2).

2.3 Conclusion

In conclusion, with the exception of S. thermophilus, each of the probiotics strains tested inhibited M. haemolytica S1. While adhesion of probiotics to BBE cells was variable, it did not correlate with antagonism of M. haemolytica S1 on BBE cells. Although probiotics are typically used as direct fed microbials in the beef industry, this in vitro study demonstrated for the first time the possibility of mitigating BRD pathogens using nasal probiotics. Future work is needed to further characterize the antimicrobial activities of probiotics against BRD pathogens, and whether commercial application can offer a mitigation strategy to reduce BRD bacterial pathogens, in place of metaphylaxis.

2.4 Materials and methods

2.4.1 Bacterial strains and culture conditions

Bacteria used in this study are listed in Table 2.1. Lactobacillus strains were cultivated in

Lactobacilli MRS broth (Difco, Detroit, MI, USA) supplemented with 0.05% L-cysteine- HCl and

incubated at 37°C for 24-48 h. Paenibacillus polymyxa and Lactococcus lactis were grown

overnight in tryptic soy broth (Difco) at 37°C. Streptococcus thermophilus was grown in brain–

heart infusion (BHI) broth (Difco) at 37°C for 24 h. Mannheimia haemolytica S1 was grown

overnight in blood agar (TSA) with 5% sheep blood (Dalynn Biologicals, Calgary, AB, Canada)

at 37°C.

2.4.2 Growth inhibitory effects of probiotic bacteria against M. haemolyica S1

Probiotic inhibition of M. haemolytica S1 was determined using a spot-on-lawn assay

described by Jacobsen et al. (1999). Overnight culture of M. haemolytica S1 was suspended in

Dulbecco’s phosphate-buffered saline (DPBS) and the absorbance at 625 nm was adjusted to

standardize bacterial concentrations (1-2 × 108 CFU per ml). A 100 µl aliquot of M. haemolytica culture was spread-plated on blood agar plates. Similarly, the absorbance values of overnight to

24-h cultures of probiotic bacteria were adjusted to obtain 107 to 108 CFU per ml. A subsample of

probiotic bacterial broth cultures was sterile-filtered (0.2 µm) to test the inhibitory effects of

cell-free broth. Triplicate 50-µl samples of each sterile-filtered and nonfiltered probiotic bacterial

culture were spotted on the surface of M. haemolytica S1 lawns. Each triplicate set of samples was

tested on an individual plated lawn of M. haemolytica S1. The plates were incubated at 37°C with

5% CO2 overnight, then zone diameters around the spots were measured. The assays were

performed twice in independent experiments on different days.

2.4.3 Collection of BBE cells

Bovine bronchial tissue was obtained from a healthy animal after processing at a

provincially inspected abattoir. Cells were isolated according to Wu and Smith (1982). Bronchial

tissue after dissection was placed in ice-cold DPBS with 100 IU penicillin-100 l µg ml-1

streptomycin and 1 µg ml-1 amphotericin B. The tissue was cut into smaller pieces, washed three

times with DPBS and then digested in 0.1% protease type XIV (w/v) at 4°C overnight. The

digestion was terminated by adding Dulbecco’s modified Eagle’s medium (DMEM)/Ham F12

medium supplemented with 10% fetal bovine serum (FBS). The tissue digest was filtered through

cell strainers (BD Falcon, Franklin Lakes, NJ, USA) with pore sizes of 100 µm and 40 µm, and

the filtrate was centrifuged at 200 × g for 10 min. The pellet was resuspended in 1 ml of

DMEM/Ham F12 with 10% FBS medium, layered onto 1.035 g ml-1 and 1.09 g ml-1 Percoll

gradients and centrifuged at 200 × g for 10 min. The cells at the interface were aspirated and

washed three times with DMEM/Ham F12 with 10% FBS, 100 IU penicillin-100 µg ml-1

-1 -l -1 streptomycin, 1 µg ml amphotericin B, 2 mmol l L-glutamine, 10 mmol l HEPES (Sigma-

Aldrich, Oakville,Ontario, Canada) and 55 mmol l-1 2-mercaptoethanol medium. One millilitre of cells was then added to a six-well tissue culture plate and incubated for 1 h at 37°C to eliminate adhering and contaminating fibroblasts or macrophages. Non adherent cells were resuspended and transferred to another six-well tissue culture plate and incubated at 37°C in 5% CO2. The medium

was changed daily until confluence. Afterwards, immunocytochemistry staining with monoclonal

mouse anti-human cytokeratin AE1/AE3 antibody (Abcam, Toronto, ON, Canada) was used to

confirm the epithelial origin of cells. Trypan blue exclusion test using an automated cell counter

(Countess, Invitrogen, Carlsbad, CA, USA) showed 96 to 98% viability. Fully confluent BBE

cells were frozen, passaged or seeded. 59

2.4.4 Assessment of bacterial adhesion to BBE cells

The BBE cells were seeded onto 12-well flat bottom tissue culture plates at 5 ×105 cells per ml per well and incubated until a complete monolayer was obtained. The percent bacterial adhesion on BBE cell monolayer was determined as described by Kalischuk and Inglis (2011).

The BBE cells were washed twice with DPBS and 1 ml of antibiotic-free DMEM/Ham F12 medium was added to each well, and the plates were incubated for 1 h before inoculation of bacteria. Overnight to 24 h cultures of the bacteria listed in Table 2. 1 were diluted to give bacterial concentrations of approx. 2 × 108 CFU per ml, as confirmed by plate counting. Monolayers were inoculated with sterile Lactobacilli MRS/tryptic soy/BHI broths (control) or bacteria to achieve a multiplicity of infection of 100 CFU per epithelial cell (MOI 100: 1), and incubated for 1 h at 37°C in 5% CO2. The monolayers were then washed three times with DPBS to release unbound bacteria.

Monolayers were lysed with 0.1% Triton X-100 in DPBS for 10 min at room temperature on an orbital shaker. The detached bacterial cells were aspirated, serially diluted with DPBS, and then plated onto appropriate agar media. The plates were incubated for 24-48 h at 37°C and colonies were counted (B1 CFU per ml). Bacterial cells initially added to each well of 12-well plates were also counted (B0 CFU per ml). The assay was performed twice in triplicate and the adhesion percentage was calculated as: % adhesion = (B1/B0) ×100.

2.4.5 Determination of antagonistic activity of probiotic strains against M. haemolytica S1

Antagonistic competition and displacement of M. haemolytica S1 by probiotic bacteria was evaluated using a method described previously (Guglielmetti et al., 2010). Monolayers of BBE cells on 12-well plates were washed twice with DPBS and incubated for 1 h with 1 ml of antibiotic- free DMEM/Ham F12 medium. In the competition assay, probiotic bacteria were added

simultaneously with M. haemolytica, and the plates were incubated for 1 h at 37°C. To test

displacement of M. haemolytica by probiotic strains, the incubation of BBE cell monolayers for 1

h at 37°C with M. haemolytica was followed by washing of unbound bacteria, addition of

individual probiotics strains, and another 1 h of incubation at 37°C. The BBE cells in both assays

were then washed, bacteria were detached and M. haemolytica were enumerated as described

above. The concentrations of bacteria were adjusted in all experiments to approx. 2 × 108 CFU per ml well. In each experiment, the following combinations were tested twice on different days using triplicate wells: (i) individual probiotic bacteria and M. haemolytica added simultaneously

(competition), (ii) individual probiotic bacteria added after M. haemolytica (displacement) and (iii) M. haemolytica alone (control).

2.4.6 Statistical analysis

Bacterial adhesion and antagonistic assays were analyzed as one-way analysis of variance

(ANOVA) using Proc Mixed Procedure of SAS. Tukey’s multiple comparison test was used to compare means. Significance was declared at P < 0.05. All statistical analyses were performed using SAS statistical software (SAS Institute Inc. Cary, NC).

2.5 Tables and Figures

Table 2.1 Bacterial strains used in the present study.

Species Strain Description Reference

Lactobacillus. Acidophilus (La) ATCC* 4356 Probiotic commercial strain Campana et al, 2012

L. casei subsp. Casei (Lc) ATCC 393 Probiotic commercial strain Saxami et al., 2012

L. helveticus (Lh) ATCC 15009 Probiotic commercial strain Giraffa et al., 2000

L. plantarum NCDO1193 (Lp) NCIMB† 8299 Probiotic from ensiled vegetable matter Zago et al., 2011

L. rhamnosus GG (Lr) ATCC 53103 Probiotic commercial strain Juntunen et al., 2001

Lactococcus lactis subsp. Lactis (Ll) DSM‡ 20250 Probiotic from Finnish Taete Schleifer et al., 1986

Mannheimia haemolytica (Mh) L024A A serotype 1 strain isolated from the Klima et al., 2014

respiratory tract of a feedlot steer that

succumbed to bovine respiratory disease

Paenibacillus polymyxa (Pp-1) JB-0501 Probiotic from livestock feed supplement Naghmouchi et al., 2013

P. polymyxa (Pp-2) ATCC 842 Species type strain Jeong et al., 2011

Streptococcus thermophilus (St) DSM 20617 Yogurt (Species type strain) Guglielmetti et al.,2010

*ATCC: American type culture collection, Rockville, MD; †NCIMB: National Collections of Industrial, Marine and Food

Bacteria, Aberdeen, Scotland; ‡DSM: Deutsche sammlung von mikroorganismen, Braunschweig, Germany. 62

Table 2.2 Growth inhibitory effects of probiotic bacteria against Mannheimia haemolytica S1.

Probiotic bacteria*

Lr La Lc Lh Lp Ll St Pp-1 Pp-2

Cell culture 19 ± 0.3 15 ± 0.6 16 ± 0.3 13 ± 0.5 17.7 ± 0.9 12 ± 0.0 0 15.3 ± 0.1 19 ± 0.4

Cell-free culture supernatant 14 ± 0.5 15.3 ± 0.1 14.3 ± 0.9 13 ± 0.0 16 ± 0.3 12 ± 0.4 0 16 ± 0.2 19 ± 0.5

*Probiotic bacteria are described in Table 1.The results are presented as mean diameter of the growth inhibition zone (mm) determined by the spot-on-lawn method (± standard deviation). The mean was obtained from two independent experiments

performed in triplicate, on separate days.

Table 2.3 Antagonistic effects of probiotic bacteria against Mannheimia haemolytica S1 adherence to bovine bronchial epithelial cells*.

Bacterial strains†

Lr La Lc Lh Lp Ll St Pp-1 Pp-2 Mh

‡Displacement 6.4 ± 0.16a 4.7 ± 0.23d 6.4 ± 0.21a 6.0 ± 0.13c 6.3 ± 0.23abc 6.3 ± 0.20bc 6.7 ± 0.02a 4.5 ± 0.29d 4.5 ± 0.16d 6.5 ± 0.07ab

§Competition 6.5 ± 0.13cb 4.9 ± 0.12e 6.2 ± 0.20cd 4.5 ± 0.07f 6.5 ± 0.11bc 6.4 ± 0.21bc 6.9 ± 0.10a 4.2 ± 0.16g 4.7 ± 0.13e 6.6 ± 0.15b

-1 *The results are reported as the mean (±SE) of M. haemolytica S1cell adherence (Log10 CFU ml ) to BBE cell monolayer obtained

from two independent experiments performed in triplicate. Mh, M. haemolytica S1 was added alone, as a negative control.

†Probiotics bacteria are described in Table 1. Different superscripted letters indicate difference between mean values in rows (P <

0.05).

‡Determined by inoculation of M. haemolytica prior to probiotic bacteria.

§Determined by inoculation of probiotic bacteria and M. haemolytica simultaneously.

Figure 2.1 Adhesion of probiotic bacteria and the bovine respiratory pathogen M. haemolytica

S1 to bovine bronchial epithelial cells.

The probiotic bacteria are listed in Table 2.1. The values are the means from two independent experiments conducted in triplicate. Different letters indicate mean values differ between different strains (P < 0.05). The vertical bars indicate standard error.

Chapter Three: Evaluation of the nasopharyngeal microbiota in beef cattle transported to

a feedlot, with a focus on lactic acid-producing bacteria

Chapter 3 has been published in Frontiers in Microbiology.

Amat S, Holman DB, Timsit E, Schwinghamer T, Alexander TW. 2019. Evaluation of the

nasopharyngeal microbiota in beef cattle transported to a feedlot, with a focus on lactic acid- producing bacteria. Front. Microbiol. 10:1988.

3.1 Introduction

Bovine respiratory disease (BRD) is the most common endemic disease in feedlot cattle,

accounting for approximately $4 billion in annual losses to the United States feedlot industry due

to the costs of treatment, prevention, and lost productivity (Cernicchiaro et al., 2013; Johnson et al., 2017). Although BRD is a multifactorial disease with a multitude of stressors that predispose cattle to viral and bacterial infections, bacterial pathogens are the principal agent in the pathogenesis of BRD. The major bacterial species involved in BRD are Mannheimia haemolytica,

Pasteurella multocida, Histophilus somni and Mycoplasma bovis. These are opportunistic

pathogens that are often present in the nasopharynx of healthy cattle as part of the commensal

microbiota (Griffin et al., 2010). However, when the host experiences compromised immunity

due to viral infection, or stressors such as maternal separation, transportation, and commingling at

auction markets and after feedlot placement, these opportunistic bacteria can proliferate in the

nasopharynx and translocate into the lung where they cause bacterial pneumonia (Rice et al.,

2007).

Bacterial pneumonia most often occurs within the first weeks of feedlot placement

(Babcock et al., 2010). Due to the segmented nature of beef production and large differences in

the operational scale of cow-calf producers and commercial feedlots in North America, beef calves

are frequently transported to a feedlot for finishing, either directly from a farm, or first through

delivery to an auction market. Some studies have shown a negative impact of shipment and

commingling at an auction market on the incidence of BRD (Step et al., 2008). The effects of these risk factors for BRD vary based on the season of the shipment (Hay et al., 2016) and distance of transportation from the farm to the feedlot (Cernicchiaro et al., 2012; Hay et al., 2014). Currently,

antibiotics are the most common management practice used to prevent (metaphylaxis) BRD in

feedlot cattle. Although usually effective, the large-scale use of antibiotics has come under greater

scrutiny due to an increase in antibiotic-resistant BRD pathogens (Timsit et al., 2017; Snyder et

al., 2017). Therefore, there is a need for the discovery and development of novel antibiotic

alternatives that are effective against BRD pathogens.

Culture-independent approaches have enhanced our understanding of the potential role of

NP microbiota in the respiratory health of cattle (Timsit et al., 2016a). The bovine nasopharynx

harbors a relatively rich and diverse microbial community which is dynamic and changes in response to various factors (Timsit et al., 2016b; Holman et al., 2015a and 2017). Previous studies

have suggested a possible association between the NP microbiota and development of BRD in feedlot cattle (Holman et al., 2015b; Zeineldin et al., 2017a). A disruption in the NP microbiota

may result in the loss of resistance to colonization by BRD pathogens or proliferation of existing

opportunistic pathogens in the nasopharynx. Thus, maintaining a stable microbial community in

nasopharynx of cattle at feedlot placement, and beyond, may decrease the risk of infection by

BRD-associated pathogens. Recent studies comparing the respiratory tract microbiota of healthy

and BRD-affected feedlot cattle, suggest that certain LAB present in the nasopharynx and the lung may be important for respiratory health (Holman et al., 2015a; Timsit et al., 2018). This was further confirmed by recent in vitro and in vivo studies that demonstrated the inhibitory effects of bovine NP-derived Lactobacillus spp. against M. haemolytica (Amat et al., 2016 and 2018). To date, however, LAB abundance within the NP microbiota of cattle has been poorly characterized.

More detailed information on these bacteria and the role they may have in respiratory health is important to better understand BRD and to propose alternatives to antibiotic metaphylaxis.

In the present study, we used 16S rRNA gene sequencing to characterize the NP microbiota

in beef cattle that were transported from a closed herd to a local auction market where they were

held for 48 h, and then transported to a feedlot. The prevalence of bovine respiratory pathogens,

including M. haemolytica, P. multocida and H. somni was also evaluated during the course of the

study using culture-dependent methods. Finally, antimicrobial activity of selected LAB isolates

from the nasopharynx of healthy feedlot cattle was tested against M. haemolytica in vitro. We

hypothesized that transport to and commingling with other cattle at the auction market would

increase the prevalence of BRD-associated bacteria and alter the NP microbiota.

3.2 Materials and Methods

3.2.1 Ethics statement

The calves used in this study were cared for in accordance to the guidelines set by the

Canadian Council on Animal Care (1993) and all experimental procedures involving cattle were

approved by the Animal Care Committee of the Lethbridge Research and Development Centre,

Agriculture and Agri-Food Canada.

3.2.2 Animal husbandry and experimental design

Thirteen Angus × Herford cross steers (initial body weight 325 ± 54 kg SD) were sourced

from a closed herd in Lethbridge, Alberta, Canada, as previously described (Holman et al., 2017).

All steers used in the present study were castrated and had no history of antibiotic or hormone

treatment, or vaccination. Calves were weaned 41 days prior to shipment to the auction market

and were bunk-fed an alfalfa-barley silage mixed diet during this period. On day 0 of the study,

calves were transported from the farm to a local commercial auction market using a cattle-hauling trailer (distance of 25 km). At the auction market, the calves were commingled with other cattle and held for 48 h before transport to the feedlot (distance of 8 km) according to regular 69

management operations by the auction market. Approximately, 4,000 head of cattle were sold

weekly at this auction market during the course of this study. At the feedlot, calves were held in

a pen separate from other cattle and fed a typical backgrounding diet which consisted of alfalfa

and barley. Animals were monitered daily for clinical signs indicative to BRD. These clinical signs include: lethargy, cough, nasal discharge and increased respiratory rate).

3.2.3 Nasopharyngeal swab sampling and isolation of BRD-associated pathogens

Deep NP swabs were collected from all calves on days 0 (at the closed herd farm prior to shipment), 2 (day of feedlot placement), 7, and 14 (i.e., 5 and 12 days after feedlot placement, respectively), as described previously (Holman et al., 2017). Swabs were kept on ice and transported to the lab where they were processed within one hour of collection. The procedures associated with swab processing for isolation and PCR confirmation of the respiratory pathogens

M. haemolytica, P. multocida and H. somni were identical to those described in our earlier study

(Holman et al., 2017). Briefly, for each animal at each time point, up to three isolates displaying typical morphology of M. haemolytica, P. multocida and H. somni were subcultured and tested by

PCR for species confirmation. An animal was considered positive for the respective bacteria if one of the isolates was PCR-positive, and negative if none of the isolates were PCR-positive. Upon plating of the swab suspension consisting of brain heart infusion (BHI) and glycerol (80:20), the suspension, including swabs, were cryopreserved at -80°C for DNA extraction as outlined in

Holman et al., (2017).

3.2.4 16S rRNA gene sequencing and analysis

The V4 region of the 16S rRNA gene was amplified and sequenced on a MiSeq instrument

(Illumina, San Diego, CA, USA) with the MiSeq Reagent Kit v2 as detailed in Holman et al.,

(2017). The 16S rRNA gene sequences were processed using DADA2 v. 1.10 (Callahan et al.,

2016) in R. 3.5.1. Briefly, the forward reads were truncated at 210 bp and the reverse reads at 200

bp. The reads were merged, chimeric sequences removed, and taxonomy assigned to each merged

sequence, referred to here as an operational taxonomic unit (OTUs) at 100% similarity, using the

naïve Bayesian RDP classifier (Wang et al., 2007) and the SILVA SSU database release 132

(Quast et al., 2012). The number of OTUs per sample and the Shannon diversity index was

calculated in R using Phyloseq 1.26.0 (McMurdie and Holmes, 2013) and vegan 2.5-3 (Oksanen

et al., 2013) was used to determine the Bray-Curtis dissimilarities and Jaccard distances. Samples

were randomly subsampled to 14,500 sequences prior to the calculation of Bray-Curtis

dissimilarities, Jaccard distances and the alpha-diversity measures. Sequences were submitted to

the NCBI sequence read archive under BioProject accession PRJNA296393.

3.2.5 In vitro growth inhibition of M. haemolytica by LAB isolates

To test inhibition of the BRD pathogen M. haemolytica, 8 LAB isolates from four genera

were evaluated. M. haemolytica was used as a model BRD pathogen due to its importance in the

development of this disease in feedlot cattle. The LAB were previously isolated from healthy

cattle (Timsit et al., 2016b) and isolates were identified as described by Holman et al., (2015b).

The M. haemolytica L024A strain was isolated from a BRD-affected steer originating from a

commercial feedlot in Alberta, Canada, and was identified as serotype 1 (Klima et al., 2014b).

Inhibition of M. haemolytica was evaluated using an agar slab method as per Dec et al. (2014) with some modifications. Briefly, 100 µL of 18 h culture of each LAB isolate was grown in Difco

Lactobacilli De Man, Rogosa and Sharpe (MRS) broth (BD, Mississauga, ON, Canada), inoculated

o onto Lactobacillus MRS agar and incubated at 37 C with 5% CO2 for 24 h. Agar slabs (10 mm in

diameter) were cut from the MRS agar after incubation using a sterile hollow punch (Tekton - 12

PC Hollow Punch Set – 6588) and placed on a lawn of M. haemolytica which was growing on tryptic soy agar (TSA) with 5% sheep blood (Dalynn Biologicals). Plates were incubation at 37oC with 5% CO2 for 24 h. The lawn of M. haemolytica was prepared by spread plating a 100 μL aliquot of M. haemolytica culture suspended in Dulbecco’s phosphate-buffered saline (pH 7.4) to obtain the target bacterial concentration of 1 × 108 CFU mL-1. After 24 h of incubation, the plates were checked for inhibition zones and the results recorded as the mean diameter (mm) of the inhibition zone for three independent experiments.

3.2.6 Statistical analysis

The number of OTUs and Shannon diversity index were analyzed by sampling day using a linear mixed model implemented in R with lme4 v 1.1.15 (Bates et al., 2014) with time as the fixed effect and animal as the random effect. Permutational multivariate analysis of variance

(PERMANOVA; adonis function; 10,000 permutations) of the Bray-Curtis dissimilarities was performed to assess the effect of sampling day on the NP microbial community structure. A pairwise adonis function (Arbizu et al., 2017) was used to determine which sampling days were most dissimilar from each other and p-values corrected for multiple comparisons using Holm's method. Differentially abundant OTUs between sampling times were identified using DESeq2 in

R (Love et al., 2014) and corrected for multiple comparisons using the false discovery rate. For this analysis samples were not randomly subsampled but OTUs found in less than 25% of the samples were removed. The relative abundance of phyla, the Lactobacillales order, families within the Lactobacillales order, and the 15 most relatively abundant genera were analyzed using generalized liner mixed model estimation procedure (PROC GLIMMIX) in SAS (ver. 9.4, SAS

Institute Inc. Cary, NC). The means were compared using LSMEANS statement and significance

was declared at P < 0·05. Lactic acid-producing bacteria are found in the Lactobacillales order,

encompassing the following six families: Aerococcaceae, Carnobacteriaceae, Enterococcaceae,

Lactobacillaceae, Leuconostocaceae, and Streptococcaceae (Mattarelli et al. 2014). . Spearman's

rank-based correlations between LAB families and the BRD-associated family of Pasteurellaceae

were calculated using the CORR procedure in SAS with the SPEARMAN option. The

Pasteurellaceae family was comprised of Mannheimia and Pasteurella genera. Possible

associations among the 15 most relatively abundant genera were predicted using a stepwise-

selected GLIMMIX model with beta-binomial distribution. The model used was: logit ( ) = ln

(ℼ/(1 - ℼ)) = b0 + btime + btime2 + b1 (X1) +….. + bn (Xn), where ℼ represents the relative abundance𝑌𝑌�

of a bacterial genus (0 to 1), Xn represents relative abundance (0 to 100%) of a bacterial genus n.

The stepwise selection method involved backward elimination and forward selection to eliminate

any variables that were insignificant (P > 0.05) to the predicted outcome.

3.3 Results

3.3.1 Isolation and detection of BRD-associated pathogens

All cattle remained healthy throughout the study and were not treated for any diseases. We

assessed the NP swabs for the presence of M. haemolytica, P. multocida, and H. somni via culturing (Table 3.1). Among these three BRD-associated pathogens, P. multocida was most frequently isolated, with 6, 2, 4, and 2 animals positive for P. multocida on days 0, 2, 7 and 14, respectively. None of the experimental animals were positive for P. multocida at any of sampling times. M. haemolytica was isolated from only a single animal on d 14 and H. somni was not detected among any of the cattle at any sampling time.

3.3.2 Effect of transport to the auction market and feedlot on the structure of the nasopharyngeal microbiota

In total, 3,064 unique archaeal and bacterial OTUs were identified among the NP samples from 1,707,932 16S rRNA gene sequences. There was a noticeable shift in the microbial community structure of the nasopharynx as the cattle were transported to the auction market and then to the feedlot. When OTU abundance (Bray-Curtis dissimilarity) was taken into account the effect was weak but significant (Figure 3.1A and 1B; R2 = 0.17; P < 0.05). Among the different

sampling times, d 0 and 14 NP samples were most dissimilar from each other (R2 = 0.18; P < 0.05).

Only d 0 and 2 and d 2 and 7 were not significantly different (R2 < 0.10; P > 0.05). We also

assessed the temporal changes in the NP microbial community structure over time using the binary

Jaccard distances (Figure 3.1C). This metric compares the microbial communities based on the

presence or absence of an OTU. Sampling time was observed to affect the community structure

(R2 = 0.15; P < 0.05), again most strongly between d 0 and 14 NP samples (R2 = 0.15; P < 0.05).

3.3.3 Longitudinal changes among the Lactobacillales families

We specifically analyzed the order Lactobacillales, which contains LAB. Overall, this

order constituted 5.18% of the 16S rRNA gene sequences in the NP microbiota. The mean relative

abundance of total LAB increased by 73% from the farm (d 0) to 5 days (d 7) after feedlot

placement (P < 0.05), although this increase was temporary as the relative abundance of LAB at d

14 was not different from the level observed on d 2 (Figure 3.2A). Within Lactobacillales, all six

families were detected from the steers during the course of the study (Figure 3.2B), including

Streptococcaceae (3.55%), Carnobacteriaceae (0.84%), Aerococcaceae (0.58%),

Lactobacillaceae (0.16%), Enterococcaceae (0.09%) and Leuconostocaceae (0.01%). There was

an increase in the relative abundance of most of these LAB families over the course of the study

(Figure 3.2C). For example, a significantly greater abundance of Streptococcaceae was observed

on d 7 compared to d 0 (P > 0.05) and Carnobacteriaceae was enriched on d 2 and d 14 compared to d 0 (P < 0.05), but not between d 7 and d 0. The relative abundance of Enterococcaceae was

lower on d 2 and 7 compared to d 0, but this family was enriched on d 14. Of note, due to higher

variation between animals, no statistical difference in abundance of Lactobacillaceae was

observed at any sampling times (P > 0.05). The relative abundance of Leuconostocaceae, the

least abundant LAB family, was unchanged from d 0 through d 14.

3.3.4 Longitudinal changes among the archaeal and bacterial phyla and genera in the nasopharyngeal microbiota

Across time and individual animals, a total of 26 different bacterial phyla were identified

with Proteobacteria (36.1%), Firmicutes (20.1%), Tenericutes (19.3%), Actinobacteria (12.7%),

and Bacteroidetes (8.6%) being the five most relatively abundant (data not shown). Overall,

Proteobacteria remained the most relatively abundant phylum throughout the study. Interestingly,

the relative abundance of most of these phyla varied significantly among individual animals on

almost all sampling days (Supplementary Figure S3.1). For example, the nasopharynx of two

animals was largely colonized with members of the Tenericutes (> 90%) on day 0.

Of the 15 most abundant genera, the relative abundance of Bacteroides, Moraxella,

Mycoplasma and Pasteurella did not differ by sampling time (Figure 3.3; P > 0.05). In contrast,

Acinetobacter, Corynebacterium and Planococcus were more relatively abundant on d 14

compared to all other sampling days (P < 0.05). Interestingly, Bifidobacterium was not detected

in any animals on day 0 but became progressively more abundant after feedlot placement and

through to d 14. The relative abundance of both Mannheimia and Streptococcus were greater on

d 2 and 7 compared to days 0 and 14 (P < 0.05). Psychrobacter spp. increased after feedlot entry and commingling at the auction market (d 2), although the relative abundance of this genus was variable during the experimental period. The relative abundance of the Rikenellaceae RC9 group was unchanged before and after feedlot placement but a lower abundance was detected on d 7 compared to d 14 (P < 0.05).

3.3.5 Associations between the 15 most relatively abundant genera

To identify potential associations among the 15 most abundant NP bacteria, we used a stepwise-selected GLIMMIX model. As shown in Table 3.2, varying degrees of positive or negative associations between these genera were detected. Moraxella, Corynebacterium,

Methanobrevibacter and Pasteurella were negatively associated with Mycoplasma (P < 0.05), with Methanobrevibacter spp. associated with the most negative regression coefficient (bME = -

0.676). Planococcus was found to be the only genus that was significantly associated with

Mannheimia. The odds ratio estimate revealed that a one unit increase in Planococcus could decrease the relative abundance of Mannheimia by 0.9 units (data not shown). The relative abundance of Pasteurella was predicted to be negatively affected by the presence of

Psychrobacter, but positively affected by Bifidobacterium.

Members of the Psychrobacter, Acinetobacter and Jeotgalicoccus genera were positively associated with Corynebacterium and Planococcus. The relative abundance of Streptococcus spp. were found to be negatively affected by Mycoplasma, Mannheimia and Acinetobacter spp. Among the 15 most relatively abundant genera, Bifidobacterium, Psychrobacter, Ruminococcaceae UCG-

005 and Planococcus were predicted to be the most interactive genera whose relative abundance

were significantly influenced by the presence of 4 or more different genera. A positive association

of sampling time with the relative abundance of these four genera was detected.

We also observed a significant interaction effect among multiple genera (Table 3.2). For

example, although both Psychrobacter and Acinetobacter had a positive association with

Corynebacterium, when analyzed individually, the interaction of these two genera (co-existence)

was negatively associated with Corynebacterium. The interactions between Mycoplasma and

Ruminococcaceae, or Planococcus and Methanobrevibacter displayed a negative association with

Psychrobacter, whereas, the interaction of Mycoplasma and Jeotgalicoccus were positively

associated with Psychrobacter.

3.3.6 Differentially abundant taxa in the nasopharyngeal microbiota following feedlot placement

In addition to the changes within the 15 most abundant genera discussed above, there was

more microbial richness (number of OTUs) in the NP microbiota at d 14 compared with d 0 as

well, although diversity (Shannon diversity index) was unaffected, due to the variability in the d 0

samples (Figure 3.4). Because the d 0 and 14 NP samples were most dissimilar based on both

Bray-Curtis dissimilarities and Jaccard distances, we identified which OTUs were most

differentially abundant between these sampling times. There was a total of 121 OTUs that were

differentially abundant between d 0 and d 14 with 105 of these OTUs enriched on d 14, having

log2 fold changes ranging from 1.6 to 10.3 (Supplementary Table S3.1). OTUs classified as

Acinetobacter, Bifidobacterium, Corynebacterium, Mannheimia, Planococcus, Prevotella, and

Psychrobacter were among the abundant OTUs that were enriched in the d 14 NP microbiota.

Notably, only 16 OTUs were more abundant in the d 0 NP samples. However, these did include

the abundant Mycoplasma, Moraxella, and Pasteurella OTUs. Differentially abundant OTUs with a log2 fold change of at least ± 6 are shown in Table 3.3.

We also compared the d 0 and 2 NP samples to identify OTUs in the NP microbiota that were altered after transport to and from the auction market, as there was a significant increase in microbial richness (number of OTUs) during this short period. In total, 25 OTUs were differentially abundant between d 0 and d 2, with 21 OTUs more abundant on d 2 and 4 on d 0

(Table 3.4). Bifidobacterium, Psychrobacter, and Streptococcus were among the most enriched genera on d 2 while Pasteurella and Sphingomonas OTUs were more abundant on d 0 prior to transport to the auction market.

3.3.7 The relationship between LAB and BRD-associated Pasteurellaceae family members

There were positive Spearman correlations between all LAB families, with the exception of an inverse association between Streptococcaceae and Leuconostocaceae (Table 3.5). Positive correlations were significant between Aerococcaceae, Carnobacteriaceae, Enterococcaceae and

Lactobacillaceae, or between Leuconostocaceae, Aerococcaceae and Carnobacteriaceae, and among Streptococcaceae and Carnobacteriaceae (P < 0.05). Although relative abundance of all

LAB families was inversely correlated with the Pasteurellaceae family, only the correlation between Enterococcaceae and Pasteurellaceae was significant (P = 0.05).

The dynamic changes in the relationship between LAB families and Pasteurellaceae over time are shown in Supplementary Figure S3.2. Enterococcaceae had a negative autocorrelation with Pasteurellaceae at all sampling points. The correlation between Lactobacillaceae and

Pasteurellaceae became increasingly negative overtime. Although a weak negative correlation

between Streptococcaceae and Pasteurellaceae was observed on d 0, the correlation shifted after

feedlot placement, becoming more positive over time. The remaining LAB families exhibited

inconsistent correlations with Pasteurellaceae at different sampling points, including after feedlot

placement.

Because of the negative associations observed for LAB families and Pasteurellaceae, we

tested whether members of LAB families, isolated from beef cattle, were capable of inhibiting the

BRD-associated pathogen M. haemolytica. All tested isolates within the genera of Lactobacillus,

Streptococcus and Enterococcus were able to inhibit the growth of M. haemolytica, with inhibition zones ranging from 14 to 21 mm (Figure 3.5). Among these inhibitory isolates, Lactobacillus strains exhibited greater inhibition of M. haemolytica (zones of inhibition ≥ 17 mm) while moderate inhibition of M. haemolytica was observed with Streptococcus and Enterococcus isolates. However, none of the Aerococcus isolates inhibited M. haemolytica.

3.4 Discussion

Increasing evidence suggests that the microbiota of the respiratory tract plays an important role in maintaining respiratory health of cattle. Additionally, it may also be a source of antimicrobials that inhibit respiratory pathogens in feedlot cattle (Timsit et al., 2016a; Amat et al.,

2016 and 2018). In North America, beef calves can be shipped to auction markets where they are held and mixed with other sources of cattle before transport to feedlots. While auction market- derived calves have been associated with increased risk of BRD, few studies have actually tested such an association at the microbiological level (Tylor et al., 2010). Previously, we investigated the effect of transporting cattle from a closed herd directly to the feedlot (Holman et al., 2017). In

the current study, we transported cattle from this same herd but included a 48 h stay at an auction

market before shipping to the feedlot.

3.4.1 The prevalence of cultured BRD-associated pathogens

Almost all calves remained negative for cultured M. haemolytica and H. somni during the

course of the study, and those that were positive for P. multocida on the farm were negative after

feedlot placement, despite the fact that they were commingeled at an auction market with other sources of cattle for 48 h prior to feedlot placement. Some studies have suggested that

commingling or mixing with different sources of cattle in the auction market increases the

possibility of BRD pathogen exposure and BRD in feedlot cattle. Ribble et al. (1995) evaluated

the correlation of commingling with the development of BRD in calves following feedlot

placement using a large number of cattle. They concluded that commingling increased the risk of

BRD in the calves following feedlot placement. Step et al., (2008) also observed significantly

greater total mortality due to BRD in auction market-derived calves compared to calves purchased

directly from a ranch. In our study, the lack of NP colonization by BRD-associated pathogens

during the course of mixing with other calves at the auction market indicated that spread of

pathogenic bacteria at the auction market was limited. This was also shown by Stroebel et al.

(2018) who observed no effect of commingling at an auction market for 24 h on the prevalence of

BRD-associated families Mycoplasmataceae and Pasteurellaceae in recently weaned beef heifers.

However, because bacterial colonization of other co-mingled calves at the auction market was not

evaluated, it is possible that spread of pathogens was limited due to these co-mingled calves not

being colonized by BRD pathogens themselves.

Step et al. (2008) also compared calves that were simultaneously weaned and shipped to a research feedlot to those that were weaned and held on a ranch for 45 d prior to being transported.

The total morbidity associated with BRD was significantly lower in calves weaned 45 d before transport to the feedlot than in calves weaned and immediately transported to the feedlot (35% vs.

6%). Although a relatively small number of cattle were followed in our study, all animals remained healthy and there was no proliferation of specific BRD-associated bacterial pathogens observed by either culturing or 16S rRNA gene sequencing. While this may indicate that commingling had a limited effect on bacterial transfer at the auction market, it is important to note that viral agents were not evaluated. In addition, the results may also be due to the weaning history and the origin of the herd the calves originated from, as well as the nature of their nasopharyngeal microbiota.

In the current study, the calves were sourced from a disease-free herd that was closed to cattle sourced outside the herd, limiting introduction of BRD-associated bacterial and viral pathogens prior to shipment. They were weaned 41 days prior to shipment to the auction market and therefore were expected to have enhanced immunity to cope with stress during transportation to and commingling at the auction market, and handling, and viral infection. Prior to shipping, these calves would have also experienced lower levels of stress, particularly stress associated with maternal-separation and eating from a bunk. It is therefore assumed that their risk of BRD was low and that they were likely less susceptible to colonization by BRD-associated pathogens.

These results are also in agreement with our earlier findings with cattle from the same farm that were directly transported to the feedlot. In that study, we were able to isolate only P. multocida from the nasopharynx and there was no increase in prevalence during the same 14 d period

(Holman et al., 2017). Overall, the data suggested that transfer of bacterial pathogens at the auction market was limited in this group of low-stress calves. 81

3.4.2 Effect of transport and auction market commingling on the nasopharyngeal microbiota

In agreement with our study, Proteobacteria, Firmicutes, Tenericutes, Actinobacteria, and

Bacteroidetes have been reported to be the most relatively abundant phyla in nasopharynx of early,

or newly weaned, and feedlot placed cattle (Holman et al., 2017; Timsit et al., 2016b and Stroebel

et al., 2018). Nevertheless, the proportion of some of these phyla observed in the present study

was different from that of some other studies. Newly weaned and auction or cow-calf ranch derived feedlot heifers were reported by Stroebel et al. (2018) harboured onaverage 3% Firmicutes

in their nasopharynx, compared to 20.1% in our study. The third most abundant phylum that we

observed was Tenericutes (19%), and this phylum has been reported to be the most abundant

phylum in the nasopharynx (Timsit et al., 2018 and Stroebel et al., 2018) and trachea (Timsit et

al., 2018) of feedlot cattle, potentially comprising more than 40% of the total microbial

community.

Overall, the structure and composition of the NP microbial community shifted after

transport to the auction market and then to the feedlot. In an earlier study, we transported beef

cattle heifers from the same farm directly to the same feedlot and noted a similar change in the

community membership following the first 2 weeks of feedlot placement (Holman et al., 2017).

The increase in microbial diversity of the NP microbiota after feedlot placement might be

due to several factors. It is possible that the feedlot was a source of bacteria that colonized the

respiratory tract through aspiration, or airborne nutrients that may have promoted growth of certain

bacterial species. In addition, the change in diet may have altered the gastrointestinal bacterial

microbiota, which may also colonize the respiratory tract through oropharyngeal transfer (Hall et

al., 2017). As mentioned above, the cattle used in our study were expected to be

immunocompetent, although it cannot be ruled out that some alteration in immunity occurred after

transportation, and that this potentially affected host modulation of the respiratory microbiome

(Hooper et al., 2012). Furthermore, potential networks of co-occurrence or co-exclusion associations predicted among the 15 most relatively abundant genera suggested that the NP microbial community membership before or after feedlot arrival, may be important factors in the susceptibility of the nasopharynx to colonization by novel bacterial species.

A large number of OTUs were increased in abundance as the cattle moved from the farm at d 0 through to d 14 when they had passed through the auction market and been in the feedlot for

12 d. Among these OTUs, the increase in Bifidobacterium was particularly striking given the complete absence of this genus in the d 0 samples. This suggests that at least one species in this genus colonized as a result of environmental exposure, contact with other cattle, or stress-related factors. While typically associated with the gastrointestinal tract of mammals, Bifidobacterium has repeteadly been identified in the respiratory tract of cattle (Holman et al., 2015a; Holman et al., 2017). Thus digesta of the gastrointestinal (GI) tract, or feces and manure, may be a source of this genus. In support of this, the relative abundance of Bifidobacterium spp. after feedlot entry

was associated with an increased relative abundance of Ruminococcaceae UCG-005, a taxon also

associated with the GI tract. Similar to many of the LAB, certain Bifidobacterium spp. are used as probiotic agents and have been investigated in humans to prevent or treat upper respiratory tract infections (Popova et al., 2012). Whether Bifidobacterium spp. in the upper respiratory tract of cattle confers a similar beneficial effect is unknown. However, given the strong positive correlation between the Bifidobacterium and Pasteurella genera in the present study, future

research should focus on assessing whether the presence of Bifidobacterium spp. increases colonization and proliferation of Pasteurella in nasopharynx of feedlot cattle.

Prevotella spp. are obligate anaerobes that have been identified from both nasopharynx and tracheal samples obtained from cattle, albeit at a lower relative abundance (Lima et al. 2016;

Timsit et al., 2018; Klima et al., 2019). The relative abundance of this genus in the nasopharynx has been reported to be similar among healthy or pneumonic dairy calves (Lima et al., 2016).

Tracheal samples from healthy feedlot cattle have also been noted to be enriched with Prevotella compared to those with BRD (Timsit et al., 2017). However, the potential role of this genus in the respiratory tract of cattle remains unknown. Prevotella is typically the most abundant genus in the rumen (Henderson et al., 2015), and Prevotella spp. respond to the acidity of the rumen, becoming more enriched when a grain-based diet is fed to the cattle (Petri et al., 2013). It is therefore plausible that the increased abundance of Prevotella in the present study may be due to the changes in the diet after feedlot placement.

Other abundant OTUs belonging to Acinetobacter, Psychrobacter and Corynebacterium genera also increased in relative abundance from d 0 to d 14. The positive associations observed among these three genera indicate that they coexist in the nasopharynx and may mutually benefit each other. These genera are often reported to be dominant in the nasopharynx of healthy feedlot cattle (Holman et al., 2017, Holman et. al, 2018; Stroebel et al., 2018; Zeineldin et al., 2017a) and so this microbial state may reflect one that is more mature and stable in feedlots. Planococcus spp. were also significantly enriched during the first 12 days of feedlot placement. This was the only genus that had a strong inverse association with Mannheimia. This bacterial genus has not been

well characterized but it warrants further research given the importance of Mannheimia spp. in the development of BRD.

Interestingly, we also observed a decrease in the abundance of OTUs assigned to the

Moraxella, Mycoplasma, and Pasteurella genera from d 0 to 14. One of the possible factors contributing to the reduction of Pasteurella spp. after feedlot placement might be the increased relative abundance of the commensal bacterial genus Psychrobacter, which displayed a strong negative association with Pasteurella. Based on the strong association of members of the

Mycoplasma and Moraxella, it is plausible to infer that reduced Mycoplasma abundance may have resulted in the decrease of Moraxella if the existence of Moraxella depends on Mycoplasma.

While several studies have reported an increase in nasopharyngeal colonization after feedlot placement (Stroebel et al., 2018; McMullen et al., 2018; Holman et al., 2019), our study only lasted

14 days, thus changes in the microbiota may have occurred subsequent to the last sampling. In contrast to Moraxella, the relative abundance of Corynebacterium and Methanobrevibacter was inversely associated with Mycoplasma. Therefore, the enrichment of these commensal genera, and perhaps also the increased diversity of the microbiota after feedlot placement, may have resulted in an undesirable nutrient landscape for M. bovis, resulting in a reduction in the relative abundance of Mycoplasma.

3.4.3 Relative abundance and antimicrobial activity of lactic acid-producing bacteria, and correlation with the BRD-associated Pasteurellaceae family

To the authors’ knowledge, this is the first study to report specifically on the longitudinal changes in the relative abundance of LAB bacteria, defined as the order of Lactobacillales, and their associations within LAB families and the BRD-associated Pasteurellaceae family in cattle

entering the feedlot. The relative abundance of the Lactobacillales order increased significantly

following the first week of feedlot placement and remained more relatively abundant on d 14

compared to d 0 at the farm. The families within the LAB order - including Aerococcaceae,

Carnobacteriaceae, and Streptococcaceae - were significantly enriched after feedlot placement.

The relative abundance of members of the Lactobacillaceae family remained stable during the course of study. Overall, positive correlations were observed among the LAB families, suggesting that cooperative relationships exist among different LAB species in the NP microbiota of feedlot cattle. Although these correlations were not statistically significant, negative Spearman’s correlation coefficients were observed for all LAB families with Pasteurellaceae. This potentially suggests that the presence of LAB in the nasopharynx may have a competitive exclusion effect on the BRD-associated Pasteurellaceae family and is supported by the inhibition assays performed.

A larger study with more animals would help determine if a true inverse relationship between LAB families and Pasteurellaceae exists in the nasopharynx of cattle.

The tested strains within the families of Lactobacillaceae, Streptococcaceae and

Enterococcaceae displayed in vitro antimicrobial activity against M. haemolytica. This finding,

along with the observed inverse associations of LAB families with Pasteurellaceae, suggests a role for the LAB in protecting against M. haemolytica colonization and proliferation in the respiratory tract. Holman et al., (2015a) observed that the relative abundance of the

Lactobacillaceae family in the nasopharynx of cattle entering the feedlot was significantly greater in animals that remained healthy compared with those that developed BRD. Furthermore, we recently identified several Lactobacillus strains isolated from the nasopharynx of healthy feedlot cattle that were able to inhibit the growth of M. haemolytica (Amat et al., 2016). The present study

suggests that bacteria other than those within the Lactobacillus genus are capable of inhibiting M. haemolytica, and offers insight into microbial competition within the upper respiratory tract of cattle. It also suggests that bacteria from the bovine respiratory tract may have potential for use as intranasal probiotics or bacterial therapeutics to modulate NP microbiota. Species within the

Lactobacillus genus have been used as probiotics for both humans and livestock (Fijan, 2014).

While certain strains of Enterococcus are pathogenic, this genus has also been used in livestock production as a direct fed microbial (EFSA, 2012). However, caution should also be applied when assessing LAB as a single group. For example, members of the Streptococcus genus are capable of degrading mucous which may adversely affect host innate immunity (Derrien et al., 2010). In addition, although Streptococcus displayed a negative correlation with the Pasteurellaceae family on d 0 at the farm, there was a positive association between the two taxa after feedlot placement.

Therefore, further research is warranted to investigate Streptococcus spp. in the respiratory tract of feedlot cattle.

In conclusion, the composition and structure of the NP microbiota underwent significant changes following commingling at the auction market and up to 12 d after feedlot placement. In particular, the NP microbiota at d 0 and at subsequent sampling times appeared to become more dissimilar. Many of these changes were driven by potential associations among several genera including increases in Acinetobacter, Bifidobacterium, Corynebacterium, Prevotella and

Psychrobacter, and decreases in Moraxella, Mycoplasma, and Pasteurella. Despite an increase in microbial richness and the relative abundance of certain genera, similar increases were not observed for BRD-associated pathogens by either sequencing or culturing. This suggests that healthy, low risk calves that have been weaned more than 40 days prior to marketing and feedlot

placement, are relatively resilient against these pathogens up to 12 days after feedlot placement.

The LAB families were also more relatively abundant in the first week of feedlot placement

compared to the farm. Most LAB families were negatively associated with BRD-associated

Pasteurellaceae family, which includes several BRD pathogens. Some LAB strains were also determined to inhibit the growth of M. haemolytica in vitro. Overall, the results of this study suggest that the NP microbiota of calves is dynamic with an increase in microbial richness following transport to an auction market and feedlot. In addition, we provided evidence of potential cooperation and exclusion taking place in the respiratory tract of cattle which may be useful for developing microbial-based strategies to mitigate BRD. Similar analysis of high-risk, immunocompromised cattle is also warranted.

3.5 Tables and Figures

Table 3.1 Calves (n = 13) positive for BRD-associate bacterial pathogens by culturing of

nasopharyngeal swabs.

Sampling time (day) H. somni P. multocida M. haemolytica

0 - 1a, 11, 15, 22, 27, 53 -

2 - 11, 15 -

7 - 11, 13, 23, 27 -

14 - 27, 44 34 a Numbers represent unique animal identifier.

Table 3.2 Associations among the 15 most relatively abundant genera in the nasopharyngeal microbiota of cattle (n =13) across time.

Interactions Ps*A Me*M My*R Je*M Pl*M Ac*R Time* Time*B Y b0 btime btime2 Mo My Co Bi Ma Ps Me Ru Ac Je Pl Ba Ri St Pa c a u y e i Je×Pl Pl a Ps*Je M 0.56 - o a - - - 0.562 - - - - 0.045 ------M - - - - y 1.63 - - 0.068 - 0.171 - - 0.676 ------0.045 ------Co 4.61 ------0.287 - 0.358 0.159 - - - - - 0.047 ------0.52 - - - 0.35 Bi 7.32 9 0.027 - - - - 0.040 - 0.095 6 - 0.232 - - - - 0.011 - 0.029 ------M - - a 1.28 ------2.255 ------0.06 0.21 - Ps 6.50 3 - - 0.003 0.157 - - - 0.322 2 0.053 0.284 ------0.009 0.015 -0.12 - - - - - M - 0.14 0.65 - - e 5.47 - - - - - 5 - - - - 0.356 0.437 0.132 - 9 ------0.197 0.238 - - - - 0.09 0.21 - Ru 5.91 6 - - - - 4 - - 0.213 - - - 0.767 - - 0.022 ------0.055 - - - Ac 5.30 - - - - 0.213 ------Je 6.02 - - - - 0.154 ------0.10 0.23 - Pl 6.31 9 ------0.106 - - 0.169 0.388 - 8 ------0.038 0.045 - 0.15 Ba 5.75 ------0.272 6 ------0.03 Ri 5.98 0 ------0.372 ------0.17 - - - St 2.98 2 - - 0.055 - - 0.038 - - - 0.882 ------0.90 - Pa 3.42 - - - - - 0 - 0.751 ------

a Associations were measured using the following stepwise-selected Generalized Liner Mixed Model: logit ( ) = ln (ℼ/1- ℼ) = b0 + btime + btime2 + b1 (X1) +….. + bn � (Xn). The values were the mean estimated regression coefficients. The negative and positive values represent𝑌𝑌 the negative association (mutual-exclusion) or

positive association (co-occupation), respectively, between the genera predicted and the independent variables (genus, time and interactions). Only those

independent variables that had significant effects (P < 0.05) on the predicted genus remained in the model and are presented in this table. Ac, Acinetobacter; Ba,

Bacteroides; Bi, Bifidobacterium; Co, Corynebacterium; Je, Jeotgalicoccus; Ma, Mannheimia; Me, Methanobrevibacter; Mo, Moraxella; My, Mycoplasma; Pa,

Pasteurella; Pl, Planococcus; Ps, Psychrobacter; Ri, Rikenellaceae_RC9_gut_group; Ru, Ruminococcaceae_UCG-005;; St, Streptococcus.

Table 3.3 Differentially abundant OTUs in the nasopharnyngeal microbiota of feedlot cattle between the d 0 and d 14 (n = 13).

Mean log2 fold OTU ID abundance change a, b FDRc Phylum Family Genus OTU66 181.4 -10.3 4.66E-57 Actinobacteria Bifidobacteriaceae Bifidobacterium OTU496 77.8 -9.1 1.98E-18 Bacteroidetes Prevotellaceae Prevotella OTU2836 68.8 -8.8 0.000281 Proteobacteria Pasteurellaceae Mannheimia OTU2280 59.5 -8.7 1.56E-37 Firmicutes Ruminococcaceae Ruminococcaceae UCG-013 OTU229 48.9 -8.4 1.01E-27 Actinobacteria Pseudonocardiaceae Saccharopolyspora OTU1569 45.1 -8.3 1.43E-20 Firmicutes Lachnospiraceae Lachnospiraceae NK4A136 group OTU2711 44.1 -8.3 8.15E-11 Proteobacteria Succinivibrionaceae Ruminobacter OTU3 39.5 -8.1 4.91E-06 Euryarchaeota Methanobacteriaceae Methanobrevibacter OTU1108 28.8 -7.6 2.14E-08 Firmicutes Aerococcaceae Facklamia OTU511 27.3 -7.6 6.27E-27 Bacteroidetes Prevotellaceae Prevotella OTU1628 27.0 -7.6 1.02E-25 Firmicutes Lachnospiraceae NA OTU1553 25.3 -7.5 0.000108 Firmicutes Lachnospiraceae Lachnospiraceae NK3A20 group OTU1457 25.0 -7.4 1.83E-27 Firmicutes Lachnospiraceae Acetitomaculum OTU1091 25.0 -7.4 1.99E-05 Firmicutes Staphylococcaceae Staphylococcus OTU1519 24.8 -7.4 8.64E-10 Firmicutes Lachnospiraceae Coprococcus OTU96 23.5 -7.4 1.39E-09 Actinobacteria Dietziaceae Dietzia OTU2871 22.7 -7.3 4.21E-12 Proteobacteria Moraxellaceae Psychrobacter OTU137 21.5 -7.2 2.63E-06 Actinobacteria Intrasporangiaceae Janibacter OTU1107 21.3 -7.2 0.00016 Firmicutes Aerococcaceae Facklamia OTU144 21.1 -7.2 1.68E-07 Actinobacteria Intrasporangiaceae Ornithinimicrobium OTU642 20.4 -7.2 7.59E-09 Bacteroidetes Rikenellaceae Rikenellaceae RC9 gut group OTU2450 19.6 -7.1 4.39E-11 Firmicutes Veillonellaceae Anaerovibrio OTU2384 18.8 -7.0 5.02E-11 Firmicutes Ruminococcaceae Ruminococcus OTU77 18.2 -7.0 1.90E-06 Actinobacteria Corynebacteriaceae Corynebacterium OTU1280 15.5 -6.8 8.27E-10 Firmicutes Clostridiaceae Clostridium sensu stricto OTU1052 15.3 -6.7 1.67E-05 Firmicutes Planococcaceae Lysinibacillus OTU2715 14.3 -6.6 2.77E-19 Proteobacteria Succinivibrionaceae Succinivibrio OTU644 14.1 -6.6 2.59E-05 Bacteroidetes Rikenellaceae Rikenellaceae RC9 gut group OTU1484 14.0 -6.6 5.07E-10 Firmicutes Lachnospiraceae Blautia OTU1629 13.5 -6.6 1.52E-12 Firmicutes Lachnospiraceae NA OTU225 13.5 -6.6 3.45E-07 Actinobacteria Pseudonocardiaceae Prauserella OTU1176 37.8 -6.5 2.91E-11 Firmicutes Streptococcaceae Streptococcus OTU1570 13.4 -6.5 0.00019 Firmicutes Lachnospiraceae Lachnospiraceae NK4A136 group OTU1143 12.9 -6.5 2.44E-06 Firmicutes Enterococcaceae Enterococcus OTU345 12.8 -6.5 2.04E-09 Bacteroidetes Barnesiellaceae NA OTU497 12.8 -6.5 5.74E-05 Bacteroidetes Prevotellaceae Prevotella OTU1469 12.6 -6.5 3.86E-05 Firmicutes Lachnospiraceae Agathobacter OTU646 12.0 -6.4 5.70E-06 Bacteroidetes Rikenellaceae Rikenellaceae RC9 gut group OTU1304 11.7 -6.4 5.69E-06 Firmicutes Clostridiaceae Proteiniclasticum OTU2939 11.5 -6.3 1.11E-05 Spirochaetes Spirochaetaceae Treponema OTU1137 11.4 -6.3 2.39E-07 Firmicutes Carnobacteriaceae Desemzia OTU830 11.0 -6.3 8.86E-06 Bacteroidetes Weeksellaceae Chishuiella OTU2451 10.8 -6.2 6.82E-06 Firmicutes Veillonellaceae Anaerovibrio OTU2938 10.8 -6.2 1.11E-06 Spirochaetes Spirochaetaceae Treponema

OTU224 10.5 -6.2 1.07E-10 Actinobacteria Pseudonocardiaceae NA OTU2868 380.6 -6.2 3.82E-12 Proteobacteria Moraxellaceae Psychrobacter OTU185 10.1 -6.1 0.000464 Actinobacteria Micrococcaceae NA OTU1473 9.9 -6.1 2.62E-05 Firmicutes Lachnospiraceae Anaerosporobacter OTU379 9.8 -6.1 7.48E-07 Bacteroidetes Muribaculaceae NA Clostridiales_vadinBB60_ OTU1312 9.3 -6.0 5.29E-08 Firmicutes group NA a Only those OTUs with a log2 fold change of at least +/- 6 are displayed. Mean abundance values are the mean abundance for each OTU among all d 0 and 14 NP samples. bNegative fold change values indicate OTUs that were enriched in the d 14 NP samples; cFDR = false discovery rate.

Table 3.4 Differentially abundant OTUs in the nasopharyngeal microbiota of feedlot cattl between the d 0 and d 2 (n = 13).

log2 (fold OTU ID Meana change) FDRc Phylum Family Genus b OTU66 52.2 -8.8 1.33E-22 Actinobacteria Bifidobacteriaceae Bifidobacterium OTU1107 29.4 -7.9 4.79E-10 Firmicutes Aerococcaceae Facklamia OTU67 23.2 -7.6 0.00014576 Actinobacteria Bifidobacteriaceae Bifidobacterium OTU1553 22.5 -7.5 0.000198397 Firmicutes Lachnospiraceae Lachnospiraceae NK3A20 group OTU1280 10.6 -6.4 0.000751528 Firmicutes Clostridiaceae Clostridium sensu stricto OTU246 10.2 -6.4 0.002568228 Actinobacteria Atopobiaceae Olsenella OTU1176 24.0 -6.2 5.08E-07 Firmicutes Streptococcaceae Streptococcus OTU1398 8.5 -6.1 0.020739752 Firmicutes Family_XIII Family XIII AD3011 group OTU1069 8.4 -6.1 0.020739752 Firmicutes Planococcaceae Sporosarcina OTU2938 7.3 -5.9 0.002568228 Spirochaetes Spirochaetaceae Treponema OTU1434 6.5 -5.7 0.02991382 Firmicutes Family_XIII NA OTU1126 5.4 -5.5 0.035196373 Firmicutes Carnobacteriaceae lloiococcus OTU2715 4.2 -5.1 0.032009852 Proteobacteria Succinivibrionaceae Succinivibrio OTU1840 18.8 -4.8 0.000909354 Firmicutes Peptostreptococcaceae Clostridioides OTU2868 121.4 -4.7 5.08E-07 Proteobacteria Moraxellaceae Psychrobacter OTU1129 26.6 -4.6 0.000372247 Firmicutes Carnobacteriaceae Atopostipes OTU2385 11.7 -4.5 0.028989764 Firmicutes Ruminococcaceae Ruminococcus OTU2029 5.7 -4.5 0.029881416 Firmicutes Ruminococcaceae Ruminococcaceae NK4A214 group OTU2869 35.3 -3.9 0.01341607 Proteobacteria Moraxellaceae Psychrobacter OTU1842 21.7 -3.1 0.02991382 Firmicutes Peptostreptococcaceae Paeniclostridium OTU1139 58.1 -3.0 0.006904558 Firmicutes Carnobacteriaceae Jeotgalibaca OTU1016 217.3 1.5 0.0154763 Firmicutes Bacillaceae NA OTU2673 138.3 1.8 0.009752217 Proteobacteria Sphingomonadaceae Sphingomonas OTU2911 147.4 1.9 0.002568228 Proteobacteria Rhodanobacteraceae NA OTU2839 611.9 4.0 0.009752217 Proteobacteria Pasteurellaceae Pasteurella aMean abundance values are the mean abundance for each OTU among all d 0 and 2 NP samples. bNegative fold change values indicate OTUs that were enriched in the d 2 NP samples whereas positive fold values show OTUs that were reduced in d 2 NP samples; cFDR = false discovery rate.

Table 3.5 Correlations between families within the order Lactobacillales, and the BRD-associated Pasteurellaceae family in the nasopharyngeal microbiota of cattle (n =13) across sampling times.

Time Aerococcaceae Carnobacteriaceae Enterococcaceae Lactobacillaceae Leuconostocaceae Streptococcaceae Pasteurellaceae 1.000 0.318a 0.476 0.214 0.101 -0.039 0.534 -0.097 Time 0.02b 0.0004 0.13 0.48 0.78 <.0001 0.49

0.318 1.000 0.792 0.366 0.479 0.334 0.043 -0.069 Aerococcaceae 0.022 <.0001 0.008 0.0003 0.01 0.76 0.62 0.476 0.792 1.000 0.531 0.508 0.278 0.274 -0.212 Carnobacteriaceae 0.0004 <.0001 <.0001 0.0001 0.04 0.05 0.13 0.214 0.366 0.531 1.000 0.366 0.200 0.019 -0.267 Enterococcaceae 0.12 0.008 <.0001 0.008 0.15 0.89 0.05 0.101 0.479 0.508 0.366 1.000 0.242 0.153 -0.084 Lactobacillaceae 0.47 0.0003 0.0001 0.008 0.08 0.27 0.55 -0.039 0.334 0.278 0.200 0.242 1.000 -0.073 -0.125 Leuconostocaceae 0.78 0.01 0.04 0.154 0.08 0.60 0.37 0.534 0.043 0.274 0.019 0.153 -0.073 1.000 -0.057 Streptococcaceae <.0001 0.76 0.05 0.89 0.27 0.60 0.68 -0.097 -0.069 -0.212 -0.267 -0.084 -0.125 -0.057 1.000 Pasteurellaceae 0.49 0.62 0.13 0.05 0.55 0.37 0.68 aSpearman’s correlation coefficient. Positive coefficient values indicate a positive correlation between two variable means; negative coefficient values indicate a negative correlation between the two variable means. bP value.

Figure 3.1 Beta-diversity of nasopharyngeal microbiota.

Principal coordinates analysis (PCoA) plot of the Bray-Curtis dissimilarities of the nasopharyngeal microbiota by sampling day for A) PC1 vs. PC2 and B) PC2 vs. PC3. C) PCoA plot of the Jaccard distances of the nasopharyngeal microbiota of cattle (n = 13) by sampling day. The percentage of variation explained by each principal coordinate is indicated on the axes.

Figure 3.2 Relative abundance of lactic acid-producing bacteria (LAB).

A) Relative abundance of total lactic acid-producing bacteria (LAB) (defined as order

Lactobacillales); B) Proportion of LAB familes within the order of Lactobacillales across time points; C) Box and whisker plots of the the percent relative abundance of 6 families observed within the oder of lactic acid bacteria (LAB, order Lactobacillale ) in the nasopharyngeal microbiota of cattle (n = 13) by sampling time. Coloured dots indicate outliers. The box in the box plots indicates the interquartile range (IQR) (middle 50% of the data), the middle line represents the median value, and the whiskers represents 1.5 times the IQR. Different lowercase letters indicate significantly different means (P < 0.05).

Figure 3.3 Box and whisker plots of the 15 most abundant genera in the nasopharyngeal microbiota of cattle (n = 13) by sampling time.

Coloured dots indicate outliers. The box in the box plots indicates the interquartile range (IQR)

(middle 50% of the data), the middle line represents the median value, and the whiskers represents

1.5 times the IQR. Different lowercase letters indicate significantly different means (P < 0.05).

Figure 3.4 Box and whisker plots of the A) number of OTUs and B) Shannon diversity index of the nasopharyngeal microbiota of cattle (n = 13) by sampling time.

The box in the box plots indicates the interquartile range (IQR) (middle 50% of the data), the middle line represents the median value, and the whiskers represents 1.5 times the IQR. Different lowercase letters indicate significantly different means (P < 0.05).

A) mean inhibitory zones from triplicate tests. B) An example of the agar slabs showing inhibitions zones on a lawn of M. haemolytica.

100

Chapter Four: Development of bacterial therapeutics against the bovine respiratory

pathogen Mannheimia haemolytica

Chapter 4 has been published in Applied and Environmental Microbiology.

Amat S, Timsit E, Baines D, Yanke J, Alexander TW. 2019. Development of Bacterial

Therapeutics Against the Bovine Respiratory Pathogen Mannheimia haemolytica. Applied and

Environmental Microbiology AEM.01359-19.

101

4.1 Introduction

Bovine respiratory disease (BRD), also known as shipping fever, remains the costliest

disease in the North American feedlot industry, despite advances in antimicrobials and vaccines

against respiratory pathogens (Taylor et al., 2010). Although BRD is a multifactorial disease with

several viral and bacterial agents involved, Mannheimia haemolytica is considered a major

pathogen in its etiology, and is therefore a primary target for both BRD mitigation and treatment in cattle (Griffin et al., 2010). As an opportunistic pathogen, M. haemolytica exists in the general

cattle population and colonizes the nasopharynx of healthy cattle. However, when cattle

experience compromised immunity due to stress and viral infection, M. haemolytica can proliferate

in the nasopharynx and then translocate into the lung where it can cause fibrinous pleuro-

pneumonia (Rice et al., 2007).

Calves arriving at feedlots are often more susceptible to respiratory bacterial infections due

to stress imposed by maternal separation, and environmental and management factors (Ives et al.,

2015). As a result, cattle determined to be at risk of BRD are frequently administered long-acting

antimicrobials upon feedlot entry (i.e. metaphylaxis) (Nickell and White, 2010). However,

antibiotic resistance has been reported to be increasing in BRD-associated pathogens (Portis et al.,

2012). In addition, recent feedlot studies conducted in both Canada (Timsit et al., 2017; Anholt et

al., 2017) and the United States (Snyder et al., 2017) revealed a high prevalence of multidrug-

resistant BRD bacterial pathogens displaying resistance towards antimicrobials used for

metaphylaxis. Emergence of antimicrobial-resistant bacteria associated with BRD presents a

significant risk to the beef industry, particularly if the efficacy of antimicrobials diminishes due to

pathogens acquiring resistance. Novel alternatives to metaphylactic antimicrobials are therefore

greatly needed. 102

Increasing evidence shows that bacterial communities residing within the respiratory tract are important to respiratory health and that disruption of the microbiota can reduce host resistance to colonization and proliferation of pathogenic bacteria (Man et al., 2017; Timsit et al., 2016).

The nasopharynx of cattle harbors a rich and diverse microbial community which is dynamic and has been shown to change in response to several management practices, including transportation to a feedlot (Holman et al., 2017), altering diet (Hall et al., 2017) and antimicrobial administration

(Holman et al., 2018). Recent studies have also suggested an association between the nasopharyngeal microbiota and development of BRD in feedlot cattle (Holman et al., 2015a;

Zeineldin et al., 2017a). This notion is further supported by studies associating a greater relative abundance of nasopharyngeal Lactobacillaceae at the time of feedlot entry with protection against

BRD (Holman et al., 2015a), and also specific inhibition of M. haemolytica in vitro (Amat et al.,

2017). Hence maintaining a stable microbial community in the nasopharynx of cattle after feedlot placement may offer protection against BRD development and bacteria colonizing the bovine respiratory tract may have potential for use as therapeutics to mitigate BRD pathogens. The objective of the present study was to develop bacterial therapeutics, with a focus on lactic acid- producing bacteria (LAB), originating from the respiratory tract of healthy cattle for mitigation of

M. haemolytica, using a step wise approach based on pathogen inhibition, cell adherence, and immunomodulatory properties.

4.2 Materials and Methods

4.2.1 Isolation of commensal bacteria from the nasopharynx of feedlot cattle

A schematic of the general study design, including bacterial isolation is shown in Figure 4.

1. Nasopharyngeal bacteria were isolated as part of previous studies (Holman et al., 2015b; Timsit et al., 2016b). Briefly, two groups of animals were used to increase diversity of bacterial isolates.

103

In a first group (Group 1), 70 crossbred recently weaned steers purchased from a local auction

market were sampled by deep nasopharyngeal swab (DNS) on day 0 (feedlot entry) and 60 days

after placement at the Lethbridge Research and Development Centre (LRDC) feedlot (Alberta,

Canada). Steers were fed diets typical of feedlots in Western Canada and remained healthy from day 0 to 60 (Holman et al., 2015b). In the second group (Group 2), 30 Angus beef steers were

sourced from a cow-calf ranch. These steers were sampled by DNS at weaning while still on the

cow-calf ranch, and then upon arrival to the LDRC feedlot and 40 d after arrival (Timsit et al.,

2016b). The DNS samples were processed for bacterial isolation using semi-selective media (De

Man, Rogosa, and Sharpe [MRS] or Rogosa plates) for LAB, as described by Holman et al.(2015b). Isolates were sub-cultured and stored in cryopreservative.

4.2.2 Identification of nasopharyngeal commensal bacterial isolates

A total of 300 banked isolates were randomly selected for inclusion in the present study.

For identification, the near full length 16S rRNA gene (> 1400 bp) was sequenced for each isolate and used for taxonomic identification as described previously (17). In instances where taxonomic identification at the species level was not possible by 16S rRNA gene sequence analysis, biochemical tests were also employed. For this, the isolates were sub-cultured on MRS

(Lactobacillus, Enterococcus) or tryptic soy agar (TSA; Staphylococcus) at 39oC. Colony

morphologies were observed after 24-48 h at both 27oC and 39oC. Anaerobic growth was tested

on MRS or TSA at 39oC in an anaerobic chamber with an atmosphere of 85% nitrogen, 10%

hydrogen and 5% CO2. Acid production from carbohydrates was determined with the API 50CHL

gallery (bioMerieux, Saint-Laurent, QC, Canada; Lactobacillus, Enterococcus) as per the

manufacturer’s instructions or using Difco purple agar base medium (BD Canada, Mississauga ,

104

ON) containing 1% carbohydrate (Staphylococcus). Confirmatory identifications were obtained

through comparison with published results.

4.2.3 Growth inhibitory effects of nasopharyngeal commensal bacteria against M. haemolytica

For all assays, M. haemolytica L024A was used. This isolate originated from a feedlot steer

that succumbed to BRD in Alberta, Canada and was confirmed as serotype 1 (Klima et al., 2014a).

Except for Moraxella and Pediococcus, isolates within all genera identified were included for

inhibition of M. haemolytica. A total of 178 commensal isolates were studied for inhibition of M.

haemolytica. Isolates were selected to include a diverse group of LAB and non-LAB bacteria from

the respiratory tract of cattle. Lactic acid-producing bacteria are defined as the order of

Lactobacillales which encompass six families: Aerococcaceae, Carnobacteriaceae,

Enterococcaceae, Lactobacillaceae, Leuconostocaceae, and Streptococcaceae (Mattarelli et al.,

2014). Growth inhibitory effects against M. haemolytica were performed using an agar slab method according to Dec et al. (2014) with some modifications. Briefly, 100 µL of an 18-h culture from each isolate grown in Difco Lactobacilli MRS broth (BD, Mississauga, ON, Canada) was

o spread as a lawn onto MRS plates and incubated at 37 C with 5% CO2 for 24 h. Agar slabs (10

mm in diameter) were cut from the 24-h incubated MRS plates using a sterile hollow punch

(Tekton - 12 PC Hollow Punch Set – 6588) and were placed with the culture side down onto a lawn of M. haemolytica on TSA plates containing 5% sheep blood. The lawn of M. haemolytica was prepared by spread-plating a 100-μL aliquot of M. haemolytica culture suspended in

Dulbecco’s phosphate-buffered saline (DPBS, pH 7.4) to obtain the target bacterial concentration of 1 × 108 CFU per mL. Up to four agar plugs were placed onto a single lawn of M. haemolytica,

including a control plug containing no bacteria. The agar plug/M. haemolytica lawns were then

105

o incubated at 37 C with 5% CO2 for 24 h. After incubation, the plates were checked for zones of

inhibition (ZOI). The ZOI were measured with a micrometer. The results were presented as the

mean diameter of the inhibition zone for three independent experiments.

4.2.4 Adherence of commensal bacteria to bovine turbinate cell monolayers

A subset of isolates (n = 47) displaying the greatest inhibition of M. haemolytica (ZOI ≥

15 mm) were evaluated for adhesion to bovine turbinate cells (ATCC-1390; American Type

Culture Collection, Manassas, VA, USA) using an assay described previously (Amat et al., 2017;

Kalischuk and Inglis, 2011) with some modifications. The isolates comprised Lactobacillus (n =

32), Bacillus (n = 2), Enterococcus (n = 3), Macrococcus (n = 1), Staphylococcus (n = 6) and

Streptococcus (n = 3) genera. The bovine turbinate cells were seeded onto 6-well flat bottom tissue culture plates at 1 × 105 cells per well and incubated in Dulbecco’s modiﬁed Eagle’s medium

(DMEM, Thermo Fischer Scientific, Oakville, ON, Canada) supplemented with 10% horse serum

(American Type Culture Collection) and 50 µg/ml gentamicin (Sigma-Aldrich, Oakville, Ontario

Canada) at 37 °C with 5% CO2 until a complete monolayer was obtained. The bovine turbinate

cell monolayer was washed twice with antibiotic-free DMEM medium. Then 2 mL of antibiotic- free DMEM was added to each well and the plates were incubated at 37 °C with 5% CO2 for 1 h

before inoculation with bacteria. Eighteen-hour cultures of bacterial isolates were diluted into

DMEM medium to give bacterial concentrations of approximately 1 × 109 CFU/mL. Then 200

µL of the bacterial suspension was pipetted into each well of cell monolayer to achieve 1 × 108

CFU bacterial cells per 105 bovine turbinate cells per well. The plates were incubated for 3 h at 37

°C with 5% CO2. After 3 h incubation, unbound bacterial cells were removed by washing four

times with DMEM medium. The monolayers were then lysed with 0.1% Triton X-100 in DPBS

for 30 min at room temperature on an orbital shaker. The detached bacterial cells were aspirated 106

and serially diluted with DPBS, and then plated onto Lactobacillus MRS agar medium. The plates were incubated for 24 to 48 h at 37 °C with 5% CO2 and colonies were counted (CFU per mL).

The assay was performed three times in independent experiments on different days.

4.2.5 Antagonistic competition activity of commensals against M. haemolytica on bovine turbinate cells

A total of 15 isolates from 3 different genera (Lactobacillus, n = 12; Enterococcus, n = 1;

5 and Staphylococcus, n = 2) that had adherence values of ≥ 5 log10 CFU/10 BT cells were

evaluated in competition assays against M. haemolytica. Antagonistic competition of M.

haemolytica was performed using a method described previously (Amat et al., 2017) with some

modifications. Monolayers of bovine turbinate cells (1 × 105 cells per well) on 6-well plates were

washed twice with DMEM medium, and then incubated with 2 mL of antibiotic-free DMEM at 37

°C with 5% CO2 for 1 h before inoculation of bacteria. M. haemolytica and probiotic bacteria suspended in DMEM medium to achieve individual concentrations of 1 × 108 CFU/mL/ well were

simultaneously added to the well, and the plates were incubated for 1 h at 37°C with 5% CO2. For

control wells, only M. haemolytica was added. At the end of the experiment, cell monolayers were

washed and lysed as described above, and the adherent M. haemolyitca were enumerated by plating

onto blood agar plates supplemented with 15 µg/mL bacitracin to inhibit Gram positive bacteria.

The reduction in M. haemolytica adhered to bovine turbinate cells was calculated as: enumerated

M. haemolytica in control wells (B0 CFU) – enumerated M. haemolytica from wells co-inoculated

with commensal bacteria (B1 CFU) / B0 * 100 %. The competition assay was replicated six times

with a minimum of four different culture days for the cell line.

4.2.6 Fluorescent microscopy of bacteria adhering to bovine turbinate cells

Bovine turbinate cells were seeded onto two-well chamber slides (Nunc® Lab-Tek® II

107

Chamber Slide System, Sigma) and incubated using the growth conditions described above until

confluent. The bovine turbinate cell monolayers were washed three times with DMEM (one mL

each time) and were then stained by adding one mL of antibiotic-free DMEM (pre-warmed) with and 5 µL DAPI (4',6-Diamidino-2-Phenylindole, Dihydrochloride) (Molecular Probes, Inc.

Eugene, OR). The slides were gently rotated to mix the DAPI stain in each chamber, followed by

0 incubation at 37 C, 5% CO2 for 90 min. After incubation, the cell monolayers were washed three

times with DMEM to remove unbound DAPI stain. One mL of DMEM was then added into each

well and incubated for 30 min before addition of labelled bacteria. For bacteria, 500-µL aliquots

of cultures grown for 18 h were centrifuged at 8000 × g for 5 min, and the pellets were re-

suspended with PBS prior to fluorescent labeling reactions. Mannheimia haemolytica was labelled

with Alexa Fluor ® 488 Microscale Protein, and Lactobacillus strains were labelled with Alexa

Fluor ® 594 microscale protein (Molecular Probes, Inc. Eugene, OR) according to the manufacture’s instruction. Lactobacillus strains were either added to the stained BT cells alone, to visualize adherence, or in combination with M. haemolytica to visualize competition. Labelled bacteria cells were suspended in DMEM to achieve concentrations of 1× 108 CFU/mL/well per

bacterium prior to addition to the well of DAPI-stained BT cell monolayers, and incubated for 1

h. Subsequently, cell monolayers were washed four times with 1 mL DMEM to remove all

unbound bacterial cells. The cell monolayers were fixed with 2% paraformaldehyde solution

(diluted in PBS) and were examined using an Olympus Fluoview FV1000 laser confocal scanning

microscope.

4.2.7 Evaluation of antibiotic susceptibility of selected isolates

A total of 15 isolates that were evaluated in the competition assays were analyzed for their

antimicrobial susceptibility. Minimum inhibitory concentrations (MIC) of 20 antibiotics were 108

determined by the microdilution method (Sensitre, Thermofisher Scientific, Nepean, ON, Canada)

using a commercially available panel (YSTP6F, TREK diagnostic systems, Cleveland, OH, USA).

The antimicrobial susceptibility testing was performed according to the procedures recommended

for the YSTP6F panel with an exception that Lactobacillus MRS broth was used for Lactobacillus

strains that did not grow well in cation-adjusted Muller Hilton broth with lysed horse blood.

Antimicrobials and the range of concentrations tested are listed in Table 4.2. Isolates were inoculated into plates using a Sensititre AIM delivery system (Sensitre, Thermofisher Scientific) and after incubation, plates were evaluated with a Vision imager (Sensitre, Thermofisher

Scientific). Reference strain Lactobacillus plantarum NCDO1193 served as quality control.

For the Lactobacillus spp. strains, the interpretation of the MIC values of clindamycin

(resistant ≥ 2 µg/mL), daptomycin (susceptible ≤ 4 µg/mL), erythromycin (resistant ≥ 8 µg/mL), linezolid (susceptible ≤ 4 µg/mL), meropenem (resistant ≥ 4 µg/mL) penicillin (susceptible ≤ 8

µg/mL) and vancomycin (resistant ≥ 16 µg/mL) were based on the interpretive criteria provided by CLSI M45 (CLSI, 2016). The L. paracasei (> 4 µg/mL) and L. plantarum (> 32 µg/mL) strains were defined as resistant to tetracycline according to the breakpoints provided by the European

Food Safety Authority (EFSA, 2012). The breakpoints provided by the EFSA (2012) were also used to define L. paracasei (> 4 µg/mL) and L. plantarum (> 8 µg/mL) strains as resistant to chloramphenicol. For L. buchneri strains, resistance to chloramphenicol (> 4 µg/mL) and erythromycin (> 1 µg/mL) were determined according to the FEEDAP Panel (European

Commission 2008), while breakpoint resistance to tetracycline (128 µg/mL) was determined according to Feichtinger et al.(2016).

For E. faecium, resistance breakpoints for tetracycline (> 4 µg/mL), chloramphenicol (≥ 16

µg/mL), vancomycin (> 4 µg/mL), and erythromycin ( > 4 µg/mL) were determined according to 109

EFSA (EFSA, 2012). The breakpoints provided by CLSI VET01S (CLSI, 2015) were used to define Staphylococcus strains (6E and 28C) as resistant to penicillin (≥ 2 µg/mL, horse), amoxicillin-clavulanate (≥ 1/0.5, cat), vancomycin (≥ 16 µg/mL, human), clindamycin (≥ 4 µg/mL, dog), erythromycin (≥ 8 µg/mL, human), chloramphenicol (≥ 32 µg/mL, human) and tetracycline

(≥ 1 µg/mL, dog). At the time of the experiment, there were no breakpoints or interpretative criteria provided by CLSI or other literature to the antibiotics azithromycin, cefepime, cefotaxime, ceftriaxone, cefuroxime, ertapenem, levofloxacin, moxifloxacin, tigecycline and trimethoprim/sulfamethoxazole.

4.2.8 Effects of Lactobacillus spp. isolates on the expression of genes associated with adaptive and innate immune response in BT cells

A total of 10 selected commensal isolates from Lactobacillus spp. were evaluated for their effects on the expression of genes associated with adaptive and innate immune response in bovine turbinate cell monolayers. These isolates were selected based on their ability to compete against

M. haemolytica for adherence to bovine turbinate cells and antimicrobial susceptibility phenotypes. The bovine turbinate cells were seeded onto 12-well flat bottom tissue culture plates at 1 × 104 cells per well and incubated using standard culturing conditions described above.

Bovine turbinate cell-monolayers were washed twice with antibiotic-free DMEM medium, and then incubated with 1 mL antibiotic-free DMEM medium at 37 °C with 5% CO2 for 1 h before inoculation with bacteria. M. haemolytica or commensal bacteria were suspended individually in

DMEM medium and added to BT cells to achieve a concentration of 1 × 107 CFU per well, then the plates were incubated for 6 h at 37 °C with 5% CO2. Controls included bovine turbinate cells without addition of bacteria. At the end of the experiment, cell monolayers were washed four times with DMEM and then lysed with 350 µL of RLT buffer (RNeasy Mini Kit, Qiagen, Valencia,

110

CA) for 10 min at room temperature on an orbital shaker. The cell lysates were aspirated and

immediately stored at -80 °C for further analysis. The immune stimulation assay was performed

four times in independent experiments on different days.

Total RNA from the bovine turbinate cell lysates was extracted using an RNeasy Mini Kit

(Qiagen) according to the manufacture’s instruction. The integrity of the extracted total RNA was evaluated with an Agilent Bioanalyzer using a RNA 6000 Nano LabChip (Agilent Technologies,

Waldbronn, Germany). Total RNA from each sample (0.5 µg each, RIN > 7) was reverse

transcribed into cDNA using the RT2 first strand kit (Qiagen). The cow innate and adaptive immunity responses RT2 profiler PCR array (Qiagen) with 84 test genes related to host response

to bacterial infection and sepsis was used to evaluate the effects of bacteria on gene expression.

Real-time PCR was performed using RT2 SYBR Green Mastermix (Qiagen) and a CFX Connect real-time system (Bio-Rad, Hercules, CA). Real-time PCR conditions were performed according to the PCR array manufacturer’s manual (Qiagen). Data normalization was performed with the housekeeping genes Glyceraldehyde-3-Phosphate Dehydrogenase (GAPDH) and Actin Beta

(ACTB) and fold change in gene expression was calculated using the 2−ΔΔCT method (Livak and

Schmittgen, 2001).

4.2.9 M. haemolytica inhibitory mechanisms of 6 Lactobacillus isolates, as candidate bacterial therapeutics

Based on the rankings of selection criteria (Figure 4.1), 6 Lactobacillus strains comprised

from four different species [L. amylovorus, L. buchneri (n = 2), L. curvatus and L. paracasei (n =

2)] were selected as the best therapeutic candidate strains. The potential inhibitory mechanisms

by which these selected strains inhibit M. haemolytica were investigated by testing their ability to

111

produce lactic acid and H2O2, and by screening their genomes for bacteriocin-encoding genes. In

addition, their effects on the cell morphology of M. haemolytica were evaluated.

Determination of lactic acid production and antimicrobial activity of lactic acid against M.

haemolytica

The selected Lactobacillus strains were inoculated individually into 5 mL MRS broth at an

OD610 of 0.05 and incubated aerobically at 37 °C with shaking at 200 rpm for 24 h. In addition,

all strains were combined in equal concentrations (OD610 = 0.05) and grown similarly to individual

strains. MRS broth without any Lactobacillus cells was used as a control. After incubation,

supernatants were collected from 1.5 mL of culture by centrifugation (10,000 × g, 5 min) followed

by filtering through 0.22 µm syringe filter to remove bacterial cells. Metaphosophoric acid (25%,

v/w) was added into the cell culture supernatant at a ration 1:5 and was mixed and immediately

stored at -20 °C till analysis. The concentrations of lactic acid (DL-lactic acid) in the cell-free supernatants were measured using a 5890A gas liquid chromatograph (Phenomenex, Torrance,

505 CA, USA) as described by Wang et al. (2006).

To determine whether the observed lactic acid range produced by these selected

Lactobacillus strains could inhibit the growth of M. haemolytica, M. haemolytica was grown in

BHI media supplemented with different concentrations of lactic acid. A 100 µL of overnight M. haemolytica culture was inoculated into 2.5 mL BHI media containing 0, 9.38, 18.75, 37.5, 75,

100, and 150 mM of lactic acid (DL-Lactic acid lithium salt, Sigma-Aldrich Canada, Oakville,

ON), and incubated at 37°C, 200 rev/min, for 24 h. After 0, 8 and 24 h of incubation, the cultures were serially diluted with DPBS (pH 7.4) and plated onto TSA blood agar plates. The plates were incubated at 37°C for 24h to determine CFU.

Determination of H2O2 production 112

Aliquots of cell-free culture supernatants of the 6 selected strains prepared for lactic acid production were subjected to H2O2 measurement. A hydrogen peroxide assay kit

(Colorimetric/Fluerometric) (ab102500) (Abcam Inc., Toronto, ON) was used according to

manufacture’s instructions.

Whole genome sequencing and screening of bacteriocin-encoding genes

Whole genome sequencing was performed on the selected strains. The details on genomic

DNA extraction, whole genome sequencing, and analysis were described previously (Amat et al.,

2019c). A web-based tool, BAGEL 4 (van Heel et al., 2018) was used to search for the genes encoding bacteriocin from the genomes of the Lactobacillus strains as described by Flórez and

Mayo (2018).

Scanning electron microscopy

The effect of selected Lactobacillus strains on the cell morphology of M. haemolytica was evaluated using scanning electron microscopy (SEM). A single colony of each Lactobacillus

strain was inoculated into 5 mL BHI and incubated for 24h in BHI (37 °C, 200 rev/min). For a

negative control, BHI containing no cells were used. After incubation, the cell-culture supernatants were obtained as described above. A 100 uL of overnight M. haemolytica culture (1-

2 × 108 CFU per mL) was centrifuged at 10,000 × g for 5 min, and the cell pellets were suspended with 1 mL cell-culture supernatants obtained from the Lactobacillus strains and incubated for 10 h (37 °C, 200 rev/min) before cell harvesting. The treated M. haemolytica cells were centrifuged at 10,000 × g for 5 min, and the cell pellets were fixed with 4% glutaraldehyde. Further sample processing and SEM imaging procedures were described previously (Amat et al., 2019d).

113

4.2.10 Statistical analysis

Competition assay data were analysed as a one-way analysis of variance (ANOVA) using

Proc Glimmix in SAS (ver. 9.4, SAS Institute Inc. Cary, NC). The LSMEANS statement was used to compare the group means. The difference in gene expression between BT cells (control) that were not co-cultured with bacteria cells and the BT cells co-cultured with bacteria cells were assessed by Student’s t-test for each gene using the Rt2 Profiler PCR Array Analysis software,

Version 3.5TM (Qiagen). The level of statistical significance was set at < 0.05.

4.3 Results

4.3.1 Isolation and identification of nasopharyngeal commensal bacterial isolates

A total of 300 isolates from MRS and Rogosa agar plates were isolated and identified using near full-length 16S rRNA gene sequences (Table 4.1). These isolates were from 14 different genera, with Bacillus (34%), Staphylococcus (30%), Streptococcus (12.3%), and Lactobacillus

(12.0%) being more predominant genera. Although both MRS and Rogosa agar plates are semi- selective for LAB, 69% of the total bacteria isolated were non-LAB species. A total of 93 isolates were taxonomically classified as LAB, and comprised the genera Streptococcus (39.8% of the total

LAB), Lactobacillus (38.7%), Enterococcus (10.8%), Aerococcus (9.7%), and Pediococcus

(1.1%).

4.3.2 Growth inhibitory effects against M. haemolytica

Of the identified bacteria, 178 isolates comprising 12 different genera were tested for their ability to inhibit M. haemolytica growth using an agar slab method (Figure 4.2). A total of 74 isolates were LAB within the genera Aerococcus (n = 4), Enterococcus (n = 5), Lactobacillus (n

= 33) and Streptococcus (n = 32) (Figure 4.2A). Of these, 88% inhibited the growth of M.

114

haemolytica, with inhibition zones ranging from 11 to 23 mm (Figure 4.2B). Approximately 48% of the tested LAB displayed relatively strong inhibition of M. haemolytica (ZOI: 17-23 mm).

Specifically, Lactobacillus displayed the greatest inhibition against M. haemolytica, with 88% of

the tested Lactobacillus isolates showing ZOI ranging between 17 and 23 mm (Figure 4. 2A and

B). Although 91% of tested Streptococcus isolates inhibited M. haemolytica, 59 % of them had

relatively weaker inhibition (ZOI: 11.1-14 mm) and 35% showed medium inhibition (ZOI: 14.1-

16.9 mm). Four of 5 tested Enterococcus isolates moderately inhibited M. haemolytica. However,

none of Aerococcus isolates inhibited M. haemolytica (Figure 4.2A).

The non-LAB isolates (n =104) tested for inhibition of M. haemolytica taxonomically

belonged to 8 different genera and were constituted mainly of Bacillus (51% of total non-LAB

isolates tested) and Staphylococcus (41%) (Figure 4.2C). Of these non-LAB isolates, 46% of

them displayed growth inhibitory effects against M. haemolytica (Figure 4.2C). Among the

inhibition-positive isolates, 48 % had relatively weak inhibition (ZOI: 11.1-14 mm), and 31%

showed medium inhibition (ZOI: 14.1-16.9 mm) (Figure 4.2D). Of Bacillus isolates tested, 55%

did not inhibit growth of M. haemolytica. Only 6% of tested isolates showed relatively strong

inhibition (ZOI >17 mm). The growth of M. haemolytica was inhibited by 49% of the tested

Staphylococcus isolates, with 76% of these inhibition-positive isolates showing weaker to medium

inhibition (ZOI: 11.1-16.9 mm). Corynebacterium and Macrococcus isolates displayed relatively

strong inhibition (ZOI: 17-19.9 mm). Moderate inhibition of M. haemolytica was also observed

with one E. coli isolate. However, isolates within the genera of Acetobacter, Micrococcus and

Rummeliibacillus did not inhibit M. haemolytica.

115

4.3.3 Adherence of selected isolates to bovine turbinate cells

A total of 47 isolates were selected for adherence to BT cells, based on their ability to inhibit M. haemolytica (inhibition zone ≥ 15 mm). These isolates were from 6 different genera

(Bacillus, Lactobacillus, Macrococcus, Enterococcus, Staphylococcus and Streptococcus). All tested isolates were able to colonize bovine turbinate cell monolayers, with mean adherence

5 ranging between 3.4 and 8.0 Log10 CFU per 10 bovine turbinate cells (Figure 4.3). Of these, 32 isolates were Lactobacillus spp. and were from 7 different species including L. amylovorus (n =

1), L. berivis (n =2), L. buchneri (n = 23), L. paracasei (n = 3) and L. hilgardi (n = 1), L. curvatus

(n = 1) and L. sunkii (n = 1) (Figure 4.3A). Among these Lactobacillus isolates, 47% displayed

5 mean adherences greater than 5.0 Log10 CFU per 10 bovine turbinate cells. Interestingly, the adherence to BT cells differed for strains within the same species, which was more obvious within the L. buchneri species (Figure 4.3A). Of the 15 non-Lactobacillus isolates, the E. faecium strains, two Staphylococcus strains (28C and 98C), and one Streptococcus strain showed mean adherences

5 greater than 5.0 Log10 CFU per 10 bovine turbinate cells (Figure 4.3B).

4.3.4 Antagonistic competition activity of selected isolates against M. haemolytica

Fifteen isolates were tested for competitive exclusion ability. The isolates were selected

based on high inhibition of M. haemolytica (ZOI > 15 mm) and strong adherence to bovine

5 turbinate cells (≥ 5.0 Log10 CFU per 10 BT cells). Reduction of M. haemolytica adherence to

bovine turbinate cells was observed with all tested isolates (Figure 4.4). The mean reduction of

M. haemolytica adherence to bovine turbinate cell monolayers ranged from 32 to 78%.

Lactobacillus amylovorus (72B) displayed the strongest inhibition of M. haemolytica adherence to BT cells, with adherence greater than 9 other tested strains (P < 0.05) In contrast, L. buchneri

116

(65A) showed the weakest antagonistic competition against M. haemolytica. The remaining

strains had relatively comparable levels of M. haemolytica adherence inhibition (P > 0.05).

4.3.5 Antimicrobial susceptibility of selected isolates

The MIC of 20 antibiotics were determined against 15 isolates. According to the

interpretive criteria provided by CLSI (M45) (CLSI, 2016), all tested bacteria were susceptible to

clindamycin, erythromycin, linezolid, meropenem and penicillin (Table 4.2). The majority of

Lactobacillus isolates were susceptible to daptomycin. Only L. amylovorus 72B grew at the

maximum antibiotic plate concentration tested (2 µg/mL). Given the resistance breakpoint for

daptomycin against Lactobacillus is defined as greater than 4 µg/mL, it was not possible to define

whether L. amylovorus 72B was daptomycin-resistant. The L. amylovorus strains were susceptible to vancomycin while the MIC values for all other Lactobacillus strains were greater than the maximum concentration tested (> 4 µg/mL) and could not be defined. Only two of the tested

Lactobacillus strains [L. buchneri (38C) and L.amylovorus (72B)] were not inhibited by levofloxacin at the maximum concentration tested (4 µg/mL). All other tested strains of L. buchneri were not inhibited by tetracycline at the maximum concentration of 8 µg/mL, which is

16-fold lower than the resistant cut off value for L. buchneri (128 µg/mL), thus limiting their evaluation of tetracycline resistance. The L. paracasei strains were susceptible to chloramphenicol.

Enterococcus faecium 64C was susceptible to tetracycline, chloramphenicol, vancomycin and erythromycin. According to the CLSI VET01S guidelines, S. chromogenes (28C) was resistant to penicillin while the S. epidermidis (6E) was susceptible to penicillin. Both of these

Staphylococcus strains were also resistant to amoxicillin-clavulanate and tetracycline, but were susceptible to vancomycin, clindamycin, erythromycin and chloramphenicol. 117

4.3.6 Stimulation of innate and adaptive immune responses in bovine turbinate cell monolayers

Among the 84 genes tested, 33 genes showed significant difference in transcriptional gene

expression between the bovine turbinate cells co-cultured with at least one bacterium tested and

the bovine turbinate cells incubated under the same conditions without bacteria (P < 0.05) (Table

4.3). Of the Lactobacillus isolates, L. curvatus (strain 103C), L. amylovorus (72B), and L.

buchneri (67A, 63A, 65G and 63B) upregulated CXCL8 gene with 3.98-fold to 8.43-fold

difference in expression compared to control (P ≤ 0.02). The transcription of IL-6 was also

upregulated by 5 of these 6 Lactobacillus isolates (fold changes ranged from 8.63 to 22.75, P ≤

0.04). The expression of NFKB1 was upregulated (P ≤ 0.03) in bovine turbinate cells co-cultured

with L. curvatus (103C isolate) and L. buchneri isolates (63A, 63B, 65G and 67A) relative to

control (fold changes ranged from 1.5 to 2.13). L. amylovorus (72B), L. buchneri (63B and 65G)

and L. curvatus (103C) induced overexpression of interferon-alpha/beta receptor alpha chain with

1.6-2.0-fold changes in gene expression compared to control (P ≤ 0.02). Of the 84 genes, very few were down regulated in bovine turbinate cells in response to Lactobacillus inoculation.

Transcription of chemokine receptor 6 was downregulated by L. paracasei (57A) and L. buchneri

(63B and 38C) isolates (P ≤ 0.04), and L. paracasei (57A) down regulated the gene encoding solute carrier family 11 member 1 by 1.94 fold (P = 0.03).

In contrast to Lactobacillus spp., significantly high immune stimulation in BT cells was observed with M. haemolytica, which upregulated 28 genes with fold changes ranging from 1.43 to 1,321 (P < 0.05). The greatest responses to M. haemolytica were the transcription of C-X-C motif chemokine 10 (1321 fold change), CXCL5 (782 fold change), myxovirus resistance 1 (330 fold change), and IL-6 (290 fold change). The genes including IRF7, DDX58, CXCL8, CD40,

118

IL15, NFKB1A, ICAM1 and LOC512672 exhibited relatively high overexpression in bovine

turbinate cells incubated with M. haemolytica (10-67 fold changes) (P < 0.05). Moderate

upregulation by M. haemolytica was observed for the genes encoding chemokine ligand 2, colony stimulating factor 2, janus kinase 2, major histocompatibility complex (class 1, A like), NFKB1,

STAT1 and TLR4 (5-10 fold changes).

4.3.7 M. haemolytica inhibitory mechanisms of 6 Lactobacillus isolates, as candidate bacterial therapeutics

A total of 6 Lactobacillus strains (listed in Table 4.4) from four different species were selected as the best candidates for the development of intra-nasal bacterial therapeutics to mitigate

M. haemolytica based on the selection criteria of inhibition (strongest inhibition), adhesion

(strongest adhesion), competition exclusion (strongest exclusion), antimicrobial susceptibility

(limited resistance), and immunomodulation (moderate stimulation of immune genes). To understand the potential mechanisms through which these selected Lactobacillus strains confer direct inhibition against M. haemolytica, we determined their lactic acid, H2O2 and putative bacteriocin production capacity, and their effect on the cell morphology of M. haemolytica.

Lactic acid production and growth inhibition of lactic acid against M. haemolytica

We determined the lactic acid production of 6 Lactobacillus strains individually, as well as when combined (Table 4.4). The concentrations of lactic acid detected from the supernatants ranged between 80 to 142 mM. The supernatants of the 6 strains combined, and L. paracasei strains (3E and 57A) contained the most lactic acid (132-142 mM), followed by L. curvatus 103C

(111 mM) and L. amylovorus 72B (103 mM). The supernatants of L. buchenri strains contained the least lactic acid, with 80 and 94 mM, respectively. No lactic acid was detected from the negative control (MRS broth).

119

The inhibitory effects of lactic acid covering the range of concentrations produced by the

6 strains was tested against growth of M. haemolytica (Figure 4. 5). At 8 h post-inoculation, M.

haemolytica growth was inhibited when the lactic acid concentrations ranged from 18.75-150 mM,

compared to 0 and 9.38 mM concentrations. Although there was cell growth in media

supplemented with 18. 75 mM lactic acid within the first 8 h incubation, this concentration of lactic

acid resulted in lower cell counts compared to 0 and 9.38 mM concentrations. After 24 h, no viable

cells were detected in media containing 100 and 150 mM lactic acid.

H2O2 production

The concentrations of H2O2 in cell-free culture supernatants obtained from Lactobacillus strains after 24 h of incubation varied among different Lactobacillus species, with L. amylovorus

being the most predominant H2O2 –producer (Table 4.4). L. buchneri (63A and 86D) and L.

curvatus (103C) strains produced similar amounts of H2O2.However, no H2O2 was detected from

the supernatants of L. paracasei strains (3E and 57A).

Encoded bacteriocins

Sequences obtained for L. paracasei (57A) were too poor to evaluate and thus were eliminated from bacteriocin evaluation. The genomes of 5 selected strains revealed that L. amylovorus 72B and L. paracasei 3E contained bacteriocin-encoding genes while L. buchneri

(63A and 86D) and L. curvatus (103C) strains did not have genes encoding bacteriocins (Table

4.4). Lactobacillus amylovorus (72B) had 5 genes encoding Enterolysin A and Helveticin J, and both which belong to the bacteriolysin class. The genome of L. paracasei (3E) contained four different bacteriocin genes including LSEI_2163, Enterolysin A, Carnocin-CP52 immunity protein, and Enterocin Xβ.

The effect of selected strains on cell morphology of M. haemolytica 120

The morphological effects of supernatants from candidate strains on M. haemolytica were

examined using SEM. Noticeable changes in the cell structure, including shrinkage of the cell

surface, irregular-rod shape, and holes in the cell envelop were observed when M. haemolytica

was incubated in culture supernatants from Lactobacillus strains, compared to non-treated cells

(Figure 4.6). Supernatant from L. amylovorus (72B) reduced cell density of M. haemolytica to the greatest extent (data not shown) and caused the most apparent destructive changes in M. haemolytica cell structure (Figure 4.6B). L. buchneri (63A) demonstrated minor cell damage compared to other strains tested (Figure 4.6C), followed by L. paracasei (3E) (Figure 4.6D). L. paracasei (57A), L. curvatus (103C) and L. buchneri (86D) exhibited similar degrees of cell

damage (Figure 4.6E-6G).

4.4 Discussion

Bacteria originating from the host target site are more likely adapted for successful re-

colonization (Shewale et al., 2014). To develop bacterial therapeutics for mitigating the BRD

pathogen M. haemolytica, we therefore performed our own screening of commensal bacteria

isolated from the upper respiratory tract of healthy feedlot cattle. This is also the site where BRD

bacterial pathogens colonize and proliferate before translocating into the lung to induce lung

infection. In previous studies comparing healthy and BRD-affected cattle phenotypes, an

increased abundance of LAB was observed in healthy animals (Holman et al., 2015a; Timsit et al.,

2018). Thus, bacterial species taxonomically belonging to LAB (order of Lactobacillales) were

originally targeted using selective media. A wide range of commensal bacterial species were

isolated and identified from two different sources of healthy feedlot cattle. Of these, 31% isolates

were LAB strains. A subset of isolates were selected fro evaluation, comprising mainly LAB 121

strains and non-LAB species such as Bacillus and Staphylococcus, which have a known history of

probiotic use (Elshaghabee et al., 2017; Borah et al., 2016).

In total, a diverse number of bacteria were able to inhibit the growth of M. haemolytica to

varying extents. This suggests that the environment within the bovine respiratory tract is highly

competitive, with multiple bacteria capable of producing factors that inhibit the opportunistic pathogen M. haemolytica. In support of this, Corbeil et al., (1985) also described several genera

from the nasal cavity of cattle that could inhibit bovine respiratory pathogens including M. haemolytica, Pasteurella multocida and Histophilus somni. In the present study, the bacteria

displaying strongest inhibition of M. haemolytica were within the genus Lactobacillus. Previously,

Lactobacillus was identified as being more abundant in the lungs of healthy feedlot calves, compared to those diagnosed with BRD (Timsit et al., 2018). Lactobacillus spp. therefore appear

to be involved in maintaining cattle respiratory health, and may achieve this by inhibiting

respiratory pathogens through the production of antimicrobial factors. Direct inhibition of pathogens is an important attribute of bacterial therapeutics and probiotics, and occurs through production of lactic acid, bacteriocins and H2O2 or through the mechanical property of auto-

aggregation (Pridmore et al., 2008; Popova et al., 2012). While few studies have investigated

bacterial therapeutics for mitigating BRD bacterial pathogens, the inhibitory effects of commensal

bacteria against the growth of a wide range of bacterial pathogens involved in human intestinal

and respiratory tract infections have been documented (Verna and Lucak , 2010; Benhsen et al.,

2013). In order to develop bacterial therapeutics with the greatest potential to mitigate M.

haemolytica only a subset of screened commensals with the strongest inhibition were further

evaluated, of which the majority were Lactobacillus spp.

122

We used bovine turbinate cells for adhesion assays because this cell type is found within the upper respiratory tract, where M. haemolytica colonizes (Frank and Briggs, 1992), and where

inhibition of its growth would be desired in order to limit proliferation and subsequent

translocation and infection of the lungs. All strains tested were capable of adhering to bovine

turbinate cells, to varying extents, which was expected given that they were originally isolated

from the nasopharynx of cattle. Surprisingly, E. faecium isolates had the strongest adherence.

Enterococcus colonizes the lower gastrointestinal tract though they are capable of adhering to

extra-intestinal epithelial cells (Sharon et al., 2017). While likely transient inhabitants of the

respiratory tract, this probably explains why Enterococcus has been consistently observed in the

upper respiratory tract of cattle (Van Tassell and Miller, 2001; Holman et al., 2015b; Timsit et al.,

2016b). Several Lactobacillus strains also displayed a high level of adherence. Although no

studies have investigated the adhesion of commensal respiratory bacteria to bovine upper

respiratory tract cells, we previously did show that commercial Lactobacillus species displayed

greater adhesion to bovine bronchial epithelial cells compared to Streptococcus and Paenibacillus

strains (Amat et al., 2017). Lactobacillus spp. have the ability to colonize both mucus (Van Tassel and Miller, 2011) and underlying epithelial cells of the respiratory tract (Saroj et al., 2016). The high level of adherence observed with some of the Lactobacillus strains to bovine turbinate cells might be attributed to both specific (surface-dependent proteins and surface-layer proteins) and non-specific (cell surface hydrophobicity and lipoteichoic acid) adhesion mechanisms (Lebeer et al., 2008). Different adherence capacities among Lactobacillus species (L. berivis vs. amylovorus), and even strains of the same species (L. buchneri), were observed in the present study. It is likely that variation in strain adherence resulted from differences in cell surface structure of the isolates, as has been described previously (Duary et al., 2011). Compared to Lactobacillus spp., the strains 123

within the genera of Bacillus, Macrococcus and Streptococcus showed weaker adhesion to BT cells, and were therefore excluded from further screening.

A total of 15 isolates, with the strongest inhibition of M. haemolytica and adhesion to bovine turbinate cells, were further evaluated for their ability to competitively inhibit M. haemolytica colonization of bovine turbinate cells. All tested strains were able to inhibit the adherence of M. haemolytica to BT cell monolayers. This likely occurred through a combination of direct inhibition and also competition for binding sites on the epithelial cells (Lee et al., 2000;

Bermudez-Brito et al., 2012). Of note, when evaluating inhibition, adhesion and competition results, no single strain was ranked as the top candidate across all of those criteria. For example,

L. amylovorus 72B showed the strongest inhibition and competition against M. haemolytica, but the adherence of this strain to bovine turbinate cells was less than L. buchneri and L. paracasei strains. In contrast, L. buchneri (63A) and E. faecium (64C) showed moderate growth inhibition of M. haemolytica but strong adherence to bovine turbinate cells and antagonistic competition against M. haemolytica, in comparison to other strains. Thus, utilizing several criteria is important in defining potentially effective bacterial therapeutics. In addition, designing therapeutics strains with varying antimicrobial properties may promote a broader efficacy of bacterial therapeutics against pathogens. In support of this, multi-strain cocktails containing probiotics with different mechanisms of action have shown better anti-pathogenic activity, and modulation of mouse gut microbiome and short chain fatty acid production (Nagpal, 2018), as well as greater attenuation of pathogen-induced inflammatory response in intestinal epithelium (MacPherson et al., 2017) compared to single-strain probiotics.

124

Safety concerns exist over probiotic or therapeutic bacteria that are resistant to antibiotics,

especially if resistance elements are encoded on mobile elements (FAO, 2016; Gueimonde et al.,

2013). The isolates used in our study were from cattle that were not administered antibiotics before

or during the time of isolation, thus pressure was low for resistance to develop or for resistant

bacteria to be selected. Indeed, when evaluated, the selected Lactobacillus strains were generally

susceptible to antibiotics with defined break points, though some antibiotics could not be fully

tested due to their breakpoint concentrations not being reached in the antibiotic panel. Although

the majority of L. buchneri strains were not inhibited by the highest concentration of tetracycline

tested (8 µg/mL), tetracycline resistance (128 µg/mL) in L. buchneri has been reported to be

intrinsic (Feichtinger, 2016). A L. buchneri strain (NRRL B-50733) that showed a tetracycline

MIC value of 32 µg/mL has been considered as safe as a silage additive for livestock (EFSA,

2017). Thus higher MIC for tetracycline in L. buchneri strains would not likely limit these bacteria being used as therapeutics, although further evaluation of the L. buchneri isolates in our study with an antibiotic panel containing higher concentrations of tetracycline would be beneficial. The L. amylovorus strain 72B had higher MIC towards levofloxacin and moxifloxacin. Currently, there are no breakpoints available for L. amylovorus strains against these two fluoroquinolone antibiotics. However, some species within the Lactobacillus genus are known to be intrinsically resistant to fluoroquinolones including levofloxacin (Hummel et al., 2007; Karapetkov et al.,

2011).

The Staphylococcus strains were resistant to either penicillin (28C) or amoxicillin- clavulanate and tetracycline (28C and 6E). Therefore, these strains were excluded from further evaluation. Despite being susceptible to the antibiotics tested, and showing strong inhibition of

M. haemolytica, the E. faecium strain (64C) was also not considered for the immunostimulation 125

assay due to E. faecium being a potential opportunistic pathogen (Gao et al., 2018) and defined

as level 2 pathogen in some countries. Considering the data from the inhibition, competition and

antimicrobial susceptibility assays, in addition to a history of safe use in both food and feed

industry (Bernardeau et al., 2006), the strains of Lactobacillus were, therefore selected for

immunomodulation properties.

Modulation of host innate and adaptive immunity by therapeutic bacteria can potentially

increase resistance to pathogen infection (Yan et al., 2011). Bovine respiratory epithelial cells are

an initial point of contact between the host and respiratory microbiota (Ackermann et al., 2010)

and were therefore used to evaluate immunomodulation. In the present study, the tested

Lactobacillus strains induced moderate gene expression of the chemokine CXCL8 in bovine

turbinate cells, showing the ability of these nasopharyngeal strains to stimulate an immune

response. In contrast, M. haemolytica induced strong over-expression of CXCL8 in bovine

turbinate cells. The protein CXCL8 is a potent chemoattractant and plays an important role in

inflammation and wound healing through activation of neutrophils and other immune cells (Jundi

and Greene, 2015). Whether CXCL8 has a beneficial role in protecting the host through the

inflammation healing process or is detrimental by promoting pathogenesis, likely depends on its

level of expression (Jundi and Greene, 2015). In support of this, Gartner et al. (2015) reported that commensal Lactobacillus spp. present in the bovine uterus upregulated CXCL8 gene expression by 2-6 fold in endothelial epithelial cells after 6-8 h co-culturing. These authors suggested that excessive expression of CXCL8 might contribute to the development of uterine disease, while a moderate stimulation of CXCL8 may be necessary for bacterial clearance.

Similar to CXCL8, expression of NFKB1 and the pro-inflammatory cytokine IL-6 in bovine turbinate cells was stimulated by most of the Lactobacillus strains tested. However, 126

upregulation of these genes in bovine turbinate cells by Lactobacillus spp. was considered as moderate, when compared to M. haemolytica. Lipopolysaccharide from M. haemolytica has been shown to induce secretion of IL-6 in pulmonary epithelial cells (Guillot, 2004) and activation of the NF-KB pathway causes excessive inflammation in lower respiratory tract, thus enhancing infection by this pathogen (Zecchinon et al., 2005). While we did not co-inoculate Lactobacillus strains with M. haemolytica for the immunomodulation assays using bovine turbinate cells, expression of IL-6 in bovine endothelial epithelial cells was moderately upregulated in response to commensal Lactobacillus spp. present in the bovine uterus (Gao et al., 2018). Future studies evaluating co-inoculation would provide insight into whether the virulence of M. haemolytica could be attenuated by bacterial therapeutics.

Overall, most of the 84 genes encoding the TLR pathway, cytokines and chemokines receptors, inflammation response, NF-KB signalling, apoptosis, and innate immune and defense response to bacteria were not influenced by the Lactobacillus strains tested. However, M. haemolytica induced overexpression of 28 genes including cytokine and chemokine receptors, inflammatory responses, NF-KB signalling and defence response to bacteria. None of the tested

Lactobacillus strains caused an excessive immune response as observed with M. haemolytica. The effects of Lactobacillus on immune modulation in bovine turbinate cells was species- and strain- specific. Similarly, immunomodulation has been shown to vary at the strain level of commensal bacteria (Hill et al., 2014; O’Toole et al., 2017). Combined, the results on immune stimulation in bovine turbinate cells suggest that Lactobacillus may have a role in modulating immunity in cattle, however, future studies are warranted to elucidate the mechanisms by which modulation occurs and its impact on respiratory pathogens.

127

All 6 of the candidate therapeutic strains produced lactic acid to varying extents. This acid

has shown to alter the membrane structure of pathogens and is a mechanism by which probiotic

Lactobacillus spp. inhibit pathogen growth (Neal-McKinney et al., 2012). In agreement of our

study, Neal-McKinney et al. (2012) reported a similar range of lactic acid production by

Lactobacillus spp. (L. acidophilus, crispatus, gallinarum and helveticus) as well as concentrations

varying according to strains tested. For our study, given that each therapeutic candidate, or a

combination of all 6, produced concentrations of lactic acid greater than the minimum inhibitory

concentration (37.5 mM), lactic acid is likely a common metabolite by which these candidates

directly inhibited M. haemolytica.

In contrast to lactic acid, production of H2O2 and encoded bacteriocins varied among the

candidate strains. Hydrogen peroxide produced from Lactobacillus has previously been shown to

have bactericidal activity against gut pathogens (Pridmore et al., 2008), and may be a mechanism

of inhibition of M. haemolytica for four of the candidate strains in our study. Only two of the

candidate therapeutics encoded bacteriocins, which can have bacteriocidal or bacteriostatic activity

(Parada et al., 2007; Mokoena, 2017). Similar to our study, it has been observed that bacteriocins

encoded by Lactobacillus is species-, strain-, and origin-dependent (Collins et al., 2017). The

different class of bacteriocins detected from L. paracasei (3E) were similar to those observed with

the probiotic strain L. paracasei SD1 (Surachat et al., 2017). Expression of these bacteriocins may have led to L. paracasei (3E) being one of the strongest inhibitors of M. haemoltytica (ZOI, > 22 mm). For L. amylovorus (72B), this strain was predicted to produce bacteriolysins including

Enterolysin A and Helveticin J. Although L. amylovorus have been reported to produce

bacteriocin Amylovorin L471 (Collins et al., 2017; Callewaert et al., 1999), these genes were not

encoded by 72B. It is interesting to note that while L. amylovorus (72B) did not have the greatest 128

adhesion to BT cells, it did have one of the greatest inhibition and antagonistic values. The 72B strain also caused the greatest morphological damage to M. haemolytica. Thus a combination of lactic acid, H2O2, and bacteriocin production may have resulted in the strong inhibition of M. haemolytica observed for 72B.

In summary, we isolated and identified commensal bacteria from 14 different genera residing in the nasopharynx of healthy feedlot cattle as part of normal flora. Of these commensal isolates, using a stepwise approach, we screened isolates comprised of 12 different genera for their ability to inhibit the growth of respiratory pathogen M. haemolytica, to adhere to bovine turbinate cells and to compete against M. haemolytica adherence to the bovine turbinate cells. We, then, evaluated the best candidates for antimicrobial resistance and their immunomodulation effects in bovine turbinate cells. Based on the data generated, 6 Lactobacillus strains from four different species (L. amylovorus strain 72B, L. buchneri strains 63A and 86D, L. curvatus strain

103C, and L. paracasei strains 3E and 57A) were selected as the best candidates for the development of intra-nasal bacterial therapeutics to mitigate M. haemolytica in cattle. Their selection was based on high inhibition of M. haemolytica directly and through competition, high adherence to bovine turbinate cells, lack of antibiotic resistance, and moderate immunomodulation in bovine turbinate cells. The potential mechanisms by which these selected 6 strains inhibited M. haemolytica were also investigated. Lactic acid production was common amongst strains but production of H2O2 and encoded bacteriocins varied. Currently, in vivo studies are being conducted to evaluate the effects of intra-nasal administration of these strains on the microbiota of feedlot cattle.

129

4.5 Tables and Figures

Table 4.1 List of bacteria identified from the nasopharynx of healthy feedlot cattle and those selected for initial inhibition of M. haemolytica using agar slabsa.

No. of isolates screened for Genus No. of isolates % of total isolates inhibition of M. haemolytica b LAB (n = 93) Aerococcus 9 3.0% 4 Enterococcus 10 3.3% 5 Lactobacillus 36 12.0% 33 Pediococcus 1 0.3% 0 Streptococcus 37 12.3% 32 Non-LAB (n = 207)

Acetobacter 5 1.7% 2 Bacillus 102 34.0% 53 Corynebacterium 3 1.0% 1 Escherichia coli 3 1.0% 2 Macrococcus 1 0.3% 1 Micrococcus 1 0.3% 1 Moraxella 1 0.3% 0 Rummeliibacillus 1 0.3% 1 Staphylococcus 90 30.0% 43 Total 300 178 a Bacteria were isolated by plating swabs onto MRS or Rogosa media and then identified using

16S rRNA gene sequencing and biochemical tests.

bLAB, lactic acid-producing bacteria.

130

Table 4.2 Minimum inhibitory concentrations (µg/ml) of antibiotics against 15 bacterial strains isolated from the nasopharynx of feedlot cattlea.

(28C) (72B) (6E) NCDO1193 (3E) (57A) (μg/mL) e (65G) (67A) (63B) (65E) (63A) (38C) (65A) (86D) (103C) (64C)

faecium

Antibiotics rang Tested planterum L. amylovorus L. buchneriL. buchneriL. buchneriL. buchneriL. buchneriL. buchneriL. buchneriL. buchneriL. curvatus L. paracasei L. paracasei L. E. S.epidermidis chromogenes S.

Amoxicillin / clavulanic acid 2:1 2/1-16/8 ≤ 2/1 ≤ 2/1 ≤ 2/ ≤ 2/1 ≤ 2/1 ≤ 2/1 ≤ 2/1 ≤ 2/1 ≤ 2/1 ≤ 2/1 ≤ 2/1 ≤ 2/1 ≤ 2/1 ≤ 2/1 ≤ 2/1 ≤ 2/1 ratio

Azithromycin 0.25-2 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 > 2 Cefepime 0.5-8 ≤ 0.5 4 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 > 8 8 8 ≤ 0.5 1

Cefotaxime 0.12-4 ≤ 0.12 1 ≤ 0.12 ≤ 0.12 ≤ 0.12 ≤ 0.12 0.25 ≤ 0.12 ≤ 0.12 ≤ 0.12 0.25 > 4 > 4 4 0.25 0.5

Ceftriaxone 0.12-2 ≤ 0.12 > 2 0.25 ≤ 0.25 0.25 0.25 0.25 ≤ 0.12 ≤ 0.12 ≤ 0.12 0.5 > 2 > 2 > 2 1 2

Cefuroxime axetil 0.5-4 ≤ 0.5 4 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 ≤ 0.5 4 4 > 4 ≤ 0.5 ≤ 0.5

Chloramphenicol 1-8 ≤ 4 2 ≤ 1 2 ≤ 1 ≤ 1 4 ≤ 1 ≤ 1 4 ≤ 1 4 2 4 4 8

Clindamycin 0.12-1 ≤ 0.12 ≤ 0.12 ≤ 0.12 ≤ 0.12 ≤ 0.12 ≤ 0.12 ≤ 0.12 ≤ 0.12 ≤ 0.12 ≤ 0.12 ≤ 0.12 ≤ 0.12 ≤ 0.12 > 1 ≤ 0.12 0.25

Daptomycin 0.06-2 0.12 > 2 ≤ 0.06 0.12 ≤ 0.06 ≤ 0.06 0.12 2 1 1 0.12 0.25 0.25 2 0.5 > 2

Ertapenem 0.5-4 1 2 2 4 2 2 2 ≤ 0.5 ≤ 0.5 ≤ 0.5 2 > 4 4 4 ≤ 0.5 ≤ 0.5

Erythromycin 0.25-2 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 1

Levofloxacin 0.5-4 4 > 4 4 4 4 4 4 > 4 4 4 1 1 1 ≤ 0.5 ≤ 0.5 ≤ 0.5

Linezolid 0.25-4 2 2 1 2 1 0.5 2 4 4 2 0.5 1 1 2 0.5 2

Meropenem 0.25-2 ≤ 0.25 0.5 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 ≤ 0.25 0.5 2 1 > 2 ≤ 0.25 ≤ 0.25

Moxifloxacin 1-8 ≤ 1 > 8 ≤ 1 ≤ 1 ≤ 1 ≤ 1 ≤ 1 2 ≤ 1 ≤ 1 ≤ 1 ≤ 1 ≤ 1 ≤ 1 ≤ 1 ≤ 1

Penicillin 0.03-4 > 4 0.12 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.06 0.5 0.5 0.25 0.5 2

Tetracycline 1-8 > 8 > 8 > 8 > 8 > 8 8 > 8 > 8 > 8 > 8 > 8 ≤ 1 ≤ 1 ≤ 1 ≤ 1 ≤ 1 > > > Tigecycline 0.015-0.12 0.12 0.03 0.12 > 0.12 0.12 0.06 > 0.12 0.06 ≤ 0.015 ≤ 0.015 0.03 0.06 0.06 0.12 0.12 0.12

131

Trimethoprim/ 0.5/9.5 - ≤ 0.5/9.5 > 4/76 ≤ 0.5/9.5 ≤ 0.5/9.5 ≤ 0.5/9.5 ≤ 0.5/9.5 ≤ 0.5/9.5 2/38 2/38 2/38 > 4/76 ≤ 0.5/9.5 ≤ 0.5/9.5 ≤ 0.5/9.5 ≤ 0.5/9.5 > 4/76 sulfamethoxazole 4/76

Vancomycin 0.5-4 > 4 ≤ 0.5 > 4 > 4 > 4 > 4 > 4 > 4 > 4 > 4 > 4 > 4 > 4 1 2 4 aBreakpoint interpretations are provided within the materials and methods.

132

Table 4.3 Selected genes that had expression altered in bovine turbinate cells after incubation with bacteria isolated from the nasopharynx of cattle.

L. amylovorus L. paracasei M. haemolytica L. buchneri (63A) (72B) (57A) L. buchneri (63B) L. buchneri (65G)

Fold P- Fold P- Fold P- Fold P- Fold P- Fold P- Gene Description change value change value change value change value change value change value

BOLA MHC class I heavy chain 3.68 0.00 1.06 0.73 1.09 0.60 1.04 0.79 1.17 0.36 1.09 0.60

Chemokine (C-C motif) ligand CCL2 2 6.36 0.00 -1.47 0.10 1.08 0.73 1.30 0.39 -1.43 0.11 -1.39 0.15

Chemokine (C-C motif) ligand CCL5 5 782 0.01 1.35 0.59 1.40 0.61 -1.01 0.76 1.29 0.67 1.08 0.94

Chemokine (C-C motif) CCR6 receptor 6 -1.76 0.13 -1.82 0.07 -1.09 0.99 -2.10 0.02 -1.69 0.04 -1.46 0.28

Chemokine (C-C motif) CCR8 receptor 8 -1.07 0.77 2.54 0.05 -1.08 0.88 -1.00 0.74 1.87 0.13 1.94 0.05

CD40 molecule, TNF receptor CD40 superfamily member 5 26.88 0.00 1.21 0.56 1.23 0.54 1.22 0.56 1.40 0.27 1.17 0.70

CD80 CD80 molecule 3.86 0.01 1.05 0.78 1.67 0.20 -1.02 0.75 -1.32 0.22 -1.05 0.97

Colony stimulating factor 2 CSF2 (granulocyte-macrophage) 5.90 0.00 -1.43 0.35 1.11 0.94 -1.29 0.40 -1.68 0.27 -1.62 0.31

Chemokine (C-X-C motif) CXCL10 ligand 10 1321 0.00 -1.67 0.35 -1.08 0.78 -1.92 0.14 -1.53 0.30 -1.48 0.34

CXCL8 Interleukin 8 35.40 0.00 5.32 0.00 8.11 0.00 2.15 0.36 3.98 0.02 4.62 0.00

DEAD (Asp-Glu-Ala-Asp) box DDX58 polypeptide 58 54.62 0.00 -1.10 0.82 -1.04 0.91 -1.26 0.42 -1.02 0.99 -1.28 0.20

Fas (TNF receptor superfamily, FAS member 6) 2.02 0.01 -1.11 0.55 -1.24 0.15 -1.14 0.28 1.02 0.95 -1.05 0.66

Intercellular adhesion molecule ICAM1 1 10.05 0.02 1.00 0.87 1.01 0.95 -1.04 0.73 -1.37 0.17 -1.05 0.73

Interferon (alpha, beta and IFNAR1 omega) receptor 1 1.43 0.02 1.45 0.05 1.65 0.00 -1.08 0.74 1.72 0.02 1.60 0.01

IFNGR1 Interferon gamma receptor 1 -1.79 0.03 -1.11 0.78 1.01 0.78 -1.36 0.47 -1.14 0.58 -1.02 0.98

IL15 Interleukin 15 19.84 0.01 -1.53 0.24 -1.60 0.47 -1.85 0.09 -2.27 0.10 -2.18 0.13

Interleukin 18 (interferon- IL18 gamma-inducing factor) 3.45 0.04 -1.23 0.69 1.14 0.75 -1.07 0.72 -1.12 0.66 -1.12 0.75

133

IL1R1 Interleukin 1 receptor, type I 1.11 0.42 1.13 0.33 1.11 0.35 -1.16 0.28 1.34 0.01 1.29 0.03

IL4 Interleukin 4 -1.83 0.33 -1.56 0.33 -1.19 0.75 -2.12 0.12 -1.69 0.27 -1.63 0.31

IL6 Interleukin 6 (interferon, beta 2) 289.5 0.01 11.68 0.01 8.04 0.02 3.73 0.11 8.63 0.04 9.13 0.04

IRF3 Interferon regulatory factor 3 4.48 0.00 -1.43 0.32 -1.10 0.96 -1.37 0.21 1.12 0.63 -1.14 0.64

IRF7 Interferon regulatory factor 7 67.01 0.00 -1.40 0.55 -1.08 0.85 -1.53 0.28 -1.08 0.53 -1.43 0.38

JAK2 Janus kinase 2 5.74 0.00 -1.10 0.71 1.06 0.71 1.04 0.82 -1.00 1.00 -1.04 0.88

LOC512 Major histocompatibility 672 complex, class I 13.20 0.00 -1.28 0.50 1.09 0.81 -1.03 0.88 1.33 0.29 1.24 0.36

LOC616 Major histocompatibility 942 complex, class I, A-like 7.21 0.00 2.18 0.23 1.18 0.51 -1.19 0.66 -1.14 0.71 1.21 0.46

Myxovirus (influenza virus) resistance 1, interferon- MX1 inducible protein p78 (mouse) 329.9 0.00 1.55 0.49 1.24 0.91 -1.02 0.62 1.08 0.87 1.28 0.81

Nuclear factor of kappa light polypeptide gene enhancer in NFKB1 B-cells 1 8.80 0.00 1.67 0.00 1.27 0.15 1.27 0.30 1.81 0.00 1.75 0.00

Nuclear factor of kappa light polypeptide gene enhancer in NFKBIA B-cells inhibitor, alpha 19.05 0.00 -1.14 0.58 1.06 0.69 1.39 0.39 -1.14 0.28 -1.25 0.10

Solute carrier family 11 SLC11A (proton-coupled divalent metal 1 ion transporters), member 1 2.18 0.05 -1.64 0.27 -1.64 0.47 -1.94 0.03 -1.69 0.08 -1.41 0.21

Signal transducer and activator STAT1 of transcription 1, 91kDa 8.17 0.00 -1.16 0.37 -1.05 0.71 -1.12 0.40 -1.07 0.57 -1.04 0.74

Signal transducer and activator of transcription 3 (acute-phase STAT3 response factor) 2.34 0.00 1.19 0.35 1.32 0.20 -1.03 0.76 1.28 0.19 1.23 0.29

TLR2 Toll-like receptor 2 2.39 0.03 1.00 0.82 1.76 0.04 -1.03 0.78 -1.13 0.84 -1.20 0.62

TLR3 Toll-like receptor 3 4.04 0.00 1.06 0.66 1.08 0.65 -1.10 0.83 -1.12 0.80 -1.17 0.85

TLR4 Toll-like receptor 4 9.28 0.00 1.20 0.75 1.28 0.62 -1.12 0.95 1.24 0.86 1.61 0.35

134

Table 4. 3. Selected genes that had expression altered in bovine turbinate cells after incubation with bacteria isolated from the nasopharynx of cattle (Continued)a

L.curvatus (103C) L. paracasei (3E) L. buchneri (38C) L. buchneri (86D) L. buchneri (67A)

Fold P- Fold P- Fold P- Fold P- Fold P- Gene Description changeb value change value change value change value change value

BOLA MHC class I heavy chain 1.06 0.70 1.01 0.87 1.01 0.90 -1.05 0.70 1.08 0.66

CCL2 Chemokine (C-C motif) ligand 2 1.17 0.51 1.05 0.82 -1.10 0.68 1.04 0.86 -1.47 0.15

CCL5 Chemokine (C-C motif) ligand 5 1.17 0.82 1.00 0.96 -1.13 0.97 1.04 0.99 1.15 0.88

CCR6 Chemokine (C-C motif) receptor 6 -1.56 0.09 -1.54 0.15 -1.79 0.02 -1.25 0.31 -1.06 0.76

CCR8 Chemokine (C-C motif) receptor 8 1.37 0.33 -1.46 0.44 1.19 0.62 1.29 0.47 2.76 0.06

CD40 molecule, TNF receptor CD40 superfamily member 5 1.28 0.46 1.20 0.62 -1.05 0.67 1.08 0.95 1.18 0.71

CD80 CD80 molecule 1.23 0.39 1.20 0.44 -1.08 0.79 -1.01 0.75 1.06 0.88

Colony stimulating factor 2 CSF2 (granulocyte-macrophage) -1.55 0.31 -1.41 0.52 -1.65 0.29 -1.16 0.69 1.43 0.70

CXCL10 Chemokine (C-X-C motif) ligand 10 -1.41 0.34 -1.28 0.56 -1.50 0.33 -1.15 0.71 1.58 0.68

CXCL8 Interleukin 8 8.43 0.02 -1.10 0.61 -1.31 0.12 1.01 0.98 6.99 0.01

DEAD (Asp-Glu-Ala-Asp) box DDX58 polypeptide 58 -1.16 0.39 -1.01 0.94 1.08 0.61 -1.07 0.85 -1.08 0.99 Fas (TNF receptor superfamily, FAS member 6) -1.03 0.73 -1.04 0.80 -1.00 0.95 -1.18 0.25 -1.02 0.88

ICAM1 Intercellular adhesion molecule 1 -1.10 0.65 1.06 0.83 1.15 0.53 1.09 0.71 -1.20 0.41

Interferon (alpha, beta and omega) IFNAR1 receptor 1 2.00 0.01 1.06 0.70 1.00 0.96 1.05 0.76 1.50 0.07

IFNGR1 Interferon gamma receptor 1 -1.16 0.50 1.03 0.73 -1.00 0.78 1.10 0.56 1.06 0.66

IL15 Interleukin 15 -1.78 0.13 -1.37 0.32 -1.61 0.12 -1.27 0.45 1.07 1.00

Interleukin 18 (interferon-gamma- IL18 inducing factor) -1.25 0.54 1.09 0.79 -1.00 0.98 1.20 0.61 1.31 0.56

IL1R1 Interleukin 1 receptor, type I 1.31 0.01 -1.04 0.68 1.04 0.76 1.08 0.53 1.19 0.20

IL4 Interleukin 4 -1.55 0.31 -1.41 0.52 -1.39 0.31 -1.23 0.75 1.43 0.70

IL6 Interleukin 6 (interferon, beta 2) 22.75 0.04 1.36 0.34 1.30 0.35 1.75 0.27 11.44 0.12

IRF3 Interferon regulatory factor 3 -1.18 0.32 -1.30 0.42 -1.19 0.47 -1.05 0.92 -1.37 0.52

IRF7 Interferon regulatory factor 7 -1.12 0.63 -1.14 0.75 -1.37 0.39 -1.11 0.75 1.41 0.70

JAK2 Janus kinase 2 1.03 0.79 1.23 0.44 1.18 0.48 1.26 0.35 1.18 0.41

LOC51267 Major histocompatibility complex, 2 class I 1.34 0.20 1.01 0.94 1.13 0.65 1.26 0.31 1.01 0.85

LOC61694 Major histocompatibility complex, 2 class I, A-like 1.47 0.15 -1.23 0.31 -1.11 0.45 -1.00 1.00 1.21 0.50

Myxovirus (influenza virus) resistance 1, interferon-inducible MX1 protein p78 (mouse) 1.42 0.66 1.07 0.81 1.47 0.52 1.03 0.82 1.13 0.84

Nuclear factor of kappa light polypeptide gene enhancer in B- NFKB1 cells 1 2.13 0.00 1.12 0.30 1.15 0.34 1.14 0.25 1.50 0.03

135

Nuclear factor of kappa light polypeptide gene enhancer in B- NFKBIA cells inhibitor, alpha 1.42 0.16 -1.07 0.54 -1.17 0.34 1.02 0.86 -1.32 0.11

Solute carrier family 11 (proton- coupled divalent metal ion SLC11A1 transporters), member 1 -1.81 0.10 -1.32 0.32 -1.30 0.25 -1.24 0.46 1.04 0.95

Signal transducer and activator of STAT1 transcription 1, 91kDa -1.09 0.57 -1.04 0.71 -1.06 0.74 -1.14 0.50 -1.12 0.45

Signal transducer and activator of transcription 3 (acute-phase STAT3 response factor) 1.40 0.12 -1.02 0.81 -1.06 0.70 1.01 0.99 1.24 0.31

TLR2 Toll-like receptor 2 1.54 0.20 1.20 0.47 -1.00 0.73 1.10 0.63 1.18 0.57

TLR3 Toll-like receptor 3 1.07 0.65 1.09 0.62 1.01 0.73 1.22 0.36 -1.11 0.54

TLR4 Toll-like receptor 4 1.32 0.69 1.28 0.63 2.07 0.14 1.06 0.81 1.35 0.56 aAll values are presented as the mean fold change in gene expression for 4 replications. The

difference in gene expression between bovine turbinate cells (control) that were not cocultured with bacteria cells and the bovine turbinate cells cocultured with bacteria cells were assessed by

Student’s t test for each gene using the Rt2 Profiler PCR Array Analysis software, Version

3.5TM, Qiagen. The level of statistical significance was set at < 0.05. Genes that show

significant differences in expression between control and at least one bacterium co-cultured with

BT cells were listed in the table. Significant changes (P < 0.05) in expression are shown in bold

type.

bFold-Change (2^(- Delta Delta Ct)) is the normalized gene expression (2^(- Delta Ct)) in the test

sample divided the normalized gene expression (2^(- Delta Ct)) in the control sample. Fold- change values greater than one indicate a positive or an up-regulation, and fold-change values

less than 1 indicate a negative or down-regulation.

136

Table 4.4 Antimicrobial properties of selected bacterial therapeutic strains (n = 6) evaluated by the measurement of their lactic acid

and H2O2 production and bacteriocin-encoding genes.

Bacteriocin-encoding genes detected from the whole Lactic acid genome sequences of the selected strains by BAGEL 4 concentration H2O2 concentrations Number of bacteriocin Strain (mM) (± SE)a (nmol/mL) (± SE)a genes Clasess of bacteriocin L. amylovorous (72B) 102.6 ± 0.61 29.0 ± 4.63 5 64.3;Enterolysin_A 70.3;Helveticin-J 6.3;Bacteriocin_helveticin_J 64.3;Enterolysin_A 70.3;Helveticin-J L. buchneri (63A) 79.9 ± 0.87 4.7 ± 4.21 0 NDd L. buchneri (86D) 93.6 ± 7.34 5.2 ± 1.85 0 ND L. curvatus (103C) 110.7 ± 4.74 9.4 ± 3.16 0 ND L. paracasei (3E) 133.6 ± 1.35 0 4 142.2;LSEI_2163 62.3;Enterolysin_A 51.2;Carnocin_CP52 97.2;Enterocin_X_chain_beta L. paracasei (57A) 131.6 ± 3.61 0 NSc ND Cocktail of 6 strains 142.1 ± 8.72 NAb NA ND Control (MRS broth) 0 0 NA ND a The results are reported as the mean (± SE) concentration of lactic acid and H2O2 produced by the selected strains over the 24 h

incubation period. The mean was obtained from triplicate samples. bNA, not applicable; cNS, Not sequenced; dND, bacteriocins not detected in genomes.

137

Figure 4.1 The schematic workflow chart.

138

This chart describes the process of isolation, identification, and screening criteria of commensal bacteria from the nasopharynx of healthy feedlot cattle, to identify candidate bacterial therapeutics for mitigating the bovine respiratory pathogen M. haemolytica.

139

Figure 4.2 Growth-inhibitory effects of bovine respiratory bacteria against M. haemolytica.

(A to D) A summary of lactic acid-producing bacteria (LAB; A) and non-LAB (C) as well as their respective zones of inhibition (B and D).. (E) Species within Lactobacillus displayed the greatest

140

inhibition. (F) Example of the agar slab method to measure inhibitory properties of screened

bacteria. Results are presented as the mean zones of inhibition (plus standard deviations [SD]) from three replicates.

141

Figure 4.3 Adherence of bovine respiratory bacterial isolates to bovine turbinate cell monolayers.

In total, 32 strains within Lactobacillus (A) and 15 strains within Bacillus, Enterococcus,

Macrococcus, Staphylococcus and Streptococcus (B) were evaluated. Results are presented as the

5 mean (± SE) of bacterial adherence (Log10 CFU) to bovine turbinate cell monolayers (10 cells) 142

obtained from three independent experiments performed on different days. Red line represents

5 the cut off value of 5 Log10 CFU per 10 bovine turbinate cells for best adhesion.

A confocal image of L. curvatus (103C; stained green with Alexa Fluor 594) adhering to bovine turbinate cells (stained blue with DAPI) is shown as an example of bacterial adherence (C).

143

Figure 4.4 Antagonistic competition of bovine respiratory bacteria (n = 15) against M. haemolytica.

The results are presented as the mean (± SE) reduction in M. haemolytica adherence to bovine turbinate cell monolayers by commensal bacteria obtained from 6 replicates (A). Different letters indicate mean values differ (P < 0.05). (B) Representative confocal image showing adherence of M. haemolytica (stained red with Alexa Fluor 488) and L. paracasei (3E) (stained green with Alexa Fluor 594) to bovine turbinate cells (stained blue with DAPI). 144

Figure 4.5 Growth inhibition effects of lactic acid on the M. haemolytica.

The values are the means of two replicates. The vertical bars indicate standard deviation (SD).

145

Figure 4.6 Scanning electron microscopy images of Mannheimia haemolytica after incubated with cell-free culture supernatants of selected bacterial therapeutic strains.

146

Bacteria were incubated with cell-free culture supernatant before fixation and microscopy.

Untreated cells (A), and cells incubated with cell-free culture supernatants of L. amylovorus 72B

(B), L. buchneri 63A (C), L. paracasei 3E (D), L. paracasei 57A (E), L. curvatus 103C (F) and L.

buchneri 86D (G). The untreated M. haemolytica cells were intact with regular rod-shape and smooth surfaces (A). However, M. haemolytica cells treated with cell-free culture supernatants of Lactobacillus strains displayed morphological cell damage characterized by shrinkage of the cell surfaces, irregular rod-shape and holes on cell envelop (B, E, F and G)

147

Chapter Five: Intranasal bacterial therapeutics reduce colonization by the respiratory

pathogen Mannheimia haemolytica in dairy calves

Chapter 5 is under review by mSystems.

Amat S, Alexandera TW, Holman DB, Schwinghamer T, Timsit E. Intranasal bacterial therapeutics reduce colonization by the respiratory pathogen Mannheimia haemolytica in dairy calves. Submitted on Sep 29, 2019. mSystems00629-19

148

5.1 Introduction

Bovine respiratory disease (BRD), also known as enzootic calf pneumonia, is one of the

most common diseases in dairy heifers and poses a significant welfare and economic burden on

the dairy industry in North America (Dubrovsky, 2019; USDA, 2012) and Europe (Mahenndran

et al., 2017). This disease is caused by many contributing factors including bacteria, viruses, as

well as management and environmental stressors (Gorden and Plummer, 2010). Among bacterial

pathogens involved in BRD, Mannheimia haemolytica, Pasteurella multocida, Histophilus somni

and Mycoplasma bovis are considered the most important (Panciera and Confer, 2010). These

pathogens can also colonize the upper respiratory tract (URT) of healthy calves without causing

disease (Francoz et al., 2015). However, when respiratory defenses are compromised due to stress

or viral infection, pathogens can proliferate in the URT and translocate into the lung causing

bronchopneumonia (Caswell, 2014). Hence, limiting the proliferation of these pathogens in the

URT may prevent BRD in dairy calves.

Currently, control of bacterial pathogen proliferation and lung infection is largely reliant on the use of antibiotics, which are often given as group medication as early as 10 days of age

(Gorden and Plummer, 2010; Teixeira et al., 2017). Unfortunately, this approach is not always

effective in preventing BRD (Windeyer et al., 2017) and can be seen as an irrational use of

antibiotics leading to the development of antibiotic resistance in the digestive and respiratory

microbiota (Catry et al., 2006). This has been shown by the recent characterization of M.

haemolytica and P. multocida displaying multidrug resistance (Corbeil et al., 1985; Timsit et al.,

2017), which in some instances were linked to resistance genes encoded within integrative

conjugative elements (Woolums et al., 2018; Klima et al., 2014b. Discovery and development of

alternatives to antibiotics to prevent BRD in dairy calves is therefore needed. 149

Increasing evidence indicates that commensal bacteria in the bovine nasopharynx may

prevent the colonization and proliferation of bacterial pathogens through a variety of mechanisms,

including direct antagonism (i.e. antimicrobial properties), competition for nutrients and adhesion

sites, and host immunomodulatory effects (Beker et al., 2018; Holman et al., 2015a;Timsit et al.,

2016a; Zeineldin et al., 2019). Among commensal bacteria, lactic acid-producing bacteria (LAB),

and more specifically Lactobacillus spp., may be important in providing colonization resistance against bacterial respiratory pathogens (Amat et al., 2017; Amat et al., 2019a; Timsit et al., 2018).

In a previous study, we used a targeted approach to characterize bacterial therapeutic (BT) candidates originating from the bovine nasopharynx (Amat et al., 2019b). In total, six

Lactobacillus strains from an initial group of 178 LAB isolates were identified to have the greatest potential as BTs based on their ability to inhibit the growth of M. haemolytica in vitro, adhere and exclude M. haemolytica from bovine turbinate cells, and modulate expression of genes related to bacterial infection in turbinate cells.

In the current study, we further characterized these BT strains by testing their effectiveness in vivo in reducing nasal colonization by M. haemolytica, modulating the nasal microbiota and stimulating an immune response in dairy calves challenged with M. haemolytica. Nasal colonization by BT strains and M. haemolytica, nasal and tracheal microbiota composition, and serum cytokine concentrations were compared between two groups of dairy calves that were either administered BTs 24 h prior to M. haemolytica challenge, or not administered BTs prior to challenge. As part of a safety assessment of the BTs, gross pathological examination of the lungs was also conducted at the end of the study period.

150

5.2 Materials and Methods

5.2.1 Ethics statement

This study was conducted in strict accordance with the recommendations of the Canadian

Council of Animal Care (Olfert et al., 1993). The research protocol was reviewed and approved by the University of Calgary Veterinary Sciences Animal Care Committee (AC17-0003).

5.2.2 Animals and husbandry

Twenty-four Holstein bull calves from six different local dairy farms were used in the study. Calves were separated from the dam immediately after birth to prevent suckling and were fed 3 L of a colostrum replacer (Calf's Choice Total HiCal, SCCL, Saskatoon, SK) within 3 h of birth to provide adequate passive immune transfer and reduce variability among calves. Calves were then transported within 3 h to a research facility, housed individually and fed another 3 L of colostrum replacer. Navels were sprayed with 7% iodine solution daily for 3 d following birth. A milk replacer enriched with a colostrum substitute (20% of total solids; Hical, SCCL) was fed to the calves twice daily until 7 days of age. From 7 days onwards, calves were fed only milk replacer and the amount of milk replacer supplied adjusted based on body weight. Clean water was available at all times and starter grain was fed ad libitum after 14 days of age.

5.2.3 Study design

The challenge study was repeated twice at the same research facility with 12 calves per replicate. The second replicate started approximately one month after the first replicate was completed. The schematic of experimental design and sampling regimen are presented in Figure

5.1. After an acclimatization period (at least 7 days), all calves were sampled using nasal swabs

(NS) to evaluate their M. haemolytica and LAB status (i.e. positive or negative; day -1). On d 0, calves were blocked by age, farm of origin, as well as M. haemolytica and LAB culture results

151

obtained on d -1. They were then randomly assigned into either BT + Mh or Mh groups (n = 6 per group). On d 1, calves in the BT + Mh group were administered into each nasal cavity 3 mL of phosphate buffered saline (PBS) containing a multi-strain cocktail of 6 Lactobacillus strains in equal concentrations (1 × 109 CFU mL-1), whereas the Mh group was administered 3 mL of PBS without bacteria. On d 2, all calves were challenged with 1 × 108 CFU mL-1 of M. haemolytica by administering 3 mL into each nasal cavity. This dose has previously been established to represent a colonization model for M. haemolytica without inducing bronchopneumonia (Frank et al., 1995;

Boudreaux, 2005). The intranasal delivery of PBS with or without bacteria was applied to the ventral meatus, 4-5 cm from nostril entrance, using a sterile 12-mL syringe fitted with a 9-cm long sterile tube (Supplementary Figure S5.1).

During the study period, calves were monitored daily by an experienced bovine veterinarian for the following clinical signs: demeanor, appetite, rectal temperature, respiratory rate, nasal and ocular discharge and presence of abnormal sounds at lung auscultation. Nasal swab samples were collected on days -1, 3, 5, 7, 9, 11, 13 and 16, and blood samples were collected on days -1, 3, 5 and 13 (as shown in Figure 5.1). Calves were euthanized on d 16, and sampled immediately by trans-tracheal aspiration (TTA) and their lungs were evaluated for gross lesions at necropsy.

Sample size calculation

The sample size (n = 12 calves per treatment group) was calculated to detect, at least, a

50% difference in the proportion of calves positive for M. haemolytica 9 d after challenge between

152

the BT + Mh and Mh groups. For this calculation, α and β were set at 0.05 and 0.20, respectively

(two-sided test).

5.2.4 Preparation of the nasal inoculum

The M. haemolytica (Guelph 1) strain used for the challenge was isolated from a BRD-

affected beef calf in a feedlot, and identified as serotype 1 (Klima et al., 2011). This strain was

plated onto brain heart infusion (BHI, Oxoid, Nepean, ON, Canada) agar and a single colony was

inoculated into 50 mL of BHI broth, followed by incubation with shaking at 37ºC for 18 h. The

culture was then centrifuged at 3000 × g for 20 min and the cell pellet was re-suspended in 10 mL

of PBS, and washed twice using PBS. The cell pellet was diluted in PBS to obtain a final

suspension containing 1 × 108 CFU mL-1 for administration to calves.

The BT cocktail included the following 6 Lactobacillus strains: L. amylovorus (72B), L.

buchneri (63A and 86D), L. curvatus (103C) and L. paracasei (3E and 57A), each at concentration

of 1×109 CFU mL-1 (Figure 5.1). These Lactobacillus isolates were plated onto Lactobacillus De

Man, Rogosa and Sharpe (MRS) agar (Dalynn Biologicals, Calgary, AB, Canada) and incubated

for 48 h at 37ºC in 10% CO2. One day prior to nasal inoculation, a single colony of each strain was inoculated individually into 5 mL Difco Lactobacilli MRS broth (BD, Mississauga, ON, Canada)

and incubated at 37ºC with agitation at 200 rpm. After 18 h incubation, each bacterial culture was

centrifuged at 7,600×g for 10 min, the supernatant discarded and the pellet re-suspended with PBS.

Aliquots of each strain were then mixed together to achieve a target concentration of 1 × 109 CFU

mL-1 to serve as the BT cocktail for administration. Inoculants for calves were prepared on the

day of administration. The concentrations of M. haemolytica and individual BT strains

(immediately prior to mixing as a cocktail) were confirmed by plating and enumeration.

153

5.2.5 Sampling and processing of NS

Nasal swab samples were collected using 10-cm long unprotected swabs, which had tips

flocked with soft nylon fiber (flocked swab; 480C ESwa Liquid Amies Collection and Transport

System, Copan, Murrieta, USA). Two NS were collected from each calf (one from each nostril).

Before sampling, each nostril was wiped with a clean paper towel to remove any debris or nasal

discharge. Immediately after sample collection, tips of the swabs were broken and placed in 1 mL

of sterile Amies transport media. The swabs were stored at 4ºC overnight and processed the

following morning.

For processing, the tip was cut from the swab and placed into 0.7 mL of BHI broth. The

transport tube containing the Amies medium was then centrifuged (2,000×g, 5 min) and 75% of

the supernatant was discarded. The pellet was re-suspended with the remaining supernatant,

transferred to the BHI tube containing the rayon tip of the swab and vortexed for 30 s.

5.2.6 Isolation and enumeration of M. haemolytica from NS

For enumeration of M. haemolytica, a 100 µL aliquot of the swab/BHI suspension was

serially diluted in PBS. The dilutions were plated onto tryptic soy agar (TSA) containing 5% sheep

blood supplemented with 15 µg mL-1 of bacitracin (to limit the growth of Gram-positive bacteria

(Catry et al., 2016) [Dalynn Biologicals]) and incubated overnight at 37ºC. Each swab from the

right and left nasal cavities was processed separately and bacterial counts for each animal were

obtained by averaging CFUs from both swabs. Up to three colonies displaying typical morphology

indicative of M. haemolytica (white-grey, round, medium-sized, non-mucoid, exhibiting β- haemolysis) were sub-streaked onto TSA with 5% sheep’s blood and incubated overnight at 37ºC.

Three glycerol stocks (1.2 mL; BHI: glycerol, 80:20%) for each M. haemolytica isolate from these plates were prepared and immediately stored at -80ºC for further analyses (i.e. pulsed field gel

154

electrophoresis, see below). In addition, a loop of each isolate was also stored in 100 µL Tris-

EDTA (TE) at -80ºC for PCR confirmation.

5.2.7 PCR identification and pulsed field gel electrophoresis typing of M. haemolytica

DNA was extracted from each M. haemolytica isolate using heat lysis at 98°C for 3 min

and used for PCR confirmation of M. haemolytica as described by Klima et al. (2014b).

Afterwards, a subset of PCR confirmed M. haemolytica isolates (n = 44) was randomly selected

(per isolate per animal per sampling time point) and typed by pulsed-field gel electrophoresis

(PFGE) to determine relatedness with M. haemolytica Guelph 1. PFGE typing was performed as

described previously Klima et al., (2014b) and PFGE profiles were analyzed using BioNumerics

7.6 (Applied Maths, Inc., Austin, TX, USA).

5.2.8 Isolation of LAB from NS

MRS agar, which is semi-selective for LAB, was used to isolate Lactobacillus spp. A 100

µL aliquot from each NS was serially diluted, plated on MRS agar, incubated for 48 h in a 10%

CO2-enriched environment at 37ºC. Up to 10 colonies from each animal were subcultured on MRS

agar. The subcultures were then processed for glycerol stocks (MRS: glycerol, 80:20%), and TE

stocks as described for M. haemolytica. DNA isolation from TE stocks was used for PCR

confirmation and subtyping (see below).

5.2.9 PCR identification and rep-PCR typing of Lactobacillus isolates

Because some non-LAB species such as Bacillus and Staphylococcus spp. grow on MRS

agar (Holman et al., 2015b), isolates within Lactobacillus were first identified using a

Lactobacillus genus-specific PCR assay. Briefly, an aliquot of 20 µL of TE stock was heat lysed

at 98oC for 3 min, and the lysate was used to identify Lactobacillus isolates using genus-specific

155

primers according to Dubernet et al. (2002). The remaining TE stock from isolates that tested positive as Lactobacillus was processed for genomic DNA extraction using a Qiagen DNeasy

Tissue kit (Qiagen Inc., Mississauga, ON, Canada) as described by Holman et al. (2016a).

The extracted genomic DNA was used for molecular typing using repetitive sequence- base (GTG)5 PCR (rep-PCR) method. The rep-PCR was performed using the (GTG)5 primer

(Versalovic et al., 1994). Briefly, the PCR mixture (20-µL) contained 10 µL of Hotstart Master

Mix (Qiagen Inc.), 2 µM of primer, 2 µL of genomic template DNA and 6 µL of H2O. DNA

fragments were amplified using an Eppendorf Mastercycler Pro Thermal Cyclers (Eppendorf

Canada, Mississauga, ON, Canada) using the following PCR conditions: initial denaturation at

94ºC for 5 min, followed by 35 cycles consisting of denaturation at 94ºC for 30 s, annealing at

40ºC for 60 s, extension at 72ºC for 10 min, and a 10 min final extension step at 72ºC.

The (GTG)5-PCR amplicons were analyzed on 1.5% (wt/vol) Seakem gold agarose (Lanza,

Rockland, ME, USA) gels in Tris-borate-EDTA buffer (Thermofisher Scientific, Ottawa, ON,

Canada). A 1 Kb Plus DNA Ladder (Invitrogen, Carlsbad, CA, USA) was used as a molecular

size marker according to the manufacturer’s directions. The electrophoresis was performed using

a Bio-Rad PowerPac 300 (Bio-Rad Laboratories, Redmond, WA) at 4.0 V/cm for 16 h at 4 ºC.

Gels were stained with ethidium bromide (5 µg/mL, Bio-Rad) and visualized under UV light using

ChemiDoc MP imaging system (Bio-Rad).

5.2.10 Blood sample collection and processing

Blood (4 mL) was collected from the jugular vein using plain collection tubes (BD

Vacutainer Blood Collection Tubes, Becton, Dickinson and Company, Franklin Lakes, NJ). Blood samples were processed within 6 h and centrifuged at 2,000 ×g for 10 min at room temperature.

Resulting sera were then stored at -80ºC until further analysis. 156

5.2.11 Quantification of cytokines in serum using enzyme-linked immunosorbent assay (ELISA)

The concentrations of the cytokines IL-6, IL-8, IL-10 and TNF-α from serum samples collected on d -1, 3, 5 and 13 were determined using commercially available ELISA kits [Bovine

IL-4, IL-6, or TNF-α Screening Set (Thermo Scientific)]; Human CXCL8/IL-8 DuoSet (R & D systems, Wiesbaden, Germany); Bovine IL-10 ELISA Kit (MyBioSource Inc. San Diego, CA,

USA) according to manufacturer’s instruction.

5.2.12 Trans-tracheal aspiration sampling and processing

Trans-tracheal aspirations were carried out as described by Timsit et al. (2017) on each calf immediately after euthanasia to characterize the tracheal microbiota. Briefly, 50 mL of sterile saline (0.9% NaCl) was introduced in a 75 cm long trans-tracheal catheter (Centracath, Vygon,

Ecouen, France) using a 50-mL syringe. Immediately after injection, gentle suction was provided by withdrawing the plunger. On average, 5 - 10 mL of tracheal fluid were recovered and immediately placed into empty sterile tubes. Samples were placed on ice and stored at -80ºC within

12 h of collection.

5.2.13 Post-mortem examination

As part of the safety assessment for BT inoculation, all calves were euthanized at the end of experiment (d 16) and necropsied by a board-certified veterinary pathologist at the Diagnostic

Services Unit of the University of Calgary. Each lung was removed and examined for gross pathology. In addition, digital photos were taken from each of the pulmonary lubes for visual examination of gross lesions and lung lesion scoring. The percent of lung tissue affected was calculated from the lung images. Briefly, transparent lattice safe on GIF was placed on top of the lung image placed on PowerPoint. The number of line-points (each line has given 2 points) was

157

used to determine the volume density of a particular lesion. The number of points falling in the

entire lung were counted followed by counting the number of points falling in the pneumonic lung.

The percent of affected lung tissue were calculated by dividing the number of points from the

particular lesion by the number of points from entire section of the lung.

5.2.14 DNA extraction from NS and TTA samples and 16S rRNA sequencing

Total DNA was extracted from the remaining NS suspensions using a method with enzyme

and mechanical cell disruption, as previously detailed (Holman et al., 2015a). For DNA extraction

from TTA samples, 2 mL of each TTA sample was transferred into a 2 mL centrifuge tube and

centrifuged at 15,000 × g for 3 min. The supernatant was discarded and DNA was extracted from

the pellet using the same method as for NS.

The nasal and tracheal microbiota were characterized through sequencing of the V4

hypervariable region of the 16S rRNA gene using the MiSeq reagent kit v2 (500 cycles) and an

Illumina MiSeq (Illumina, San Diego, CA, USA) as previously described (Holman et al., 2017).

The 16S rRNA gene sequences were processed using DADA2 v. 1.8.0 (Callahan et al., 2016) in R

v. 3.5.0. Briefly, primer sequences were removed, forward and reverse reads were truncated at 220

bp, and reads with a maximum number of expected errors greater than 2 were eliminated. The

forward and reverse reads were then merged and chimeric sequences removed. The RDP naive

Bayesian classifier (Wang et al., 2007) and the SILVA SSU database release 132 (Quast et a.,

2012) were used to assign taxonomy to each merged sequence, referred to here as operational taxonomic units (OTUs) at 100% similarity. The Shannon diversity index, the number of OTUs per sample and the pairwise Bray-Curtis dissimilarities were calculated using the R packages vegan v. 2.5-2 (Oksanen et al., 2007) and phyloseq v. 1.24.0 (McMurdie et al., 2013).

158

Twelve extraction control samples were also included to assess potential contamination

during DNA extraction and sequencing. Any OTUs that were identified as predominant in the

negative controls were removed prior to analysis.

5.2.15 Estimation of Lactobacillus spp. abundance from NS using qPCR

Because of the semi-selective nature of MRS plates, accurate counts of Lactobacillus spp.

could not be achieved. Instead, Lactobacillus abundance was estimated via qPCR. A Lactobacillus

group-specific PCR primer, S-G-Lab-0677-a-A-17 was used to selectively amplify the 16S rRNA gene from the members of Lactobacillus genus (Heilig et al., 2002). Each qPCR mixture contained

1X iQ SYBR Green Supermix (Bio-Rad Laboratories Inc.), 0.4 μM of each primer, 0.1 μg/μL BSA

(New England Biolabs,180 Pickering, ON, Canada), and 25 ng of DNA extracted from the NS, in a total volume of 25 μL. A CFX96 Touch Real-Time PCR Detection system (Bio-Rad

Laboratories Inc.) with the following conditions was used: an initial denaturation at 95ºC for 3 min, followed by 40 cycles at 95ºC for 25 sec, 50ºC for 30 sec, and then 72ºC for 45 sec. Standard curves (102 to 106 gene copies) were produced using the pDrive cloning vector (Qiagen Inc.)

containing the PCR product from Lactobacillus strain (L. paracasei 57A). A melt curve analysis

was performed following amplification for all qPCR reactions to ensure only target genes were

amplified.

5.2.16 Statistical analysis

The mean M. haemolytica counts for each animal were obtained by averaging CFUs from

the two NS collected per animal (right and left nasal cavities). Bacterial counts, rectal temperatures

and respiratory rates were compared between treatment groups (BT + Mh and Mh). The M.

haemolytica counts were natural log-transformed prior to statistical analysis. All data were

analyzed as a randomized complete block design with repeated measurements using a Linear

159

Mixed-Effects Model. This model accounted for the repeated measurements and also adjusted for day of sampling and M. haemolytica counts before challenge (i.e. baseline). Animal (n = 24) was included as a random effect. The parameters were estimated using Restricted Maximum

Likelihood. All analyses were done in R version 3.5.0 (R Core Team, Vienna, Austria) and a P <

0.05 was considered significant.

The number of 16S rRNA gene sequences in each NS and TTA was randomly subsampled

to 3,850 and 25,500 sequences, respectively, prior to the calculation of alpha-diversity metrics and

the Bray-Curtis dissimilarities. Permutational multivariate analysis of variance (PERMANOVA)

with 10,000 permutations was used to determine the effect of treatment, and sampling time on the

microbial community structure. Differentially abundant OTUs in the nasal microbiota of the

BT+Mh and Mh groups were determined using DESeq2 (Love et al., 2014). Only those OTUs

found in at least 10% of samples were included in the analysis and the Benjamini-Hochberg

procedure was used to account for multiple comparisons.

The relative abundance of specific taxa from the NS and TTA samples, the qPCR data, as

well as serum cytokine concentrations were analyzed using the GLIMMIX procedure in SAS (SAS

9.4, SAS Institute Inc., Cary, NC). The individual, block, treatment, and time were included in the

CLASS statement. The models were “generalized” due to the specification of residual distributions

that were not Gaussian normal. Models were “mixed” due to the inclusion of fixed effects

(treatment-nested-in-time and time) and random effects (block and individual). Variance

heterogeneity was modeled using a “RANDOM_RESIDUAL_/GROUP = Treatment*Time”

statement. Residual distributions and covariance structures were selected for each genus based on

the model fit statistics, i.e., Bayesian information criterion (BIC). Preliminary models that

specified the beta-binomial distribution did not converge. Therefore, alternative distributions were 160

tested: Gamma, inverse Gaussian, lognormal, shifted t, Gaussian normal, exponential, and geometric.

Path analysis was performed to evaluate the potential interrelationship structure among nasal microbial community. Path analysis is a member of the structural equation modeling tools that enable to identify causal relationship between the measured variables (Schwinghamer et al.,

2017). For path modeling, data on the 10 most relatively abundant genera in the nasal microbiota were broken into two subsets based on treatment group. Pearson’s product-moment correlation coefficient detects the strength and direction of linear relationships among variables and the use of

Spearman’s correlation coefficients allows for nonlinear (monotonic increasing or decreasing) relationships among the studied variables. Therefore, path analysis was based on the matrix

Spearman rank-based correlation coefficients for each subset that were produced using SAS PROC

CORR with the SPEARMAN option (SAS 9.4). In the initial model, time and the relative abundance of Lactobacillus spp. (the BT inoculation) were exogenous (predictor or independent) variables. In path diagrams, single-headed arrows that represent causal paths point from but never point to exogenous variables. The sources of variability in the exogenous variables are not included in the model. The initial model hypothesized that the experimental variables (time and

Lactobacillus) predicted variability in the relative abundance of the studied genera (that were therefore hypothesized to be endogenous or criterion variables). As such, the initial model was entered into SAS PROC CALIS:

Time. Lactobacillus ↓ {Mannheimia, Moraxella, Acinetobacter, Bifidobacterium, Streptococcus, Lactobacillus, Prevotella, Bacteroides, Klebsiella} Path models were modified by adding and subtracting paths and covariance using the PATH and

PCOV statements, respectively, based on suggestions made by Lagrange modifier statistics that 161

were obtained by the inclusion of the MODIFICATION option in the PROC CALIS statement.

Time was exogenous and endogeneity was not considered with respect to time. Lower values of

Schwarz’s Bayesian criteria (SBC) indicated superior model fitting, in comparison to the initial hypothetical model, and therefore the plausibility and appropriateness of the respective modified model structures.

5.3 Results

5.3.1 Animal health

None of the calves displayed abnormal clinical signs (e.g. lethargy, diarrhea, cough, nasal

discharge and increased respiratory rate) throughout the study period. Mean rectal temperatures and respiratory rates were within normal ranges (Table S5.1) and did not differ between BT + Mh

and Mh groups (P > 0.05) (Table S5.2). Gross examination of the lungs at d 16 revealed that 10

calves had lung consolidation. However, percentages of consolidation were low (median = 4.6%;

min = 0.8%; max = 22.6%) and did not differ between groups (P > 0.05) (Supplementary Figure.

S5. 2).

5.3.2 Isolation and enumeration of M. haemolytica from NS

M. haemolytica-specific PCR analysis confirmed that > 98% of isolates that were subcultured were M. haemolytica. Bacterial counts obtained from culturing the NS revealed that

BT administration was significantly associated with a 3.3 natural log CFU reduction in M.

haemolytica per NS (P = 0.02; Table 5.1). As shown in Figure 5.2A, there was a significant

increase in M. haemolytica counts in both groups on d 3 (i.e. 24 h after M. haemolytica inoculation)

compared to d -1 (P < 0.05). However, M. haemolytica counts remained significantly lower in BT

+ Mh calves on d 3, 5 and 7 compared to Mh calves (P < 0.05). Effects of time (P < 0.01), replicate

162

(P = 0.01) and interactions between time and treatment (P = 0.02) on M. haemolytica counts were also detected (Table 5.1).

5.3.3 PFGE typing of M. haemolytica isolates

The PFGE typing of PCR-confirmed M. haemolytica isolates showed that the inoculated

Guelph 1 strain, as indicated by pulsotype, was not isolated from calves prior to its administration on d 3 (Supplementary Figure S5.3). The Guelph 1 strain was the most dominant M. haemolytica isolated from calves after day 3, indicating a high rate of colonization of calves with this strain.

However six additional pulsotypes were observed showing that some calves were colonized with different strains of M. haemolytica, but these strains were less frequently observed.

5.3.4 Lactobacillus abundance in NS determined by qPCR

As MRS agar is only semi-selective and allows the growth of non-LAB such as Bacillus and Staphylococcus spp., qPCR was used to estimate the abundance of Lactobacillus from the metagenomic DNA extracted from NS. No significant difference in total Lactobacillus copy numbers per NS was observed between the two groups with an exception of d 5, when the BT +

Mh group had significantly greater abundance of Lactobacillus 16S rRNA copies per swab, compared to the Mh group (P < 0.05) (Figure 5.2B).

5.3.5 Rep-PCR typing of Lactobacillus isolates

An example of (GTG)5 fingerprints is shown in Figure S5.4. Molecular typing of PCR- confirmed Lactobacillus spp. revealed that BT strains were not detected in any NS collected on d

-1. In addition, the BT strains were not detected in any Mh calves. On d 3, 5, 7 and 13, respectively,

(GTG)5 fingerprints of Lactobacillus isolates were observed that matched the BT strains, but detection was variable (Supplementary Table S5.3). Of the fingerprints that were representative of

163

the BT strains, only L. amylovorus 72B was detected in calves beyond d 7, but detection was limited to two calves.

5.3.6 Effects of BT on the composition and diversity of the nasal microbiota

In total, 8,869,675 16S rRNA gene sequences (sequences per sample: median = 55,055, min = 3,853 and max = 91,448) were obtained from 179 NS samples with 4,824 unique archaeal and bacterial OTUs identified. PERMANOVA revealed that BT administration (R2 = 0.011; P =

0.02), sampling time (R2 = 0.052; P < 0.01) and replicate (R2 = 0.028; P < 0.01) had significant,

but relatively minor effects on the microbial structure of the nasal microbiota (Figure 5.3).

Individual animal had the largest effect (R2 = 0.247, P < 0.001), explaining 24.7% of the variability observed. Although a treatment effect on the microbial community structure was not observed during the first 13 days, there was a larger and significant effect on d 16 (R2 = 0.15; P = 0.01) with

19 OTUs differentially abundant between the two groups (Table S5.4; FDR < 0.05).

Twenty-five different bacterial phyla were observed among all NS samples but only

Proteobacteria, Actinobacteria, Firmicutes, and Bacteroidetes had a relative abundance greater

than 0.5% (Figure 5.4A). As shown in Fig 4B, the relative abundance of these phyla changed over

time. The relative abundance of Proteobacteria, which includes the genus Mannheimia, increased

between d -1 and d 3 and then decreased afterwards, whereas the relative abundance of

Actinobacteria increased throughout the study period to become the most relatively abundant

phylum at d 16.

On d -1, of the 10 most relatively abundant genera, Lactococcus was the predominant genus

with a relative abundance of 10.9%, followed by Moraxella (9.0%), Acinetobacter (6.0%),

Lactobacillus (4.3%) and Bifidobacterium (4.0%) among all animals (Figure 5.4C). However, after

M. haemolytica inoculation (i.e. from d 3 to 16), there was a 29-fold increase in the relative 164

abundance of Mannheimia, which became the most relatively abundant genus (14.4%; Figure

5.4C) across treatments. During that period, the relative abundance of Lactococcus declined from

10.9% to 4.23% and Lactococcus spp. became the third most abundant genus after Mannheimia and Moraxella.

A comparison of the 10 most relatively abundant genera between the two groups revealed that the relative abundance of these genera varied by sampling time (Figure 5.4D). For example, the nasal microbiota of BT+Mh calves harbored a greater abundance of Lactococcus on d 7 but

Lactococcus was more relatively abundant in the Mh calves on d-1 and d 16 (P < 0.05). The relative abundance of Lactobacillus did not differ by treatment group at any sampling day except for d 11 and 16 during which Mh calves had a significantly greater relative abundance of Lactobacillus compared to BT+Mh calves (P < 0.05). The relative abundance of Klebsiella spp. was lower on d

7 but greater on d 9 in BT+Mh calves compared to Mh calves (P < 0.05). None of the other relatively abundant genera differed significantly between the two groups at any sampling time (P

> 0.05).

In terms of alpha diversity, the number of OTUs per sample (richness) was affected by both treatment (P = 0.04) and time (P < 0.01) (Figure 5.5A). The Mh calves had a significantly higher number of OTUs on d 11 compared to the BT+Mh calves (P < 0.01). No difference was observed between groups for the Shannon diversity index except on d 16, with greater diversity in the Mh calves (P = 0.04; Figure 5.5B).

5.3.7 Effects of BT on the recursive structure of causal relationships among the 10 most relatively abundant genera in the nasal microbiota

To evaluate whether the causal relationships among genera in nasal microbiota changed in response to the intranasal inoculation of BT, path analysis was performed on the relative

165

abundance data of the10 most abundant genera from all sampling days. The matrices of Spearman

correlation coefficients that were used as the basis for path analysis are provided as Tables S5.5

and S5. 6. Two path models (Figure 5.6, Table 5.2) were selected based on the model fit statistics

(Table 5.2). Based on the values of the squared multiple correlation, R2, model 1, constructed

based on the data from Mh group, explained 11.90 to 51.47% the variance in relative abundances

of the endogenous genera, while model 2 (BT + Mh group) explained 8.43 to 44.94% of the

variance in relative abundances of the endogenous genera (Table 5.7).

Path modelling (Figure 5.6, Table 5.2) indicated a distinct causal relationship among the

10 most relatively abundant genera of nasal microbial community between Mh and BT + Mh

groups. The relative abundances of seven genera, including Acinetobacter, Bacteroides,

Bifidobacterium, Lactococcus, Mannheimia, Prevotella, and Streptococcus were endogenous

(variance explained by factors inside the model such as relative abundance of other genera, and time) while the relative abundances of three genera (Klebsiella, Lactobacillus, and Moraxella) were exogenous (variance explained by unmeasured factors outside of the model) in the Mh group.

Within the nasal microbial community of Mh calves, 16.2% of the variance in the relative abundance of Mannheimia was explained by a combination of other genera and sampling time

(Table S5.7). For example, the genus Lactobacillus was positively linked with the relative abundance of Bifidobacterium, which in turn promoted Lactococcus. Lactococcus, on the other hand, was negatively associated with the relative abundance of Mannheimia. In addition,

Prevotella was predicted to have indirect positive effect on Mannheimia by negatively impacting

the Mannheimia inhibitor Lactococcus. The relative abundance of Prevotella was predicted to be

positively affected by Streptococcus, Acinetobacter and Klebsiella and time.

166

Interestingly, the path model for BT + Mh calves differed from Mh calves indicating that

causal relationships between the 10 most abundant genera were likely altered by BT

administration. Mannheimia was found to be an exogenous variable in the BT + Mh group whose

relative abundance was not affected by any of the other genera or time. The relative abundance of

Moraxella was also an exogenous variable in the BT + Mh group. Unlike the Mh group, no

associations between Lactococcus, Prevotella and Mannheimia were predicted. Of note, both

Lactobacillus and Mannheimia were predicted to be positively associated with Streptococcus. The

relative abundance of Klebsiella was predicted to decline over time in BT + Mh group but not in

the nasal microbiota of Mh calves.

5.3.8 Effects of BT on the composition and diversity of the tracheal microbiota

For the TTA samples collected on d 16, there were 1,585 OTUs identified among 1,895,098

sequences (sequences per sample: median = 78,297, min = 25,968, max = 141,708).

PERMANOVA indicated that the tracheal microbiota structure on d 16 was not significantly

affected by BT treatment (R2 = 0.065; P = 0.10) or replicate (R2 = 0.071; P = 0.07) (Figure 5.7A).

The richness (number of OTUs) and diversity (Shannon diversity index) of the tracheal microbiota

also did not differ between the BT+Mh and Mh calves (P > 0.05; Figure 5.7B and C).

Twenty-nine bacterial phyla were detected among the TTA samples with six having a relative abundance greater than 0.5% across treatments: Actinobacteria (65.6%) Proteobacteria

(9.8%), Firmicutes (9.7%), Bacteroidetes (9.2%), Fusobacteria (2.1%) and Tenericutes (1.5%)

(Figure 5.7D). The Mh calves had a significantly greater relative abundance of Proteobacteria and

Fusobacteria compared to BT + Mh calves (P < 0.05). Actinobacteria, however, tended (P = 0.07) to be more abundant in BT + Mh group.

167

A total of 392 genera identified within the tracheal microbiota, however, only 13 had an overall relative abundance greater than 0.5%. Of the 10 most relatively abundant genera, only three (Mannheimia, Lactobacillus and Prevotella) were also among the most relatively abundant genera detected in the nasal microbiota (Figure 5.7E). Of note, the relative abundance of

Mannheimia (P < 0.0001) and Pseudomonas (P < 0.01) was significantly greater in Mh calves compared to BT + Mh calves but the relative abundance of Lactobacillus did not differ between these two groups. Though relative abundance was found to be highly variable the mean abundances of Ureaplasma and Caviibacter genera were also significantly greater in the Mh group relative to

BT + Mh group (P < 0.001).

5.3.9 Effects of bacterial therapeutics on serum cytokine concentrations

No difference in concentrations of serum cytokines IL-6, IL-8 and IL-10 was observed between the two groups at any sampling times (P > 0.05) (Figure 5.8). TNF-α was not detected from the serum samples of any calf at any sampling times (detection limit of the ELISA kit used was 0.1 ng/mL).

5.4 Discussion

Increasing evidence indicates that commensal members of the bovine respiratory microbiota have competitive relationships with opportunistic pathogens and may be involved in a microbiota-mediated defense against respiratory infection (Zeineldin et al., 2019; Amat et al.,

2019a,b). Therefore, we recently screened commensal bacteria (n = 178) isolated from the respiratory tract of healthy cattle for the development of intranasal BT to mitigate the BRD pathogen M. haemolytica, as an alternative to metaphylactic antimicrobial use. By ranking BT candidates based on in vitro inhibition of M. haemolytica, as well as epithelial cell colonization,

168

and immunomodulation, we identified six Lactobacillus strains for evaluation in cattle. In the

study reported here, we further evaluated the in vivo potential of these selected strains when

administrated intranasally in young dairy calves that were subsequently challenged with M.

haemolytica.

Intranasal administration of the BT cocktail did not adversely affect calf health, implying

that the dose of inoculation tested was tolerated by young dairy calves. The Lactobacillus strains

we selected for nasal inoculation upregulated (moderate level) gene expression of IL-6 and IL-8 in bovine turbinate cells in vitro (Amat et al., 2019b). However, in the present study, there were no significant differences in serum concentrations of IL-6, IL-8 and IL-10 between Mh and BT +

Mh calves. Unlike the previous in vitro experiment where these selected Lactobacillus strains were co-cultured with turbinate cells without the presence of M. haemolytica, in the present study, all calves were challenged with M. haemolytica after BT inoculation. There were no groups of calves that only received BT nor a control group that received neither BT nor M. haemolytica. Therefore, it was difficult to extrapolate the findings from in vitro experiment into this challenge study.

The BT cocktail reduced nasal colonization by M. haemolytica, which was expected based on our previous in vitro studies (Amat et al., 2017; Amat et al., 2019b). Indeed, all six

Lactobacillus strains inoculated have been shown to inhibit M. haemolytica adherence to bovine turbinate cells by 32% to 78% (Amat et al., 2019b). Furthermore, these strains produced 80 to 142 mM of lactic acid after 24 h of incubation, which is sufficient to inhibit the growth of M. haemolytica (Amat et al., 2019b). Finally, genome sequencing showed that two of the strains (L. amylovorus 72B and L. paracasei 3E) (Amat et al., 2019b,c) encoded bacteriocins which have bactericidal or bacteriostatic activity (Pridmore et al., 2008; Parada et al., 2007). Therefore, the reduction in M. haemolytica colonization could have partially resulted from competition for 169

adherence to the nasal mucosa or direct inhibition of M. haemolytica. It is also possible that the

BTs modified the nasal microbiota in a way that conferred increased colonization resistance to M. haemolytica. A previous study showed that microbiota manipulation using probiotic Lactobacillus

strains reduced the colonization of vancomycin-resistant Enterococcus (VRE) in challenged mice

(Li et al., 2019). The authors proposed that the compositional alteration observed from phylum to

species level by probiotic strains may have contributed to decreased VRE colonization. The altered

microbial community structure observed in our study, as well as differences observed in the causal

relationships between the10 most relatively abundant genera of nasal microbiota between BT +

Mh and Mh treatments, suggest that microbiota-mediated colonization resistance may also have resulted in increased resistance to M. haemolytica colonization and proliferation. This may explain why there was a reduction in colonization of M. haemolytica from days 3-7, while Lactobacillus was only greater in BT + Mh calves on d 5.

A single administration of BTs was performed to align with management strategies of calves, which are typically handled and processed only once early in life. As observed by qPCR, a transient increase in the number of Lactobacillus-specific 16S rRNA gene copies occurred, but only on d 5. This implies that the majority of strains in the BT mixture did not colonize the nasal cavity of calves for more than two days after inoculation. However, L buchneri 86D and L. amylovorus 72B were isolated from two calves on days 7 and 13, respectively, indicating successful colonization by these BT strains in these calves. The differences observed between strains and the variation in calf colonization are difficult to explain but may be due to several factors. Firstly, the method of delivery may have resulted in colonization variation as PBS suspensions rapidly dissipated upwards and downwards of the application site. Improved delivery methods my thus reduce variability. Secondly, the source of calves may also have resulted in 170

different microbiotas between calves and as a result, altered resistance to BT colonization. In

support of this, most of the variability observed in the composition of the nasal microbiota was

explained by differences among calves. This is not surprising as the calves originated from

different farms and it has previously been shown that the dam vaginal microbiota influences the

respiratory microbiota of calves (Lima et al., 2019). In addition, the presence of similar species to

those inoculated may limit colonization of probiotics. For example, indigenous Bifidobacterium

longum in the human gut was shown to prevent colonization of exogenous B. longum AH1206

strain (Maldonado-Gómez et al., 2016). Zmora et al. (2018) also observed that the successful

colonization of probiotic strains in the human gut was negatively affected by the presence and

abundance of indigenous members of the same genera to which the probiotic strains belonged.

Indeed, we detected Lactobacillus spp. in calves prior to inoculation via qPCR. Thus, enhanced

colonization may be achieved if the BTs are delivered soon after birth and before the establishment

of the respiratory microbiota. Finally, the BT strains originated from post-weaned feedlot beef

steers (Amat et al., 2019b). As the nasal microbiota can differ between pre-weaned dairy calves

(Gaeta et al., 2017) and post-weaned feedlot beef cattle (Holman et al., 2015b), it is possible that these BT strains were better adapted to beef cattle.

Despite colonization being transient, an impact of BT on the nasal microbial community was observed even after the majority of BT strains were no longer detectable in the NS. Although probiotics are frequently reported as poorly established members of the gut microbiota (Lee et al.,

2014; Eloe-Fadrosh et al., 2015), the effect of probiotics on the gut microbiota structure and diversity has been reported to be evident from 2 weeks up to 5 months after cessation of probiotic consumption (Zmora et al., 2018; Zhang et al., 2014; Faust and Raes, 2012). The underlying mechanisms by which the BT strains may modulate the nasal microbiota long-term is challenging 171

to explain. However, one of the mechanisms could be associated with the impact of introducing

BT strains on the microbial interaction network among members of the nasal microbiota.

Microbial communities in most niches form complex ecological interaction webs and such

interactions are important in maintaining microbiota homeostasis and symbiotic relationships

between microbes and host (Suez et al., 2018). Here, we observed significantly different structures

of causal relationships between the 10 most relatively abundant genera as evident by the path

models. The biological significance of these causal relationship structures observed remains to be

determined. However, it has been recently suggested that probiotic bacteria may promote intestinal

microbiota homeostasis by enhancing species-species interactions and increasing the number of

connecters and/or module hubs within the network (Yang et al., 2017). Hence, it is tempting to

speculate that BT inoculation may have also positively affected the microbial relationship and

thereby promoted microbial homeostasis, which may have contributed to reduced microbial

diversity and richness observed in the last week of study.

Path analysis models can help to predict potential mutualistic and antagonistic interactions

among bacterial species (Weng et al., 2017). In the present study, the relative abundance of

Mannheimia in Mh calves was predicted by path modeling to be negatively affected by

Lactococcus and time (direct), and Bifidobacteria and Lactobacillus (indirect). The fact that

Lactococcus could antagonize M. haemolytica is in agreement with a previous study (Timsit et al.,

2018) which showed that Lactococcus was more abundant in the lower respiratory tract of healthy feedlot cattle compared to cattle with BRD (which had a higher abundance of M. haemolytica).

Furthermore, we have previously shown that Lactococcus lactis could inhibit the growth of M. haemolytica in vitro (Amat et al., 2017). Lactococcus spp. were relatively abundant among all

calves (10.9%) on day -1. This indicates that species within Lactococcus may have a role in 172

maintaining respiratory homeostasis in the early life of calves. Surprisingly, in the BT + Mh group,

Mannheimia was found to be an exogenous variable whose relative abundance was not affected

by the other tested genera, including Lactobacillus. This finding is difficult to explain, and further

research should be conducted to better understand how the Lactobacillus strains that we inoculated

affected communication and microbial networks among nasal bacteria. It is possible that limiting

the path models to the 10 most relatively abundant genera excluded minor genera that could have

affected Mannheimia endogenously in the BT + Mh calves. Regardless, our study showed that a

single dose of BTs could affect the respiratory microbiota of calves, highlighting that BTs may be

an effective method to modulate the respiratory microbiota. Thus, further research on the method

and timing of administration, as well as additional BT species are warranted to enhance pathogen

resilience of the bovine respiratory microbiota.

To the authors’ best knowledge, this is the first study to profile the tracheal microbiota in

young dairy calves (< 5 weeks of age). Despite being removed from the dam and dairy farm within

3 h post birth, and housed in a clean and controlled environment, these calves harbored a diverse

(393 different genera identified) and rich tracheal microbial community. 16s RNA gene

sequencing of TTA samples collected on d 16 revealed that the tracheal microbiota contained more phyla than the nasal cavities (29 vs. 26) and only 3 out of 10 most abundant genera were shared between these two anatomical sites. This suggests that a self-sustaining and established microbiota

is present in the lower respiratory tract of newborn calves. Comparing the microbial profiles

between the BT + Mh and Mh groups, no significant differences were observed for both the beta

and alpha diversity indices. However, there were alterations in the relative abundance of specific

phyla and certain genera in the BT + Mh group relative to the Mh group.

173

The M. haemolytica challenge design used in the present study was a colonization model and not a disease model, explaining the absence of abnormal clinical signs during the study period.

To experimentally induce disease (i.e. pneumonia), M. haemolytica has to be inoculated directly into the lower respiratory tract (bypassing the upper respiratory tract) (Lhermie et al., 2016) or should be inoculated after a viral infection (i.e. bovine herpes virus 1) (Word et al., 2019; Jricho et al., 1986). Ten calves nevertheless had lung lesions and M. haemolytica was found to be more abundant in Mh versus BT + Mh group. Therefore, it is possible that some M. haemolytica reached the lower respiratory and created lesions, especially in the Mh calves. It is interesting to speculate that the BTs reduced colonization ofM. haemolytica in the upper respiratory tract of BT+Mh calves thereby reducing the amount of M. haemolytica translocating to the lungs. It is generally accepted that proliferation of BRD pathogens is a prerequisite to lung infection (Griffin et al., 2010). Indeed, some parenteral metaphylactic antibiotics given to feedlot cattle upon arrival at the feedlot have been shown to reduce the prevalence of M. haemolytica in the nasopharynx of calves, which may be related to their efficacy (Zaheer et al., 2013). Thus should BTs reduce proliferation or the abundance of M. haemolytica in the upper respiratory tract, they may also limit translocation of the pathogen to the lungs. A next logical step would be to investigate if the nasal inoculation with

BTs can prevent the development of BRD due to M. haemolytica.

In summary, a single dose inoculation of intranasal BTs developed from six Lactobacillus strains originating from the bovine respiratory microbiota was able to reduce nasal colonization by M. haemolytica in experimentally challenged dairy calves. Administration of BTs also altered the recursive structure of causal relationships among the 10 most relatively abundant genera within the nasal microbiota of calves. A lower relative abundance of M. haemolytica in the trachea of calves that received BT was observed at the end of the study period. Finally, the administration 174

of BT did not stimulate a systemic immune response based on the cytokines tested. Overall, the results of this study, for the first time, demonstrated the potential use of intranasal BTs to mitigate the BRD pathogen M. haemolytica in cattle. Further research should be conducted to investigate if intranasal BT inoculation can prevent respiratory disease, which is highly relevant to the cattle industry.

175

5.5 Tables and Figures

Estimated mean bacterial counts (Natural log CFU per nasal Independent variable and level a,b swab) (SE) P-value Treatment Mh (n=12) Ref. BT+ Mh (n=12) -3.31 (1.36) 0.024 Day of sampling -0.25 (0.08) 0.003 Treatment × time interaction 0.26 (0.11) 0.025 Replicate# 2.38 (0.84) 0.010 a Intercept = 7.93 ± 1.049 (P < 0.001)

b Random effect = calf# (n = 24)

176

Table 5.2 Path models fitted to data corresponding to statistics that are Schwarz’s (Bayesian) information criteria of three path models fitting subsets of experimental data.

Unmodified model: Modified model 1 (Mh Modified model 2 (BT (Time, Lactobacillus) group): Time → + Mh group): → (Mannheimia, (Bacteroides, Klebsiella → Moraxella, Prevotella, (Lactobacillus, Acinetobacter, Mannheimia, Acinetobacter, Bifidobacterium, Acinetobacter), Bacteroides, Streptococcus, Lactobacillus → Lactococcus) Lactobacillus, Bifidobacterium, Time → (Lactococcus, Prevotella, Bacteroides, (Klebsiella, Klebsiella, Bacteroides, Klebsiella) Bifidobacterium) → Bifidobacterium), Acinetobacter, Lactobacillus → Lactococcus → (Bifidobacterium, Mannheimia Streptococcus), Acinetobacter → Bifidobacterium → Streptococcus, Bacteroides, Streptococcus → Acinetobacter → (Bacteroides, Bifidobacterium, Prevotella), Mannheimia → (Bifidobacterium, Streptococcus, Prevotella) → Bacteroides → Lactococcus, Prevotella, Acinetobacter ↔ Lactococcus ↔ Moraxella Prevotella, Bacteroides ; 0 = {(Lactobacillus, ↔ Prevotella, Moraxella, Klebsiella) Lactobacillus ↔ ↔ Time, Acinetobacter; 0 = Lactobacillus ↔ {Mannheimia ↔ Moraxella} (Time, Moraxella, Lactobacillus), Moraxella ↔ (Time, Lactobacillus), Lactobacillus ↔ Time} Mh 281.1053 154.4841 235.4901 BT + Mh 283.2544 220.6100 163.6963

Model fit statistics (model 1 to Mh group data, model 2 to BT + Mh group data) Iterations 7 9 Chi-square = 32.6909, =38.3309, p 2= 0.7519 p 2= 0.4545 RMSEA <𝜒𝜒39 .0001 <𝜒𝜒38 .0001 Bentler-Bonett NFI 0.8863 0.8724 Bentler-Bonnett non-NFI 1.0383 0.9980 Stability coefficient of reciprocal causation 0 0

177

Figure 5.1 Study design and sampling regimen.

178

Figure 5.2 Bacterial counts of M. haemolytica and abundance of Lactobacillus spp.

Counts of M. haemolytica were determined by culturing (A), and estimation of total Lactobacillus spp. quantity was determined by qPCR of Lactobacillus 16S rRNA gene copies (B) from nasal swabs of calves that received an intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT + Mh; n = 12) or only M. haemolytica (Mh; n = 12). Each box indicates the interquartile range

(IQR) (middle 50% of the data), the middle line represents the median value, and the whiskers represents 1.5 times the IQR Coloured dots indicate outliers. * Significant difference between treatments (P < 0.05).

179

Figure 5.3 Beta diversiry of nasal microbiota.

Principal coordinates analysis (PCA) plots of the Bray-Curtis dissimilarities by treatment and sampling day for the nasal microbiota of calves that received an intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT + Mh; n = 12) or only M. haemolytica (Mh; n =

12). The percentages of variation explained by the principal coordinates are indicated on the axes.

180

181

Figure 5.4 Relative abundance of the 5 most relatively abundant phyla and 10 most relatively abundant genera in the nasal microbiota

of calves that received an intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT + Mh; n = 12) or only M.

haemolytica (Mh; n = 12).

(A) Comparison between Mh and BT + Mh calves at the phyla level; (B) comparison of phyla by groups and sampling day; (C)

comparison before (d -1) and post BT and Mh inoculation (average from d 3 to 16); (D) comparison of genera by groups and sampling day.

182

Each box indicates the interquartile range (IQR) (middle 50% of the data), the middle line represents the median value, and the whiskers

represents 1.5 times the IQR Coloured dots indicate outliers. Significant difference between treatments at P < 0.05 (*) and P < 0.01

(**). 183

184

Variances of measured variables (relative abundances of genera and time) with standard errors and P values are shown in squares. Causal

relationships that are implied by the model are shown as solid lines with arrows that indicate the direction of causation. Causal paths are

labelled with the standardized path coefficients, standard errors, and P values. Covariance terms are shown in ovals with dashed arrows

between the variables that co-vary. Green line represents the positive effect, whereas the red line represents the negative effect. The

thickness of the solid line represents the strength of the effect.

185

Figure 5.7 Description of the tracheal microbiota of calves that received an intranasal inoculation of either bacterial therapeutics and

M. haemolytica (BT + Mh) or only M. haemolytica (Mh).

186

Principal coordinates analysis plots of the Bray-Curtis dissimilarities (A); percentages of variation explained by the principal coordinates are indicated on the axes), boxplots of observed OTUs (B), Shannon diversity index (C), the six most relatively abundant phyla (D),

and the 10 most relatively abundant genera (E).

187

Figure 5.8 Serum cytokine concentrations (IL-6, IL-8 and IL-10) by sampling day in dairy calves that received an intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT +

Mh; n = 12), or only M. haemolytica (Mh; n = 12).

188

Chapter Six: Intranasal administration of bacterial therapeutics induces longitudinal

modulation of the nasopharyngeal microbiota in post-weaned beef calves

Chapter 6 is being submitted to Microbiome.

Contributing authors: Amat S, Timsit E, Workentine M, Schwinghamer T, van der Meer F,

Alexander TW.

189

6.1 Introduction

Bovine respiratory disease (BRD) is the most significant health condition affecting beef

calves, and accounts for economic losses due to costs associated with treatment and prevention,

and reduced productivity (Cernicchiaro et al., 2013; Johnson et al., 2017). In North America, the

management of beef cattle typically involves their shipment from pastures to feedlots for

production. During feedlot placement, cattle are most susceptible to BRD, with the majority of

cases occurring within the first 60 days of feedlot placement (Booker et al., 2008). Primary viral

infections or stressors, that include weaning, shipping, and co-mingling with new pen mates, are

proposed to reduce host immunity during transition to feedlots (Taylor et al., 2010). Consequently,

opportunistic bacterial pathogens residing in the upper respiratory tract proliferate and translocate

to the lungs causing bronchopneumonia (Rice et al., 2007). The main BRD-associated bacteria

include Mannheimia haemolytica, Pasteurella multocida, Histphilus somni, and Mycoplasma

bovis (Confer, 2009). As a result of increased BRD susceptibility, commercial feedlots rely on antimicrobial-driven approaches to prevent BRD infections in cattle (Cameron and McAllister,

2016).

Long-acting injectable antimicrobials are commonly administered to cattle entering

feedlots (i.e. metaphylaxis) for BRD prevention (Nickell and White, 2012). For example, the

macrolide tulathromycin was used at feedlot entry by 45.3% of feedlots in the United states for

BRD mitigation (USDA, 2013). Metaphylactic antimicrobials treat lung infections that may be

prevalent in calves entering feedlots and also prevent infection during the course of their

bioactivity. It is also likely that metaphylactic antimicrobials reduce prevalence and proliferation

of BRD pathogens in the upper respiratory tract, a prerequisite to lung translocation (Zaheer et al.,

2013). However, antimicrobial resistance in BRD pathogens has increased over the last ten years

190

(Alexander et al., 2013; Snyder et al., 2018; Anholt et al., 2018) with resistance to tulathromycin

recently being detected in more than 70% of M. haemolytica and P. multocida isolated from feedlot calves (Timsit et al., 2017). In addition, resistance elements have been detected in mobile elements from the BRD-associated Pasteurellaceae family, conferring multi-drug resistance to antimicrobials used for both prevention and treatment of BRD in feedlot cattle (Klima et al., 2014;

Klima et al., 2019). Antimicrobial resistance in BRD pathogens therefore threatens the efficacy of currently used antimicrobials in beef cattle. Alternatives to antimicrobials are therefore needed for use in novel feedlot management strategies.

The respiratory microbiota contribute to host health by providing colonization resistance against pathogens and maintaining homeostasis (Man et al., 2017). It is proposed that disruption of the bovine respiratory microbiota can promote the proliferation of BRD pathogens (Zeineldin et al., 2019). Indeed, several management factors including transportation to a feedlot (Holman et al., 2017), diet composition (Hall et al., 2017), and antimicrobial administration (Holman et al.,

2019) alter the upper respiratory tract microbiota of cattle. Recently, we observed that several LAB were inversely correlated with Pasteurellaceae in the nasopharynx of cattle transported to an

auction market and subsequently a feedlot (Amat et al. 2019a). Genera within the LAB order

Lactobacillales have been shown to be reduced in cattle that develop BRD (Holman et al., 2015;

Timsit et al., 2018). Additionally, isolates of LAB originating from the bovine respiratory tract

have been shown to directly inhibit BRD pathogens (Corbeil et al., 1985; Amat et al., 2019a).

These data support that LAB are important community members of the bovine respiratory tract and may be integral to providing colonization resistance against BRD pathogens. Consequently, we have previously developed BTscomprised of six Lactobacillus strains that were characterized

for their in vitro inhibition and exclusion of M. haemolytica, and adherence to and 191

immunomodulation of bovine turbinate cells (Amat et al 2019b). The objective of the current

study was to evaluate the longitudinal effects of these BTs on the nasopharyngeal microbiota of

beef calves, after intranasal administration. The effect of the BTs were also compared to those of

tulathromycin administered subcutaneously.

6.2 Materials and Methods

6.2.1 Animals and experimental design

Animals used in this study were cared for in agreement with the Canadian Council for

Animal Care guidelines (Olfert et al., 1993). All the procedures and protocols with respect to animal handling and sampling were reviewed and approved by the Animal Care Committee at the

Lethbridge Research and Development Centre, Agriculture and Agri-Food Canada (Lethbridge,

Alberta, Canada).

Sixty crossbred beef heifers approximately 6 months-old (initial BW = 266 ± 13 kg) and

originating from a single cow-calf ranch, were purchased from a local auction market and

transported to the Lethbridge Research and Development Centre feedlot (<10 km distance). Upon

arrival, the calves were weighed and nasopharyngeal (NP) swabs were collected (d -1). Calves

were then blocked by weight and randomly assigned to three treatment groups (n = 20 per treatment): i) BT group received an intranasal cocktail of six Lactobacillus strains suspended in phosphate buffered saline (PBS) in equal concentrations (3 × 109 CFU per nostril), ii)

metaphylaxis (MP) group received a subcutaneous injection of tulathromycin (2.5 mg/kg BW),

and iii) control (CTRL) group received intranasally PBS without bacteria. The calves were housed

individually in pens throughout the study. and were fed once daily with a diet containing 75%

192

barley silage, 22.5% dry roll barley, and 2.5% standard feedlot supplement. Calves had free access

to drinking water.

6.2.2 Preparation of BT inoculum

The BT cocktail was a mixture of six Lactobacillus strains (1×109 CFU mL-1): L. amylovorus (isolate 72B), L. buchneri (63A and 86D), L. curvatus (103C) and L. paracasei (3E

and 57A). These isolates were inoculated on Lactobacillus De Man, Rogosa and Sharpe (MRS)

agar (Dalynn Biologicals, Calgary, AB, Canada) and incubated for 48 h at 37°C in 10% CO2. One

day prior to nasal inoculation, a single colony of each strain was inoculated into 5 mL MRS broth

and incubated at 37°C with agitation at 200 rpm. After 18 h of incubation, each bacterial culture

was centrifuged at 7,600 × g for 10 min, the supernatant discarded and the pellet re-suspended

with pre-warmed (37°C) PBS to achieve a target concentration of 1 ×109 CFU per mL-1 using pre- established OD600 values. The BT cocktail was prepared by mixing the six Lactobacillus isolates

at equal ratios in PBS. One hour prior to inoculation, 3 mL of the BT cocktail was loaded in a

sterile 10 mL syringe and the syringe tip was covered with sterile needle to prevent any leakage.

For the control group, 3 mL PBS was loaded similarly into a syringe without bacteria.

6.2.3 Administration of BT cocktail and tulathromycin

On day 1, calves were restrained in a squeeze chute and administered treatments. Sterile

laryngo-tracheal mucosal atomization devices (LMA® MADgic® Laryngo-Tracheal Mucosal

Atomization Device without syringe, Cat# MAD 700, Teleflex, Morrisville, NC) were fitted to

loaded syringes, and the atomization device was inserted into each nostril of calves (approximately

15 cm) and sprayed until the syringe was empty. One atomization device was used for the two

nostrils of each calf. A total of 6 mL of BT inoculum (3 mL per nasal cavity) was administrated

193

to calves in the BT treatment group. For control group animals, PBS was sprayed into the nostrils

similar to the BT inoculum (3 ml per nostril). The metaphylaxis group received a single

subutaneous injection of long-acting tulathromycin (2.5 mg/kg body weight).

6.2.4 Nasopharyngeal swab sampling and processing

In addition to sampling at feedlot arrival (d -1), NP samples were collected from the right

nostril of each calf in the study on days 1, 2, 4, 7, 14, 28, and 42. The NP sampling procedures

were described previously (Holman et al., 2017). Prior to sampling, the right nostril was wiped clean with 70% ethanol. Extended guarded swabs (27 cm) with a rayon bud (MW 124, Medical

Wire & Equipment, Corsham, England) were used for sampling. Swabs were taken while the animals were restrained in a squeeze chute. Swab tips were then cut and placed in a sterile 1.5 mL tube on ice. Samples were transported to the lab and processed within one hour of collection. At

the lab, the swab tip was transferred into a cryovial containing 1 mL brain heart infusion (BHI)

with 20% glycerol and vortexed.

Isolation and detection of bovine respiratory pathogens

Aliquots of swab suspension was plated for isolation and detection of BRD-associated pathogens including M. haemolytica, P. multocida, and H. somni. Culturing, isolation, and PCR identification of the pathogenic isolates were described previously (Holman et al., 2017; Amat et al., 2019a). The remaining swab suspensions in BHI glycerol stock were stored at -80°C for DNA extraction.

194

6.2.5 Genomic DNA extraction, 16S rRNA gene sequencing and analysis

The genomic DNA was extracted from the swab suspension according to the methods of

Holman et al., (2017 and 2019). From the extracted DNA, the V4 region of the 16S rRNA gene

was amplified using primers 515-F (5′-GTGYCAGCMGCCGCGGTAA-′3) and 806-R (5′-

GGACTACNVGGGTWTCTAAT-′3) (Holman et al., 2019). The amplicon was sequenced on a

MiSeq instrument (Illumina, San Diego, CA, USA) with the MiSeq Reagent Kit v2.

After quality check with FastQC 0.11.5 and MultiQC 1.0 (Ewels et al., 2016) primers and

low quality sequences were trimmed off the raw sequence reads using cutadapt 1.14 (Martin,

2011). The trimmed reads were used to construct amplicon sequence variants (ASVs) using dada2

1.10.0 (Callahan et al., 2016) in R 3.5.1 (R Core Team, 2018). Unless otherwise stated, all dada2 functions were used with default parameters. Reads were first filtered with dada2::filterAndTrim with a max expected error of 1. Error rates were learned for the forward and reverse reads separately and these error rates were used to infer exact sequences (error correct) for each sample from dereplicated, trimmed reads using pooled=TRUE for the dada2::dada. Following this, the forward and reverse reads were merged using dada2::mergePairs. Chimeras were removed with dada2::removeBimeraDenovo and taxonomy was assigned using the naïve Bayesian classifer

(Wang et al., 2007) as implemented in dada2::assignTaxonomy trained with the Silva training set

version 132 (https://doi.org/10.5281/zenodo.1172782). Species level assignment was done with

dada2::addSpecies which uses exact matching to assign species where possible. ASVs were

aligned with ssu-align 0.1.1 (Nawrocki, 2009). Quantification of bacterial, Lactobacillus spp., and

antibiotic resistance determinants using quantitative PCR

195

Real-time PCR was used to quantify copies of 16Sr RNA genes in DNA from nasal swabs

that were specific to all bacteria, or Lactobacillus, to estimate the abundances of total bacteria and

Lactobacillus genera, respectively. Total 16S rRNA gene copies were amplified using primers

515F and 806R described above. Lactobacillus-specific 16S rRNA gene copies were amplified

using a genus-specific primer described previously (Heilig et al., 2002). In addition to 16s rRNA

genes, the macrolide resistance gene msr(E) and tetracycline resistance gene tet(H) were

quantified from the DNA extracted from nasal swabs. The primers for msr(E) and tet(H) were

reported previously by Klima et al., (2014) and Zhu et al., (2013), respectively.

To generate standards for PCR, amplicons of each target gene were cloned into competent

E.coli cells using the TOPO Cloning Reaction Kit (Invitrogen) according to the instructions of the manufacturer. Plasmids containing amplicon inserts were purified by the QIAprep spin miniprep kit (Qiagen, Hilden, Germany) and then serially diluted. Each real-time PCR mixture (25 μL) contained 1X iQ SYBR Green Supermix (Bio-Rad Laboratories Inc.), 0.4 μM of each primer, 0.1

μg/μL BSA (New England Biolabs,180 Pickering, ON, Canada), and 25 ng of DNA. For each PCR reaction, the amount of DNA extracted from the NP swabs was normalized to 10 ng/µL. The quantification of target genes were performed on a CFX96 Touch Real-Time PCR Detection system (Bio-Rad Laboratories Inc.) with the following conditions: an initial denaturation at 95°C for 3 min, followed by 40 cycles at 95°C for 25 sec, 50°C for 30 sec, and then 72°C for 45 s. For quantification of total and Lactobacillus-specific 16S rRNA gene copies and the resistant gene copies, standards were prepared for each gene using the respective p-Drive plasmid containing inserted amplicons and concentrations of 106, 105, 104, 103, and 102 copies per reaction (in duplicate). Melt curve analyses were performed on all PCR reactions to ensure specific

196

amplification. The temperature range was 60°C to 95°C and fluorescence was measured at 0.5°C

intervals.

6.2.6 Statistical analysis

Statistical analysis of sequence data was done with R 3.6.0 with phyloseq 1.28.0

(McMurdie et al., 2013), and vegan 2.5.5 (Oksanen et al., 2017). Plots were created with ggplot2

3.1.1. Sequences matching mitochondria or chloroplast were removed along with any sequences

that were not assigned to Bacteria. A filtered copy of the ASV sequence table was created that

retained ASVs present (count ≥ 2) in at least 1% of the samples. This served to reduce noise for

downstream analysis. The full version of the sequence table was used for alpha diversity, which was assessed with the Shannon diversity index using 0.3.1(Willis and Martin, 2018). Richness was

estimated with a Poisson model using breakaway 4.6.8 (Willis and Bunge, 2015). To calculate beta

diversity the ASV counts (filtered table) were normalized by with a variance stabilizing transform

(using DESeq2 1.24.0) (Love et al., 2014) with size factors calculated using GMPR (Chen et al.,

2018) then sample-sample distances were determined with the Bray-Curtis metric and visualized

with detrended correspondence analysis (DCA). Permutational multivariate analysis of variance

(PERMANOVA) with 10,000 permutations was used to determine the effect of treatment, and

sampling time on the microbial community structure. For statistical testing the model included an

interaction between day and treatment to detect differences between treatment groups over time.

The betta function from breakaway 4.6.8 (Willis et al., 2017), which models both observed and

unobserved diversity, was used to test alpha diversity. Corncob 0.1.0 (Martin et al., 2019) was

used to test for differentially abundant taxa by fitting a model with and without the interaction

term. This identified taxa that showed a change from baseline in the BT and MP groups differed

197

significantly from the CTRL group. Corncob uses a beta-binomial-based model that controls for correlated observations (taxa) and overdispersion.

Generalized linear models (SAS PROC GLIMMIX) were built on the prevalence data of

M. haemolytica, P. multocida and H. somni determined by culturing. The Bernoulli/binary distribution was selected for the models. Separate F-tests of the treatment effect were produced, using the SLICE option in the LSMEANS statement, for the respective days of observations." The relative abundance of phyla and qPCR results data was analyzed using the GLIMMIX procedure in SAS (SAS 9.4, SAS Institute Inc., Cary, NC). The individual calf, treatment, and time were included in the CLASS statement. The models were “generalized” due to the specification of response distributions that were not Gaussian normal. Models were “mixed” due to the inclusion of fixed effects (treatment-nested-in-time and time) and random effects (individual). Variance heterogeneity was modeled using a “RANDOM _RESIDUAL_ / GROUP = Treatment*Time” statement. Response distributions and structures of the variance-covariance matrix were selected for each genus based on the model fit statistics, i.e., the Bayesian information criterion (BIC).

Preliminary models that specified the beta-binomial distribution did not converge. Therefore, alternative distributions were tested: Gamma, inverse Gaussian, lognormal, shifted t, Gaussian normal, exponential, and geometric.

Ecological network modeling was performed to evaluate the directed microbial interactions among the nasopharyngeal microbial communities using BEEM-static in R. Briefly, based on the abundance profile of all observed genera derived from 8 sampling time points, generalised Lotka-

Volterra models (gLVMs) coupling biomass estimation and model inference in an expectation maximization-like algorithm (BEEM) was used to construct three (CTRL, BT, MP groups) 198

ecological network models as described by Li et al. (2017). The interaction network inferred by

BEEM-statistic was visualized by the plots generated using graph package of R.

Path analysis was performed to model the interrelationships of selected bacterial species within the NP microbial communities. Path analysis is a member of the structural equation modeling tools that enables the identification of causal relationships between measured variables

(Schwinghamer et al., 2017). The construction of an initial path model is based on theories regarding the causal relationships. These path diagrams are used to illustrate the strength and direction of the causal relationships between variables. To identify biological interactions within the NP microbiota, and the changes in biological interactions in response to the BT and MP treatment, 16 genera were selected: 12 that exhibited the greatest change in relative abundance in

BT and MP groups relative to CTRL group over the course of study, and four BRD-associated genera (Mannheimia, Pasteurella, Histophilus, and Mycoplasma). The relative abundance data for these 16 genera were sorted into three subsets based on the treatment groups (BT, CTRL and MP).

The methods of path modeling followed those of Schwinghamer et al. (2017). The CORR procedure in SAS (SAS 9.4, SAS Institute Inc., Cary, NC, USA) was used to calculate the matrices of Spearman rank-based correlation coefficients that were used as input for path modeling with

SAS PROC CALIS. The initial hypothetical model was:

𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝑢𝑢𝑚𝑚 𝐷𝐷𝐷𝐷𝐷𝐷 _ 005 � � → � ⋮ � 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 Three modified path models were developed𝑅𝑅 𝑅𝑅𝑅𝑅𝑅𝑅based𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 on the addition𝑈𝑈𝑈𝑈𝑈𝑈 and subtraction of paths and covariance terms, based on Wald statistics and Lagrange multiplier statistics that were calculated using the MODIFICATION option in the PROC CALIS statement. Modifications were selected

199

based on lower (better) values of Schwarz's BIC and a value equal to zero for the stability criterion

of reciprocal causation.

6.3 Results

6.3.1 Calf health and weight gain

The calves were monitored daily for clinical signs of BRD throughout the study. Elevated

rectal temperature (≥ 39.7 °C) was detected in five calves (three from CTRL and two from BT;

Supplementary Table S6.1) that recovered in response to a single injection of the antibiotic micotil

(tilmicosin). The remainder of the experimental calves were healthy during the course of study.

The calves were weighed first at arrival (d-1) and subsequently on a bi-weekly basis for the 42

days of study. The average daily gain was not different among treatments (P = 0.506)

(Supplementary Figure S6.1).

6.3.2 Prevalence of BRD-associated pathogens determined by NP swab culturing

Presence of M. haemolytica, P. multocida, and H. somni was evaluated by culturing the NP

swabs collected during the first 28 days of study (Figure 6.1). Overall, M. haemolytica had the

highest prevalence of the three pathogens on d-1 (24-40% across treatments). No significant difference was detected among treatment groups (P > 0.05) at any sampling time point. However, there was a high tendency (P = 0.06) for reduced M. haemolytica prevalence in MP-treated calves on d 7. For P. multocida, prevalence in CTRL (10-40%) and BT (5-36%) calves was similar (P >

0.05). In MP calves, prevalence of P. multocida was lower (0-5%), and it was significantly lower on days 7 and 14 compared to CTRL and BT calves (P < 0.05). Overall, the prevalence of H. somni remained low throughout the study for all treatment groups and was not different between treatment groups at any sampling point (P > 0.05). Only the MP group had colonization rates of

200

0% for P. multocida (days 1, 4, and 14) and M. haemolytica (d 1), which occurred after metaphylactic treatment.

6.3.3 Total bacteria and Lactobacillus in NP swabs determined by qPCR

The abundances of total bacteria and Lactobacillus in nasal swabs was estimated by quantifying the gene copy numbers of general and Lactobacillus-specific 16S rRNA (Figure 6.2).

An interaction between treatment × time affected the total bacterial number (P = 0.02; Figure 6.

2A). Compared to BT and CTRL calves, tulathromycin injection reduced bacteria in NP swabs on days 4, 7, and 42 (P < 0.05). Total bacteria in NP swabs of BT and CTRL calves were not different over the course of the study (P > 0.05).

The estimated number of Lactobacillus in NP swabs was affected by the interaction of treatment × time (P = 0.01). The mean total Lactobacillus 16S rRNA copy number per swab from the BT group increased (p < 0.01) from d -1 to d 1 (24 h post-BT inoculation) (Figure 6.2B), and

then decreased to similar levels in CTRL and MP calves by d 2 (P > 0.05). On d 42, Lactobacillus

was reduced in the MP group compared to BT and CTRL calves (P < 0.0001).

6.3.4 Structure and composition of the NP microbiota

16S rRNA gene sequencing overview

The raw SV table contained 10,400 SVs with a total of 7,161,751 reads assigned to 476 samples. The median number of sequences per sample was 15,037 ± 3,482.21 with a minimum of

0 and maximum of 26,536. After filtering, the SV table contained 531 SVs with a total of 6,349,998 reads. The median number of sequences per sample was 13,613 ± 4,100.6 with a minimum of 22 and maximum of 26,027.

The community structure of the NP microbiota

201

PERMANOVA revealed that an interaction of treatment × time had an effect (R2 = 0.04, P

= 0.001) on the microbial structure of the NP microbiota (Figure 6.3). However, time had a larger

effect on microbial structure (R2 = 0.141, P = 0.001) compared to treatment (R2 = 0.034, P = 0.001).

As indicated by the DCA plots (Figure 6.3) the microbiota tended to cluster by early (d -1 to 2),

mid (d 4-7) and late (d14-28) time points. Clustering according to treatment was most evident on

d 28 and 42.

Alpha-diversity, as assessed by richness and Shannon diversity index, revealed that both

indices were affected by a treatment × time interaction (P < 0.05) (Figure 6.4). After BT

inoculation, the NP microbiota of BT calves had reduced richness throughout the study, compared

to CTRL and MP calves (P ≤ 0.014), except for on d 7, when richness was similar in BT and CTRL

groups (P = 0.213). In contrast, MP calves had increased richness on days 7-42 when compared

to BT calves, and days 7 and 14, when compared to CTRL calves (P ≤ 0.001). Similar to richness, the Shannon diversity index was reduced in BT calves from days 7-42, compared to CTRL and

MP treatments (P ≤ 0.0001). The Shannon diversity was greater in MP calves compared to BT and

CTRL calves, but only on days 1 and 7 (P ≤ 0.016).

Composition of the NP microbiota

Across time and treatment groups, a total of 14 different bacterial phyla were identified, among which Proteobacteria (36.4 %) Tenericutes (22.2 %), Firmicutes (17.4 %), Actinobacteria

(12.9 %), and Bacteroidetes (9.9 %) were the most relatively abundant, and together constituted

98.8% of the sequences. The diversity of genera within each phylum varied with the relative abundance of a single genus ranging from <1% to 100% of a phylum. Overall, the ten most relatively abundant genera across treatments and time included Mycoplasma (22.2%), Moraxella

(18.8%), Pasteurella (3.7%), Mannheimia (3.6%), Corynebacterium_1 (2.7%), 202

Ruminococcaceae_UCG-005 (2.4%), Psychrobacter (2.3%), Jeotgalicoccus (2.1%), Histophilus

(1.1%) and Planococcus (1.1%) (data not shown).

6.3.5 Changes in microbial composition following BT and tulathromycin treatment

Changes in the five most relatively abundant phyla

Noticeable changes in NP microbial composition at phylum level were observed in

response to treatment and time effects (Figure 6.5). The relative abundance of Proteobacteria in

CTRL calves varied over the 42 days of study, with a gradual increase in the first 7 days followed

by a decline in the remaining 5 weeks of the study. In BT calves, the relative abundance of

Proteobacteria was similar to CTRL calves, except for on d 28, when it was increased (P = 0.02).

In MP calves, however, Proteobacteria had a lower relative abundance compared to CTRL calves

on days 4, 7 and 14 (P ≤ 0.008). The relative abundance of Firmicutus did not differ between BT

and CTRL groups at any sampling time (P > 0.05). However, Firmicutes was increased in MP

calves within the first 14 days antibiotic injection compared to CTRL calves (P < 0.05). Similarly,

Bacteroidetes became significantly enriched in MP calves on days 7 and 14 relative to the CTRL

group (P ≤ 0.022). Compared to CTRL calves, the abundance of Actinobacteia was reduced in

BT calves on d 42 (P = 0.011), while it was increased in MP calves on d 1 (P = 0.038). The relative abundance of Tenericutes was not affected by treatment (P > 0.05).

Changes in lactic acid-producing bacteria at family level

The LAB taxonomically belong to the Lactobacillales order (Mattarelli et al., 2014). Five different LAB families including Aerococcaceae, Carnobacteriaceae, Enterococcacease,

Lactobacillaceae and Streptococcaceae were detected in the present study (Supplementary Figure

203

6.2). Overall, no treatment effects on the abundance of these families were observed, although

the relative abundance of Lactobacillaceae was >10% in four calves from the BT group.

Changes in microbial composition at genus level

Beta-binomial regression analysis allowed for identification of compositional differences of NP genera between treatment groups, across time. In total, we identified 28 genera within BT

and MP groups whose change in relative abundance from d-1 (baseline) was significantly different

from changes that occurred in the CTRL group (P < 0.05). As shown in Figure 6. 6, 4 of the 10 most abundant genera (Ruminococcaceae_UCG-005, Psychrobacter, Jeotgalicoccus and

Planococcus) were included among these taxa that differed between treatment groups. In general, most of the 28 taxa became less abundant in the BT group, following BT inoculation. Among which, Ruminococcaceae_NK4A214_group, Paeniclostridium,

Lachnospiraceae_NK4A136_group, and Cellvibrio experienced a consistent decline during the entire post-BT inoculation period. For some taxa, the magnitude of change in relative abundance from the baseline varied among sampling times. For BT calves, the most significant reduction from baseline (d-1) was observed for Ruminococcaceae_NK4A214_group (d 2),

Lachnospiraceae_NK4A136_group (d 1), Facklamia (d 28), Celllvibrio (d 2) and Acetitomaculum

(d 4). Lactobacillus was the only taxa in the BT group that experienced an increase in relative abundance that was >1, occurring on d 1. In contrast to BT group, most of the 28 genera became enriched in the MP calves, with several taxa having increases in relative abundances which were

>2. The most immediate and consistent enrichment following tulathromcyin injection was observed for Jeotgalibaca.

Changes in relative abundance of Lactobacillus spp.

204

To identify whether the significant enrichment of genus Lactobacillus observed on d 1

(Figure 6.2B) was due to the inoculated BT Lactobacillus strains, species-level taxonomic

identification was performed on sequencing data from the Lactobacillus genus (Supplementary

Figure S6.3). Within Lactobacillus, 6 different taxonomic OTUs classified as L. acetotlerans/fructivorans, L.acidophilud/amylovorus, L. amylovorus/buchneri, L. curvatus/graminis, L. fermentum/mucosae, and L. ruminis were identified. Overall, the abundance of these OTUs varied however L. acidophilus/amylovorus, L. amylovorus/buchneri, and L. curvatus/graminis were generally only detected up to 48 h post-BT administration. This would suggest that the BT strains within the species L. amylovorus, L. buchneri, and L. curvatus accounted for the increase in these OTUs on days 1 and 2.

Changes in relative abundance of BRD-associated genera

The relative abundances of Mannheimia, Pasteurella, Histophilus and Mycoplasma which encompass BRD-associated pathogens, were not affected by treatment (P > 0.05; Figure S6.4).

6.3.6 Microbial interactions and dynamics of the NP microbiota

Interaction network structure among all observed OTUs

To evaluate overall dynamics of microbial communities, ecological modeling was used to analyse the interaction of all genera. As shown in network plots (Figure 6.7), distinct microbial interaction network structures were observed between CTRL, BT and MP groups. Compared to the CTRL group, the interaction network of microbiota from BT calves was more complex with a greater number of genera-genera interactions. In contrast, there was large decrease in genera- genera interactions among the microbial community of calves in the MP group, with only 22 genera identified in the network model. Even among these 22 genera, the interactions were

205

connected by two completely separate hubs, indicating that the tulathromycin injection diminished the interaction network among the NP microbial community.

Structure of causal relationships among 16 selected genera

Based on observing the distinct changes in overall microbial network structures of NP microbiota in response to both BT and MP treatments, we further evaluated the relationship of 16 targeted genera using causal structure-based path modeling. This path model analysis provides greater detail on the interaction between species, as it predicts causality and direction of the causality, and accounts for the time effect and unmeasured effects. Three path models (CTRL,

BT, and MP) are described in Table 6.1 and are depicted as diagrams of causal relationships of the relative abundances in Figure 6.8.

Details on model fit statistics and the number of iterations performed by the CALIS procedure are shown in Table 1. The p-values for the chi-square statistic for the modified path models were > 0.05, indicating good model fitting. The root mean square error of approximation

(RMSEA) was < 0.05, and the 90% confidence limits of the RMSEA were also <0.05. Therefore, the null hypothesis that the modified path models closely fit the data was retained. The values of the Tucker-Lewis index were also > 0.9, where the value of a true model would be expected to be

The variables that constituted the subsets of exogenous (independent) and endogenous

(dependent) variables varied in the three path models (Table 6.1). Day was initially hypothesized to be exogenous, and no path or covariance was added to change the position in the model of Day from being an independent explanatory variable. The process of model modification positioned the abundance of Histophilus as an exogenous variable in the modified path models, and it was a

206

predictor variable with a statistically significant standardized total effect on Alloprevotella

( , = 0.18 ± 0.07, P = 0.005) in Model 1 (of the BT group) (Table S6.2), while it was not a

𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 predictor𝑏𝑏� variable in Model 2 (CTRL) or Model 3 (MP) (Figure 6.8).

The relative abundances of seven genera (Acetitomaculum, Atopostipes, Jeotgalibaca,

Lactobacillus, Mannheimia, Mycoplasma, and Psychrobacter) were explained in the three

modified path models. In Model 1 (BT group), Pasteurella was exogenous and therefore the

sources of variance of Pasteurella abundance were outside of the model. In Model 2 (CTRL

group), however, the variance of Pasteurella was explained by predictor variables: Alloprevotella,

Atopostipes, Christenseneelaceae_R7_group, Lactobacillus, Pseudomonas, Psychrobacter,

Rikenellacease_RC9_gut_group, Ruminococcaceae_UCG005, Histophilus, and Day.

Robust relationships

Path relationships that were statistically significant in the three modified models were considered to be robust. As such, the negative and statistically significant (all p <.0001) values of the standardized direct effect from Day to Atopostipes in the modified path models represent a robust causal relationship. Therefore, under the conditions of this experiment, the relative abundance of Atopostipes is expected to decrease monotonically over time. In contrast, the standardized direct effect from Day to Psychrobacter was positive in the modified path models, and therefore the relative abundance of Psychrobacter is expected to increase monotonically over time, under conditions similar to the experiment. Robust causal relationships were also identified between observed genera. For example, the standardized positive and statistically significant (P

<.0001) total effects from Psychrobacter to Jeotgalibaca ( , = 0.50 ± 0.06; , = 0.44 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐽𝐽𝐽𝐽𝐽𝐽𝐽𝐽 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐽𝐽𝐽𝐽𝐽𝐽𝐽𝐽 𝑏𝑏� 𝑏𝑏� ± 0.07; , = 0.39 ± 0.07) (Table S6.2); and from Ruminococcaceae_UCG005 to 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 3 𝑏𝑏�𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐽𝐽𝐽𝐽𝐽𝐽𝐽𝐽 207

Acetitomaculum ( , = 0.19 ± 0.04; , = 0.48 ± 0.07; , = 0.63 ± 0.07) and 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 3 𝑏𝑏�𝑅𝑅𝑅𝑅𝑅𝑅 𝑅𝑅 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑏𝑏�𝑅𝑅𝑅𝑅𝑅𝑅 𝑅𝑅 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑏𝑏�𝑅𝑅𝑅𝑅𝑅𝑅 𝑅𝑅 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 Jeotgalibaca ( , = 0.23 ± 0.04; , = 0.43 ± 0.07; , = 0.50 ± 0.06). These 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 3 �𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝐽𝐽𝐽𝐽𝐽𝐽𝐽𝐽 �𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝐽𝐽𝐽𝐽𝐽𝐽𝐽𝐽 �𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝐽𝐽𝐽𝐽𝐽𝐽𝐽𝐽 positive path relationships𝑏𝑏 can be interpreted𝑏𝑏 as commensal biological𝑏𝑏 relationships: where the

bacteria that is a source of variance experiences no benefit or harm, but the bacteria that extracts

variance can benefit by increasing monotonically in relative abundance.

Co-occurrence of path relationships for treatment groups

Path relationships among the observed genera were identified in Model 1 (BT group) and

Model 2 (CTRL group). For the BT (Model 1) and CTRL (Model 2) groups, there were

statistically significant (p ≤ .0001) and positive standardized total effects from

Christensenellaceae_R7_group to Rikenellaceae_RC9_gut_group ( , = 0.25 ± 0.07; 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑏𝑏�𝐶𝐶ℎ𝑟𝑟𝑟𝑟 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 , = 0.19 ± 0.05) and from Pseudomonas to Acetitomaculum ( , = 0.42 ± 0.04; 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑏𝑏�𝐶𝐶ℎ𝑟𝑟𝑟𝑟 𝑅𝑅𝑅𝑅 𝑅𝑅𝑅𝑅 𝑏𝑏�𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 , = 0.41 ± 0.08), Atopostipes ( , = 0.54 ± 0.04; , = 0.22 ± 0.05), 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑏𝑏� 𝑏𝑏� 𝑏𝑏� Phascolarctobacterium ( , = 0.31 ± 0.08; , = 0.22 ± 0.09), and 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑏𝑏�𝑃𝑃𝑠𝑠𝑒𝑒𝑒𝑒 𝑃𝑃ℎ𝑎𝑎𝑎𝑎 𝑏𝑏�𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑃𝑃ℎ𝑎𝑎𝑎𝑎 Rikenellaceae_RC9_gut_group ( , = 0.28 ± 0.04; , = 0.23 ± 0.06). The 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑏𝑏�𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝑏𝑏�𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑃𝑃ℎ𝑎𝑎𝑎𝑎 standardized total effect from Pseudomonas to Mannheimia was negative ( , = -0.02± 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑏𝑏�𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 0.04, P = 0.69; , = -0.12 ± 0.04, P = 0.002). Standardized total effects from 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 Psychrobacter were𝑏𝑏� positive and statistically significant (P = 0.022) to Acetitomaculum

( , = 0.68 ± 0.05; , = 0.15 ± 0.04), Alloprevotella ( , = 0.36 ± 0.07; 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑏𝑏� 𝑏𝑏� 𝑏𝑏� , = 0.52 ± 0.07), Atopostipes ( , = 0.88 ± 0.05; , = 0.07 ± 0.02), 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑏𝑏�𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑏𝑏�𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑏𝑏 �𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 Christensenellaceae_R7_group ( , = 0.20 ± 0.07; , = 0.06 ± 0.02), and 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐶𝐶ℎ𝑟𝑟𝑟𝑟 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐶𝐶ℎ𝑟𝑟𝑟𝑟 𝑏𝑏� 𝑏𝑏� Rikenellaceae_RC9_gut_group ( , = 0.05 ± 0.02; , = 0.07 ± 0.02). Furthermore, 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑏𝑏�𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝑏𝑏�𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 208

Ruminococcaceae_UCG005 statistically significantly (P <.0001) and positively affected

Atopostipes ( , = 0.24 ± 0.05; , = 0.74 ± 0.05), Christensenellaceae_R7_group 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑏𝑏�𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑏𝑏�𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 ( , = 0.40 ± 0.07; , = 0.65 ± 0.06), and Rikenellaceae_RC9_gut_group 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑏𝑏�𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝐶𝐶 ℎ𝑟𝑟𝑟𝑟 𝑏𝑏�𝑅𝑅𝑅𝑅𝑅𝑅 𝑅𝑅 𝐶𝐶ℎ𝑟𝑟𝑟𝑟 ( , = 0.58 ± 0.05; , = 0.78 ± 0.06). 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 2 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝑏𝑏� In the MP group (Model𝑏𝑏� 3), there were fewer path relationships that co-occurred in the BT

(Model 1) and CTRL (Model 2) groups. The causal relationships between genera that were

unidirectional for the CTRL and BT groups were frequently bi-directional in the MP group. The

relative abundances of seven genera become exogenous variables in Model 3 (Table 6.1). One

genus in the MP group (Atopostipes) exhibited standardized total effects that co-occurred in the

BT group. The standardized total effects from Atopostipes were positive and statistically

significant (p ≤ 0.02) to Acetitomaculum ( , = 0.32 ± 0.07; , = 0.75 ± 0.08) and 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 3 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑏𝑏� 𝑏𝑏� Jeotgalibaca ( , = 0.06 ± 0.03; , = 0.10 ± 0.04). 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 3 �𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝐽𝐽𝐽𝐽𝐽𝐽𝐽𝐽 �𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝐽𝐽𝐽𝐽𝐽𝐽𝐽𝐽 The standardized𝑏𝑏 total effects from𝑏𝑏 the relative abundance of Lactobacillus to the relative

abundances of other bacteria are shown in Table S6.3. In this case, the effects from Lactobacillus

to Acetitomaculum, Acinetobacter, Alloprevotella, and Jeotgalibaca were positive in both BT and

CTRL group. In Model 3 (MP), the relative abundance of Lactobacillus did not affect the relative

abundances of the selected genera. Based on CTRL data (Model 2) Lactobacillus was modeled to

increase monotonically the relative abundance of Acetinomaculum, which inhibits Mannheimia

(Fig 7). The effect of Lactobacillus on Mannheimia was therefore indirect in Model 2. In the BT

group (Model 1), Lactobacillus had a direct positive effect and an indirect negative effect on

Mannheimia. The indirect inhibition of Mannheimia was mediated by Phascolarctobacterium,

which inhibited Mannheimia. The standardized total effects from Lactobacillus to Mycoplasma

209

and Pasteurella were negative in the CTRL group, but these effects were not statistically

significant in the BT group.

6.3.7 Antimicrobial resistance determinants in the NP microbiota

The macrolide resistance gene msr(E) increased in the NP microbiome of MP calves during the first 28 days of the study (p < 0.01) (Figure 6.8A). The abundance of msr(E) was greater

in MP calves compared to CTRL and BT calves during the last two weeks of study (p < 0.05). The

abundance of the tetracycline resistant gene, tet(H) was not affected by treatment or time (p > 0.05)

(Figure 6.8B).

6.4 Discussion

Studies have suggested that mutualistic and antagonistic interactions take place within the

microbial community of the bovine respiratory tract (Corbeil et al., 1985; Timsit et al., 2018).

These interactions may contribute positively or negatively to microbiota-mediated colonization

resistance against respiratory pathogens (Timsit et al., 2016a; Zeineldin et al., 2019). Specifically,

in the nasopharynx, LAB have been negatively associated with Pasteurellaceae, and certain LAB

strains were capable of directly inhibiting BRD bacterial pathogens (Amat et al., 2019b; Corbeil

et al., 1985). These data suggested that LAB have potential as BTs for mitigating BRD-associated

pathogens. In the present study, we evaluated the effect of inoculating 6 previously characterized

Lactobacillus strains (Amat et al., 2019b) directly into the upper respiratory tract of cattle, on the

microbiota, and compared those effects to a common metaphylactic antimicrobial, tulathromycin.

Throughout the study, all calves remained healthy, with the exception of 5 that were treated for

BRD. However, the rates of BRD were not affected by treatment, and weight gains was similar across all treatments, indicating that the BTs did not adversely affect inoculated calves.

210

6.4.1 Colonization by BTs

A single application of BTs was selected, in order to fit within modern management systems of beef cattle, which employ a single processing event upon arrival at feedlots. Despite the BT strains originating from the nasopharynx of feedlot cattle and displaying strong in vitro adhesion to bovine turbinate cells (Amat et al., 2019b), real-time PCR and 16S rRNA SV analyses indicated that colonization by BT strains was transient, lasting up to 48 h. Similar findings from gastrointestinal studies support the difficulty that exogenous strains have in colonizing established microbial communities (Lee et al., 2004; Eloe-Fadrosh et al., 2015). The presence of similar indigenous species can affect colonization by exogenous bacteria (Zmora et al. 2018), potentially by limiting resource availability (Libertucci and Yong, 2019; Maldonado-Gomez et al., 2016 ).

For example colonization by Bifidobacterium longum (AH1206) was impeded in the gastrointestinal tract when endogenous B. longum was already present (Maldonado-Gómez et al.,

2016). Indeed, Lactobacillus were detected prior to inoculation of BTs though these were mainly assigned to OTU L. fermentum/mucosae. Thus inoculating the BTs earlier in the life of cattle may increase colonization potential if done prior to establishment of similar Lactobacillus spp.

6.4.2 Longitudinal effects of BTs and tulathromycin on the respiratory microbiota

Changes in the composition of the NP microbiota were observed after BT administration.

Despite BT treatment resulting in both increases and decreases of 28 taxa that significantly changed from baseline levels prior to inoculation (Figure 6.6), the strongest BT effects on these taxa were inhibitory. The BT strains have previously been shown to produce lactate, hydrogen peroxide, or encode bacteriocins (Amat et al., 2019b) which may have led to inhibition of resident bacteria. These direct impacts however would have only occurred within the 48 h the BTs colonizing the nasopharynx. It is interesting to note despite being transient, a single administration

211

of the BTs had a long-term impact on structure and diversity of the NP microbiota. Both a reduction

in richness and diversity was observed up to 42 days after BT inoculation. Information regarding

the effect of probiotics or BTs on respiratory microbiota in animal models is limited, however

similar effects have been reported for probiotics targeting the gastrointestinal microbiota. For

example, Zhang et al., (2014) observed that the gut microbial community was impacted up to two

weeks after administration of a L. casei probiotic strain. Other studies have reported prolonged

effects on the gut microbiota for up to one (Zmora et al., 2018) and five (Suez et al. 2018) months

following cessation of a probiotic cocktail containing 11 strains fed to humans. The underlying

mechanisms by which the BT Lactobacillus strains exerted long-term modulation of the

microbiota are difficult explain but are likely related to their initial effects on community members.

Bacterial communities in most niches form complex ecological interaction webs, and such

interactions are important in maintaining microbiome homeostasis and a symbiotic relationship

between microbe and host (Faust and Raes, 2012). We therefore evaluated the microbiome wide

community networks to gain insights into the changes in microbial community interactions in NP

microbiota in response to BTs and antibiotic administration. The interaction network among all

observed genera was predicted based on interaction networks from sequencing data. After

observing the distinct interaction network structure among treatment groups, we decided to further

investigate the causal networks among targeted taxa using structure equation modeling. Mainali

et al., (2019) was the first to apply causal models to detect interaction networks in the human

microbiome using conditional Granger causality. These authors argued that the causal models may

provide more accurate prediction of interaction networks among microbial community relative to

standard correlation/network analysis, as correlation is neither necessary nor sufficient to establish causation and environmental filtering can lead to correlation between non-interacting taxa. 212

The path model of the BT group revealed that a moderate degree of alterations had taken place in the causal relationship structure amongst observed genera, as seen by the changes in the magnitude of the direct and indirect effects from one genera to another. It is possible that the indirect effects of the BTs are what caused the prolonged effects on community diversity. Given the computational complexity of the path models, they were limited to 16 genera. However, it was interesting that ecological network analysis of all observed genera showed a more complex network for the BT group compared to both CTRL and MP calves. Yang et al., (2017) suggested that probiotics (Paracccus marcusii DB11 and Bacillus cereus G19) promote intestinal microbiota homeostasis by enhancing species-species interactions and increasing the number of connecters and/or module hubs within the network. Thus, BT administration may have promoted NP microbiota homeostasis by strengthening and promoting species-species interactions. This in turn may have reduced colonization potential by new bacteria, which is supported by the decreased richness observed for BT calves. While we did not sample calves prior to arrival, several studies have shown that diversity of the NP microbiota increases within days after feedlot placement and it has been suggested to be linked to susceptibility to BRD (Holman et al. 2017). It would therefore be interesting to measure the effect of BT inoculation in calves prior to transport in future studies, to evaluate whether they promote stabilization of the respiratory microbiota during transport.

Tulathromycin also altered the NP microbiota structure, diversity and composition, compared to CTRL calves. Measured by real-time PCR, the total bacteria per NP swab was reduced in MP calves on days 4 and 7. This was likely due to inhibition of members of the phylum

Proteobacteria, which decreased in MP calves. Interestingly, despite this reduction, Shannon diversity and richness were increased in MP calves on d 7, compared to the other treatments. While

Holman et al. (2019) did not see an increase in diversity of the NP microbiota of calves 213

administered tulathromycin, they did observe an increase in diversity in calves administered

oxytetracycline. In addition, like our study, Holman and colleagues (2019) observed an increase

in Ruminococcaceae_UCG-005, Rikenellaceae_RC9_gut_group, Phascolarctobacterium,

Facklamia, Jeotgalibaca, and Acinetobacter following tulathromyin injection, supporting that injectable antimicrobials may lead to an increase in microbial richness and diversity of the respiratory microbiota.

Most often the diversity of microbial community is claimed to be positively associated with the stability of microbiota (Larsen and Classen 2018) and health (Rinninella et al., 2019). Typically however, studies showing positive associations between bacterial diversity and health related to the gastrointestinal tract (Jernberg et al., 2010; Arboleya et al., 2012). However, it has also been argued that diversity in host-associated microbial communities may not always be associated with microbial community stability and health. For example, higher bacterial diversity and richness were observed in upper respiratory tract of children with invasive pneumococcal disease compared to healthy children (Camelo-Castillo et al., 2019). It was proposed by the authors that the higher diversity and richness of the NP microbiota was associated with impaired immune response. It was interesting in our study that antimicrobial treatment increased diversity of the respiratory tract, whereas antimicrobial administration can reduce diversity in the gastrointestinal tract. Perhaps this is refelection of increased exposure to exogenous bacteria that the respiratory tract faces compared to the digestive tract.

Functional properties of the gut microbiota including colonization resistance at the community level are believed to be maintained by the active interactions among genetically distinct and diverse microbial species, which allows the microbial community to perform complex metabolic activities. A single mono-species population or multi-species but with no 214

interconnectivity could not provide such collective functions of microbiota (Venturelli et al.,

2018). In addition, the functional activities and stability of a microbiota is influenced by the positive and negative feedback loops generated as a result of the cooperation (Elias et al., 2012;

Nadell et al., 2009) and competition (Foster and Bell, 2012) among the different microbial species

(Faust et al., 2012). Also, evolution of species-species interactions has been reported to determine microbial community productivity in new environments (Fiegna et al., 2014). Although proportionally equal positive and negative interactions between the bacterial species within NP microbiota was observed in both CTRL and BT group, such cooperative and competitive interactions were more intensive in BT group compared to CTRL group. In contrast, there was only positive interactions observed between these fewer species that remained in network model as interconnected species in the MP group.

The stability of a mammalian microbiota depends on how the species interacts with one another (Foster et al., 2017). Weak and competitive interactions are stabilizing and they limit positive feedback loops and the possibility that, if one species decreases, it will result in the decrease of others. Cooperative interactions determine the productivity of microbiome which is the efficiency of converting resources into energy (Foster et al., 2017). In the MP group, there was only a cooperative interaction observed in NP microbiota and the competitive interaction was missing, suggesting that NP microbiota in MP cattle would most likely experienced dysbiosis.

Whereas, in BT calves, both cooparative and competitive interactions were present with higher magnititude. This indicated that the NP microbiota in BT calves was most likely stable with normal community functions. It has been argued that the asymmetry of an unhealthy microbiome can relate to non-neutral states created by strong stressors, reducing host ability to contain certain bacteria and resulting in overgrowth of abundance and multiplication of species (Li and 215

Convertino, 2019). We observed that the abundance of most of the observed significant taxa (n

=28) were significantly enriched in calves received antibiotic. The overgrowth of these taxa therefore might be due to the non-neutral state of microbiota induced by antibiotic.

While the increase in diversity following tulathromcyin was unexpected, we hypothesize that it was related to the effects of the antimicrobial on the community network. For MP calves,

7 genera were determined to be exogenous to the developed path model, which had fewer interactions than those developed for the BT and CTRL groups of calves. In addition, the ecological network was also far less complex, and together these data indicate a substantial perturbation of the ecological network in MP calves. Likewise, Yang et al., (2017) reported that the use of the antimicrobial florfenicol resulted in deterioration of the ecological network among intestinal microbiota of sea cucumbers, leading to the homeostatic collapse of microbiota. Thus, the deterioration of the ecological network by tulathromycin may have made the NP microbiota in

MP calves more permissive to exogenous bacteria colonization, increasing diversity.

6.4.3 Longitudinal effects of BTs and tulathromycin on BRD-associated pathogens

The MP treatment reduced the prevalence of culturable M. haemoltyica. In support of this, tulathromycin injection has previously been shown to reduce NP colonization of M. haemolytica in feedlot cattle (Zaheer et al., 2013). In contrast, despite showing strong inhibitory properties against M. haemotytica in vitro, the BTs did not reduce prevalence of M. haemolytica. However, the BT strains were tested for competitive exclusion in vitro (Amat et al., 2019b). The fact that

24% of calves in the BT group were M. haemoltyica-positive prior to administration suggests that the BT strains may not be effective for displacement of M. haemolytica.

Despite MP calves being the only group to have several time points where no M. haemolytica or P. multocida could be cultivated from NP swabs, sequence analysis of

216

Mannheimia, Pasteurella, Histophilus, and Mycoplasma revealed no differences in relative

abundance of these BRD-associated genera. For Mannheimia, Pasteurella, and Histophilus, this

was likely a result of inter-animal variation and overall limited relative abundance of these genera.

While a previous study did show that tulathromycin injection reduced NP Pasteurella, the calves

in that study were ranch-derived and had higher levels of Pasteurella at feedlot entry (Hollman et

al., 2019). The limited number of BRD cases and reduced relative abundance of BRD-associated

genera in our study was likely attributed to the calves being at lower risk for BRD, having come from a single source and being placed into individual stalls instead of being co-mingled.

It was interesting that Histophilus was exogenous to microbial path models for all three

treatments. H. somni has previously been shown to have a 32.5 times greater chance of being

isolated from feedlot calves after 40 days on feed (Timsit et al., 2017). This suggests that genera

that were not included in the models, or external factors associated with feedlots, may be related

to its colonization of cattle. Lactobacillus had a direct positive effect on Mannheimia, for BT

calves, though indirect negative effects were also observed. This finding is difficult to explain but

may be related to limiting model analysis to 16 genera. Although the relative abundance of

Mannheimia did not increase as a result of BT treatment, future studies with Lactobacillus-based

BTs should be tested for their effects on Mannheimia. In contrast, for CTRL calves, Lactobacillus

had overall negative effects on Mannheimia and Pasteurella though they were not direct. Thus this

supports previous data showing the importance of indigenous Lactobacillus in bovine respiratory

health (Timsit et al., 2018; Amat et al. 2019a)

217

6.4.4 Longitudinal effects of BTs and tulathromycin on antimicrobial resistant determinants

The macrolide resistance gene msr(E) encodes a multidrug efflux pump conferring

resistance to macrolides, including tulathromycin (Desmolaize et al., 2011). This gene has been

detected in BRD pathogens and has been detected within integrative conjugate elements (Klima et

al., 2014; Klima et al., 2019). Given the increase in macrolide resistance in feedlot BRD pathogens

over the last 10 years (Alexander et al., 2013; Timsit et al., 2017), evaluation of resistance genes in respiratory bacteria is important, to maintain proper selection of antimicrobials. Despite altering the microbiota, BT-treated calves did not select for bacteria that encoded msr(E) or tet(H). In

contrast, msr(E) increased in MP calves showing that tulathromycin administration selected for

bacteria carrying this resistance gene. Similarly, tulathromycin use has previously been associated

with increased antimicrobial resistance determinants in NP microbiomes of feedlot cattle (Holman

et al., 2018 and 2019).

In summary, a single dose of intranasal BTs, which were developed from bovine

respiratory commensal Lactobacillus spp., induced longitudinal modulation of the NP microbiota

in post-weaned beef calves, with no adverse effects on animal health and growth performance.

While no differences in the relative abundances of BRD-associated genera were observed between treatments, the ecological networks of NP bacteria from BT-treated calves became more integrated. It was proposed that this resulted in a more stable microbiome with increased resilience against exogenous microorganisms. In contrast, disruption of the microbiome after tulathromycin treatment reduced resilience, leading to increased diversity in MP calves. Overall, this study showed that the bovine respiratory microbiota can be altered by administration of BTs and may therefore provide new opportunities to enhance respiratory resistance against BRD pathogens.

218

Future studies should evaluate the optimal administration of BTs as better resilience against BRD pathogens may result from administration prior to calves being shipped to feedlots.

219

6.5 Tables and Figures

Table 6.1 Manifest variables in the modified path models.

Model 1 (BT) Model 2 (CTRL) Model 3 (MP) n = 158 n = 160 n = 157 Iterations 9 7 7 Endogenous Acetitomaculum, Acetitomaculum, Acetitomaculum, Atopostipes, Acinetobacter, Acinetobacter, Jeotgalibaca, Lactobacillus, Alloprevotella, Alloprevotella, Atopostipes, Mannheimia, Mycoplasma, Atopostipes, Christensenellaceae_R7_gr Pasteurella, Psychrobacter Christensenellaceae_R7_ oup, Jeotgalibaca, group, Jeotgalibaca, Lactobacillus, Mannheimia, Lactobacillus, Mycoplasma, Pasteurella, Mannheimia, Phascolarctobacterium, Mycoplasma, Pseudomonas, Phascolarctobacterium, Psychrobacter, Pseudomonas, Rikenellaceae_RC9_gut_gr Psychrobacter, oup, Rikenellaceae_RC9_gut_ Ruminococcaceae_UCG005 group, Ruminococcaceae_UCG0 05

Exogenous Day, Histophilus, Day, Histophilus Day, Acinetobacter, Pasteurella Alloprevotella, Christensenellaceae_R7_group, Histophilus, Phascolarctobacterium, Pseudomonas, Rikenellaceae_RC9_gut_group, Ruminococcaceae_UCG005

98.4328, p = 0.4687 88.7043, p = 0.8578 101.7159, p = 0.3255 2 RMSEA,𝜒𝜒 0.0053 (0.0000, 0.0426) 0.0000 (0.0000, 0.0235) 0.0195 (0.0000, 0.0477) 90% confidence limits Tucker- 0.9995 1.0161 0.9901 Lewis

220

*Significant difference between treatments (P < 0.05). 221

The results are presented as estimated mean ± SEM. *Significant difference between treatments (

* represents P < 0.05; ** represents P < 0.01; **** represents P < 0.0001) 222

Figure 6.3 Beta diversity of nasopharyngeal microbiota.

Detrended corrospondence analysis (DCA) plots of the Bray-Curtis metric in nasopharyngeal samples obtained from obtained from cattle received either intranasal bacterial therapeutics (BT),

PBS (CTRL) or subcutaneous metaphylaxis (MP) (n = 20 per group) over the course of 42 days.

The percentages of variation explained by the DCA are indicated on the axes.

223

Figure 6.4 Alpha diversity of nasopharyngeal microbiota.

Alpha diversity of nasopharyngeal samples obtained from cattle received either intranasal bacterial therapeutics (BT), PBS (CTRL) or subcutaneous metaphylaxis (MP) (n = 20 per group) over the course of 42 days. Diversity estimates by day, colored by treatment, with error bars showing estimated variance. Top plot shows the mean richness estimate and the bottom panel shows the mean estimated Shannon diversity.

224

(A) overall comparison between treatment groups; (B) comparison by treatment groups and sampling day.

225

Figure 6.6 Taxa (n = 28) that showed a significant change from baseline (day-1) in BT and MP groups above and beyond any changes in the control (CTRL) group over the course of 42 days.

226

Figure.8A. Model 1 (BT group)

228

Figure.8B. Model 2 (CTRL group)

229

Figure.8C. Model 3 (MP group)

230

Figure 6.8 Path diagram of models showing the recursive structure of causal relationships among

the 16 selected genera in the nasopharyngeal microbiota of calves that received an intranasal

inoculation of either bacterial therapeutics (BT) (A) or PBS (CTRL) (B), or subcutaneous

metaphylaxis (MP) (C) (n = 20 per group).

Variances of measured variables (relative abundances of genera and time) with standard errors and

P values are shown in squares. Causal relationships that are implied by the model are shown as

solid lines with arrows that indicate the direction of causation. Causal paths are labelled with the

standardized path coefficients, standard errors, and P values. Covariance terms are shown in ovals

with dashed arrows between the variables that co-vary. Green line represents the positive effect, whereas the red line represents the negative effect. The thickness of the solid line represents the strength of the effect.

231

The results are presented as estimated mean ± (standard error of the mean). **Significant difference between treatments (P < 0.01). 232

Chapter Seven: General discussion, conclusions and future directions

7.1 General discussion

Although bovine respiratory disease (BRD) has been researched for more than a century,

rates of BRD-associated morbidity and mortality have not improved (Taylor et al., 2010). Given

the segmented nature of the beef industry, commercial feedlots in North America rely mainly on

antibiotic-based approaches to prevent and control BRD in feedlot cattle, in order to mitigate the

economic and welfare impacts of this disease (Duff and Galyean, 2007; Hilton, 2014). Cattle

arriving at feedlots are more susceptible to respiratory infections due to numerous host and

environmental stressors that they encounter during transition from cow-calf farm to the feedlot.

Therefore, modern management practices utilize metaphylactic administration of long-acting antibiotics to treat existing lung infections associated with bacterial pathogens at the time of arrival and to prevent infections after feedlot placement (Nickell and White, 2010). However, metaphylaxis is associated with the development of antimicrobial resistance (Catry et al., 2016;

Holman et al., 2019) and concerns over the emergence of multidrug-resistant BRD pathogens have prompted research into antimicrobial alternatives for the feedlot industry. The objective of this thesis was to develop intranasal BTs to mitigate the BRD pathogen Mannheimia haemolytica in feedlot cattle, as alternatives to metaphylactic antimicrobials.

Under immunological stress, BRD pathogens can proliferate in the upper respiratory tract

and translocate to the lungs, leading to bronchopneumonia (Caswell, 2014). Metaphylaxis

treatment inhibits upper respiratory tract colonization by BRD pathogens (Zaheer et al., 2013).

The premise of this thesis was therefore that intranasal inoculation of BTs, with inhibitory

properties towards BRD pathogenic bacteria, would reduce pathogen proliferation and enhance

233

respiratory health. While several bacterial pathogens are implicated in BRD, M. haemolytica

serotype 1 is most frequently isolated from affected cattle (Klima et al., 2014). As such, this bacterium was selected as a model BRD pathogen for BT development. Studies had previously shown the inhibitory effects of probiotics against human respiratory pathogens (Santagati et al.,

2012). Therefore, commercial probiotic bacteria, which were easily accessible, were evaluated for their potential to inhibit growth and cell adhesion of M. haemoltyica in vitro, as a preliminary experiment (Chapter 2). Except for S. thermophilus, all tested Lactobacillus and Paenibacillus probiotic strains inhibited the growth of M. haemolytica. They also displayed the ability to colonize bronchial epithelial cells and compete against M. haemolytica on the epithelial cells. This suggested that application of bacteria to the upper respiratory tract of calves had potential to inhibit proliferation of M. haemolytica.

In recent years, using probiotics to promote animal health and productivity has been

increasingly explored in ruminants, swine and poultry (Cameron and McAllister, 2019) due to the

notion of reducing non-therapeutic antibiotic usage and to save important drugs for serious

infections. Although animal performance has been the major focus of probiotic research in cattle,

improvements in high-throughput DNA sequencing has allowed for investigations into the

therapeutic application of therapeutic bacteria against bovine mastitis (Assis et al., 2015; Beecher

et al., 2019), reproductive health (Genís et al., 2018), and intestinal infections (Li et al., 2016 and

2018). Importantly, high-throughput sequencing has allowed in-depth study of host microbiotas

and has provided information on bacteria associated with certain host phenotypes (e.g. sick or

healthy) or life cycle stages (Shreiner et al., 2015). For example, Holman et al. (2015) observed

that Lactobacillaceae were reduced in feedlot calves that later become affected with BRD. Given

the inhibitory properties of Lactobacillus observed in Chapter 2, associations between LAB and 234

the BRD-related family Pasteurellaceae were investigated in transported calves using high

throughput sequencing in order to identify potential BT targets.

In Chapter 3, it was shown that the NP microbial community structure and abundance of

LAB families experienced significant changes as a result of auction market exposure and

transportation to a feedlot. Inverse correlations between LAB families and the BRD-associated

family Pasteurellaceae were observed and several bovine isolates from Lactobacillaceae,

Streptococcaceae and Enterococcaceae families displayed inhibition against M. haemolytica in

vitro. These results suggested that the presence of LAB in the nasopharynx had a competitive

exclusion effect on Pasteurellaceae members. Importantly, this study provided insight into

potential cooperative and competitive interactions taking place among bacteria within the NP

microbiota of cattle, with certain genera predicted to enhance or inhibit BRD-associated bacteria.

This finding was supported by Corbeil et al. (1985), who identified bacteria within

Corynebacterium, Stephylococcus and Acenitobacter spp. that enhanced BRD pathogen growth

and Bacillus that inhibited their growth. It was also identified in Chapter 3 that an increase in NP

microbial diversity occurred during the transition period from farm to two days after feedlot

placement. Similar results have been shown before for cattle transported to an auction market and

then to a feedlot (Stroebel et al. 2019) or directly to a feedlot (Holman et al 2017). Thus, the

respiratory microbiota of cattle become more permissive to exogenous bacteria colonization at the

time of feedlot placement, which might be linked to their increased susceptibility to BRD in

feedlots.

Probiotic strains originating from the human respiratory tract have been shown to reduce

respiratory tract infectious (i.e. Otis media) in humans when applied topically as a nasal spray

(Roos et al., 2001 and 2011; Marchisio et al., 2015). Although the commercial strains tested in 235

Chapter 2 both adhered to bovine bronchial cells and inhibited M. haemolytica in vitro, the strains were allochthonous (microbes that are not from the indigenous flora) to the NP microbial community of cattle. There is a greater chance for autochthonous bacteria (originating from the host target site) being adapted for successful colonization of the host target than allochthonous strains (Shewale et al., 2014). Therefore, in Chapter 4, bacteria originating from the respiratory tract of healthy feedlot calves were enrolled for development of BTs. Based on Chapter 3 showing inverse correlations between LAB and Pasteurellaceae, LAB isolates were specifically targeted.

A series of criteria were applied in a step-wise fashion to identify candidate BTs among respiratory isolates (Chapter 4). First, direct inhibition of M. haemolytica was evaluated, given the importance of this criterion in selecting BTs (Dunne et al., 2001; de Melo Pereira et al., 2018), and the rapid nature of the agar plug screening that allowed broad testing. The bacteria capable of inhibiting M. haemolytica were diverse, supporting the concept that the environment within the bovine respiratory tract is highly competitive, as shown in Chapter 3. Among the different genera tested, species within Lactobacillus displayed the strongest inhibition of M. haemolytica.

Lactobacillus has been identified as being more abundant in the lungs of healthy feedlot calves than in those diagnosed with BRD (Timsit et al., 2018) implying that this genus is involved in both upper and lower respiratory tract health. Therefore, the majority of isolates for downstream screening were from Lactobacillus.

Adhesion and colonization of the target site of the host are important features necessary for

BTs to deliver their antagonistic and immune modulatory effects (Collado et al., 2006), and this was the second selection criterion evaluated. Bovine turbinate cells were used as they originated from, and are prevalent in, the upper respiratory tract of cattle. Although, all the tested isolates in

Chapter 4 were cultivated from the bovine respiratory tract, their adherence abilities varied. 236

However, a majority of Lactobacillus isolates displayed relatively high adherence to BT cell monolayers, suggesting that these Lactobacillus strains had the best potential to colonize the bovine respiratory tract. Lactobacillus isolates were also best at inhibiting M. haemolytica colonization of turbinate cells in competition assays, the third selection criterion, and an important aspect of BTs in order to protect the host tissue from invading pathogens (Reid et al., 1990). Also an important attribute of candidate BTs is their lack of antimicrobial resistance, which could exasperate the problem of resistance in feedlots. Therefore antimicrobial susceptibility testing was employed and isolates with acquired resistance were screened out from further evaluation. Lastly, a select group of only Lactobacillus strains were analyzed for immunomodulation, which is an additional mechanism by which BTs enhance host resistance to pathogen infection (Yan and Polk,

2011), and it is often included as part of BT selection criteria (de Melo Pereira et al., 2018). Except for Histophilus somni (Lin et al., 2016), the immune-stimulatory effects of bacteria originating from bovine respiratory tract had not been evaluated using bovine turbinate cells. So it was difficult to narrow down potential target genes that would be influenced by the tested strains, which led to using a PCR array to test the effects of strains on turbinate cell response. Excessive immune stimulation by probiotic bacteria in susceptible individuals has been reported and is an undesirable effect (Doron and Snydman, 2015). None of the tested Lactobacillus strains caused an excessive immune response, in comparison to M. haemolytica.

Lactobacillus spp. have been labelled as having qualified presumption of safety “QPS” status (EFSA, 2007), and safety concerns over the use of Lactobacilli in animals has rarely been reported (Gaggìa et al., 2010) making them easier to commercialize. Based on their history of safety, and the employed selection criteria, 6 Lactobacillus strains were identified as having the greatest potential as BTs to inhibit M. haemolytica in Chapter 4. These strains were lastly 237

characterized for inhibitory mechanisms and were found to produce lactate at concentrations

capable of inhibiting M. haemolytica and two strains also carried bacteriocins. Because it has been

reported that multi-strain probiotics are more effective than single strains in enhancing host health,

through increasing diversity of probiotic actions (Chapman et al., 2010; Wang et al., 2015), all 6

strains were further evaluated in cattle as a combined BT cocktail (Chapters 5 and 6).

In a challenge study using dairy calves, the BT strains were tested for reduction in M.

haemolytica colonization (Chapter 5). Dairy calves were used because of their availability throughout the year, and because BRD is a highly relevant disease to the dairy industry. To minimize the inter-animal variations in host immunity, respiratory microbiome, and pathogen exposure, the calves were removed from the dams shortly after birth. Intranasal administration of the BT cocktail did not adversely affect calf health, suggesting that the dose of inoculation tested was tolerated by young dairy calves. Despite transient colonization, the BT strains reduced nasal colonization by M. haemolytica, modulated the nasal microbiota, and altered the structural relationship among the nasal microbiota. Thus for the first time, the intranasal use of BTs were shown to inhibit a principal BRD pathogen in vivo. A particularly important finding was a lower relative abundance of M. haemolytica in the lungs of calves administered the BTs. This implied that through inhibition of M. haemolytica in the upper respiratory tract, the BTs reduced

translocation of M. haemolytica to the lung. Such translocation is thought to be a prerequisite to

BRD pathogenesis (Rice et al., 2007) and was a key assumption for the development of the BTs.

Lack of prolonged colonization of BTs in dairy calves was unexpected given their

adherence to turbinate cells in Chapter 4. Although adhesion to mucous was not evaluated for the

BT strains, Lactobacillus spp. are known to colonize both mucus and underlying epithelial cells

of the human respiratory tract (Van Tassell and Miller, 2011; Saroj et al., 2016). It was speculated 238

that adhesion and colonization of the BT strains may increase in beef cattle, the host species the

BTs were originally isolated. It was also possible that syringe-based intranasal delivery resulted in rapid dissipation of the BTs in the dairy calves reducing site localization. Combined, the BTs were therefore next tested in beef calves originating from an auction market, and were inoculated using a nebulization device in an attempt to enhance colonization by BTs.

In Chapter 6, BT colonization of auction market-derived beef calves was evaluated.

Originating from the auction market, the calves used in this study were considered medium-risk, and were expected to have an increased chance for pathogen proliferation in the upper respiratory tract. In North American feedlots, medium, high, and ultra-high risk calves are often administered injectable antimicrobials to prevent BRD on arrival as part of management strategies used by conventional feedlots (Ives and Recheson, 2015). As such, antimicrobial alternatives employed in conventional feedlots should fit within management regimes. This was the reason that the BTs administered to calves in Chapters 5 and 6 were only inoculated once, to mimic potential use in feedlot operations where cattle are processed at entry. One group of calves was administered tulathromycin, to compare BT administration to a commonly prescribed metaphylactic antimicrobial used in feedlots.

Despite being medium-risk, proliferation of BRD-associated bacteria in the nasopharynx of cattle was not commonly observed. Based on bacterial culturing, tulathromycin reduced prevalence of M. haemolytica and P. multocida in the calves at a few time points, which is likely related to the reduction in BRD when this antimicrobial is used in feedlots (Zaheer et al., 2013).

However the relative abundance of four genera associated with BRD (Mannheimia, Pasteurella,

Histophilus, and Mycoplasma) were not different between treatments based on 16S rRNA sequencing, partially due to inter-animal variation observed. Similar to inoculation of dairy calves 239

(Chapter 5), BT colonization was transient in the auction market-derived beef calves (Chapter 6).

This might have been related to colonization resistance conferred by existing microbiota (Li et al.,

2019) and the presence of indigenous Lactobacillus that shared similar phylogenetic traits with

BT strains, thus limiting nutrient and niche availability (Maldonado-Gomez et al., 2016; Zmora et

al., 2018). Thus timing of BT administration to cattle is an important consideration.

A particularly important finding was that despite being transient, a single intranasal dose

of BTs induced longitudinal modulation of the NP microbiota. This resulted in reduced NP

microbial diversity and richness, which typically increase when calves are transported to feedlots

and are thought to have a role in increased BRD susceptibility (Homan et al., 2017). Upon analysis

of the community structure it was observed that the BTs increased the structure of cooperative and

competitive microbial relationships. Microbial communities with enhanced cooperation have been

shown to be more resistant to pathogen colonization (Sassone-Corsi and Raffatellu, 2015). In contrast, tulathromyin resulted in increased microbial richness and a breakdown of cooperative and competitive microbial relationships. The long-term effects BTs and tulathromycin on microbial cooperation are unknown but it was proposed that use of BTs prior to feedlot

transportation may be most beneficial if the BTs enhance NP community cooperation and

consequently, inhibit pathogen proliferation. In support of this, the NP microbiota profiles of

children with invasive pneumococcal disease have been shown to have higher bacterial diversity

and richness compared to healthy children (Camelo-Castillo et al., 2019). Overall however, in the

short-term, the BTs did not have equivalence or superiority over tulathromycin when considering

prevalence of BRD pathogenic species that were cultured. Thus their use in feedlots should only

be considered as part of a series of management practices to reduce antimicrobial use, and not as

a replacement of antimicrobials. 240

7.2 Future work

The challenge study using dairy calves showed that BTs reduced colonization of M.

haemolytica. Thus a logical next step would be to test the developed BTs in field studies utilizing

commercial feedlot cattle that are at high risk of respiratory pathogen proliferation. However,

several studies to further characterize the BTs should take place beforehand. First, optimal timing

for BT administration should be investigated. The results from Chapters 4 and 5 indicated that the

BTs were effective at inhibiting M. haemolytica through competition, whereas Chapter 6 showed that the BTs may not be effective at displacement. Inoculation of BTs prior to M. haemolytica

proliferation may therefore improve their effectiveness and as such, testing the effects of BT

administration prior to feedlot transportation would be valuable. Because colonization was

transient, inoculation of BTs immediately prior to transportation would increase the chance of the

BTs being prevalent in the upper respiratory tract at the time of feedlot arrival. However, given

the enhanced community cooperation of the respiratory microbiota that was observed after BT

inoculation, testing the administration of BTs 2-3 weeks prior to transportation would provide important information on whether increased resilience of the NP microbiota against BRD

pathogens occurs. In addition to timing of inoculation, the concentration of BTs administered to

calves should be optimized by evaluating dose responses (Allexander et al., 2013).

The dairy calf study also highlights the need for further evaluation of the developed BTs

on commercial dairy farms. Bovine respiratory disease has a significant impact on the dairy

industry (Mahendran et al., 2017; Dubrovsky et al., 2018). However, as for feedlot calves, timing

of BT inoculation should be further tested in dairy calves. The effect of early life administration

of BTs on the development of the respiratory microbiota, as well respiratory disease resistance in

later life, can be a focus of dairy calf studies. Unlike the beef industry, dairy calves are processed 241

at a young age thus there is opportunity to inoculate them within a day of birth. Additionally, given the observation that the tracheal microbiota was affected by BT administration, further investigations on the impact of intranasal BTs on lung microbiome homeostasis are warranted.

Overall, the research presented in this thesis provided a framework for developing bacterial-based therapeutics to mitigate M. haemolytica and to modulate the respiratory microbiota in cattle. The stepwise screening approach that was used for BT candidate selection could also serve as platform for developing BTs against other respiratory pathogens, including P. multocida and H. somni, or even against other infectious agents of cattle. Modifications to the screening process however could enhance BT identification. For example, despite using turbinate cells to select BT candidates with high adherence in vitro, in vivo colonization was limited. Therefore including additional selection criteria such as adherence to mucin or tissue explants may help identify BTs that are more likely to colonize cattle (Nishiyama et al., 2016).

Particularly beneficial to evaluating the effects of BTs was application of 16S rRNA gene sequencing. This allowed for insight into community modulation resulting from BT application and showed important causal relationships of bacteria within the respiratory microbiota. Future application of models similar to the ones developed in this thesis could lead to development of communities of bacteria for inoculation, rather than individual strains to enhance respiratory health of cattle. Potentially, communities of bacteria that are interdependent may be more competitive and have a better chance of colonization (Davey and O'toole, 2000).

7.3 Concluding remarks

By developing effective BTs against M. haemolytica, the beef industry will have one more technology to use for mitigating BRD, and reducing antimicrobial use. This project provided a 242

framework for the discovery of BTs against BRD pathogens. While further evaluation of the developed BTs is warranted, it was evident that the BTs may not be capable of directly replacing metaphylactic antimicrobials for BRD prevention within conventional feedlot systems. However, overall, the research presented here supports the premise that BTs have potential to enhance bovine respiratory health and may have application in the beef industry as part of larger management protocols aimed at reducing BRD.

243

References

Ackermann MR, Derscheid R, Roth JA. 2010. Innate immunology of bovine respiratory disease. Vet

Clin North Am Food Anim Pract 26(2):215-228. doi:10.1016/j.cvfa.2010.03.001.

Alexander TW, Cook S, Klima CL, Topp E, Mcallister TA. 2013. Susceptibility to tulathromycin in

Mannheimia haemolytica isolated from feedlot cattle over a 3-year period. Frontiers in

Microbiology 4:297.

Alugongo GM, Xiao J, Wu Z, Li S, Wang Y, Cao Z. 2017. Review: Utilization of yeast of

Saccharomyces cerevisiae origin in artificially raised calves. J Anim Sci Biotechnol 8:34.

doi:10.1186/s40104-017-0165-5.

Amat S, Holman DB, Timsit E, Schwinghamer T, Alexander TW. 2019a. Evaluation of the

Nasopharyngeal Microbiota in Beef Cattle Transported to a Feedlot, With a Focus on Lactic Acid-

Producing Bacteria. Front Microbiol. https://doi.org/10.3389/fmicb.2019.01988.

Amat S, Timsit E, Baines D, Yanke J, Alexander TW. 2019b. Development of Bacterial Therapeutics

Against the Bovine Respiratory Pathogen Mannheimia haemolytica. Applied and Environmental

Microbiology AEM.01359-19. https://doi.org/10.1128/AEM.01359-19.

Amat S, Holman DB, Timsit E, Gzyl KE, Alexander TW. 2019c. Draft genome sequences of 14

Lactobacillus, Enterococcus, and Staphylococcus isolates from the nasopharynx of healthy feedlot

cattle. Microbiol Resour Announc 8:e00534-19. https://doi.org/10.1128/MRA.00534-19.

Amat S, Baines D, Timsit E, Hallewell J, Alexander TW. 2019d. Essential oils inhibit the bovine

respiratory pathogens Mannheimia haemolytica, Pasteurella multocida and Histophilus somni and

have limited effects on commensal bacteria and turbinate cells in vitro. J Appl Microbiol, 126:

1668-1682. doi:10.1111/jam.14238.Amat S, Subramanian S, Timsit E, Alexander, TW. 2017.

244

Probiotic bacteria inhibit the bovine respiratory pathogen Mannheimia haemolytica serotype 1 in

vitro. Lett Appl Microbiol 64:343-349. https://doi.org/10.1111/lam.12723.

Amat S, Timsit E, Alexander TW. 2018. 356 Intranasal administration of novel bacterial therapeutics

reduces colonization of the bovine respiratory pathogen Mannheimia haemolytica in challenged

calves. J Anim Sci 96 (Issue suppl_3, 7):177–178. doi: 10.1093/jas/sky404.386.

Amat S, Timsit E, Holman DB, Alexander TW. 2016. 0472 Characterization of bovine nasopharyngeal

lactic acid bacteria and their in vitro antimicrobial activities against the respiratory pathogen

Mannheimia haemolytica. J Anim Sci 94 (Issue suppl_5, 1):225-226. doi: 10.2527/jam2016-0472.

Angen O, Mutters R, Caugant DA, Olsen JE, Bisgaard M. 1999a. Taxonomic relationships of the

[Pasteurella] haemolytica complex as evaluated by DNA–DNA hybridizations and 16S rRNA

sequencing with proposal of Mannheimia haemolytica gen nov, comb nov, Mannheimia

granulomatis comb nov, Mannheimia glucosida sp nov, Mannheimia ruminalis sp nov,

Mannheimia varigena sp nov. International Journal of Systematic Bacteriology 49:67-86.

Anholt RM, Klima C, Allan N, et al. 2017. Antimicrobial Susceptibility of Bacteria That Cause Bovine

Respiratory Disease Complex in Alberta, Canada. Front Vet Sci 4:207.

doi:10.3389/fvets.2017.00207.

Arbizu M. 2017. Pairwiseadonis: Pairwise multilevel comparison using adonis. R Package Version 00,

Arboleya S, Binetti A, Salazar N, Fernández N, Solís G, Hernández-Barranco A, Margolles A, de Los

Reyes-Gavilán CG, Gueimonde M. 2012. Establishment and development of intestinal microbiota

in preterm neonates. FEMS Microbiol. Ecol 79:763-772. doi: 10.1111/j.1574-6941.2011.01261.x.

245

Assis BS, Germon P, Silva A, Even S, Nicoli J, Loir YL. 2015. Lactococcus lactis V7 inhibits the cell

invasion of bovine mammary epithelial cells by Escherichia coli and Staphylococcus aureus.

Beneficial Microbes 6:879-886.

Babcock AH, Renter, DG, White BJ, Dubnicka SR, Scott HM. 2010. Temporal distributions of

respiratory disease events within cohorts of feedlot cattle and associations with cattle health and

performance indices. Prev Vet Med 97(3-4):198-219.

Bailey D, Stenquist NJ. 1996. Preconditioning calves for feedlots Available: Managing for Today’s

Cattle Market and Beyond, Livestock Marketing Information Center (LMIC) (1996)

http://ag.arizona.edu/arec/emc/TodaysCattlePub.html.

Bansil,R, Turner BS. 2018. The biology of mucus: Composition, synthesis and organization, Advanced

Drug Delivery Reviews 124:3-15.

Bartlett JA, Fischer AJ, McCray PB. 2008. Jr Innate immune functions of the airway epithelium. Contrib

Microbiol 15:147-163.

Baruch J, Cernicchiaro N, Cull CA, Lechtenberg K F, Nickell JS, Renter DG .2019. Performance of

multiple diagnostic methods in assessing the progression of bovine respiratory disease in calves

challenged with infectious bovine rhinotracheitis virus and Mannheimia haemolytica, Journal of

Animal Science 97(6): 2357-2367. https://doi.org/10.1093/jas/skz107.

Bassel L, Caswell J. 2018. Bovine neutrophils in health and disease. Cell Tissue Res.

https://doi.org/10.1007/s00441-018-2789-y.

Bates D, Mächler M, Bolker B, Walker S. 2014. Fitting linear mixed-effects models using lme4. arXiv

preprint arXiv:14065823.

246

Bavananthasivam J, Dassanayake RP, Kugadas A, Shanthalingam S, Call DR, Knowels DP, Srikumaran

S. 2012. Proximity-dependent inhibition of growth of Mannheimia haemolytica by Pasteurella

multocida. Appl Environ Microbiol 78:6683-6688.

Beecher C, Daly M, Berry DP, Klostermann K, Flynn J, Meaney W, Hill C, McCarthy TV, Ross RP,

Giblin L. 2009. Administration of a live culture of Lactococcus lactis DPC 3147 into the bovine

mammary gland stimulates the local host immune response, particularly IL-1β and IL-8 gene

expression. Journal of Dairy Research 76:340-348.

Behnsen J, Deriu E, Sassone-Corsi M, Raffatellu M. 2013. Probiotics: properties, examples, and specific

applications. Cold Spring Harb Perspect Med 3(3):a010074. doi:10.1101/cshperspect.a010074.

Beker M, Rose S, Lykkebo CA, Douthwaite S. 2018. Integrative and Conjugative Elements (ICEs) in

Pasteurellaceae Species and Their Detection by Multiplex PCR. Front Microbiol. 2018;9:1329.

https://doi.org/10.3389/fmicb.2018.01329.

Belkaid Y, Hand TW. 2014. Role of the microbiota in immunity and inflammation. Cell 157(1):121-41.

Bentley S, Sebaihia M. 2007. Bacterial therapeutics Nat Rev Microbiol 5:170-171.

Bermudez-Brito M, Plaza-Díaz J, Muñoz-Quezada S, Gómez-Llorente C, Gil A. 2012. Probiotic

Mechanisms of Action. Ann Nutr Metab 61:160-174.

Bernardeau M, Guguen M, Vernoux JP. 2006. Beneficial lactobacilli in food and feed: long-term use,

biodiversity and proposals for specific and realistic safety assessments, FEMS Microbiology

Reviews 30 (4): 487-513.

Bhatt K, Timsit E, Rawlyk N, Potter A, Liljebjelke K. 2018. Integrative conjugative element ICEHs1

encodes for antimicrobial resistance and metal tolerance in Histophilus somni. Front Vet Sci 5:153.

doi:10.3389/fvets.2018.00153.

247

Bogaert D, de Groot R, Hermans P. 2004. Streptococcus pneumoniae colonisation: the key to

pneumococcal disease. Lancet Infect Dis 4:144-154.

Bogovič-Matijašić B, Rogelj I. 2011. Bacteriocins of probiotics and enteric cytoprotection. In: Malago

JJ, Koninkx JFJG, Marinsek-Logar R (eds) Probiotic bacteria and enteric infections. Springer,

Dordrecht, pp 313-354. ISBN 978-94-007-0385-8.

Booker CW, Abutarbush SM, Morley PS, et al. 2008. Microbiological and histopathological findings in

cases of fatal bovine respiratory disease of feedlot cattle in western Canada. Can Vet J 49:473-

481.

Borah D, Gogoi O, Adhikari C, Kakoti BB. 2016. Isolation and characterization of the new indigenous

Staphylococcus sp. DBOCP06 as a probiotic bacterium from traditionally fermented fish and meat

products of Assam state. Egyptian Journal of Basic and Applied Sciences 3(3):232-240.

Boudreaux CM, Corstvet RE, Cooper RK, and Enright FM. 2005. Effects of cecropin B transgene

expression on Mannheimia haemolytica serotype 1 colonization of the nasal mucosa of calves.

Am. J. Vet. Res. 66:1922-1930. https://doi.org/10.2460/ajvr.2005.66.1922.

Brandtzaeg P. 2007. Induction of secretory immunity and memory at mucosal surfaces. Vaccine

25:5467-5484.

Brooks KR, Raper KC, Ward CE, et al. 2011. Economic effects of bovine respiratory disease on feedlot

cattle during backgrounding and finishing phases. Prof Anim Sci 27:195-203.

Buffie CG, Pamer EG. 2013. Microbiota-mediated colonization resistance against intestinal pathogens.

Nat Rev Immunol 13(11):790-801. doi:10.1038/nri3535.

Buhman MJ, Perino LJ, Galyean ML, Wittum TE, Montgomery TH, Swingle RS. 2000. Association

between changes in eating and drinking behaviors and respiratory tract disease in newly arrived

calves at a feedlot. Am J Vet Res 61:1163-1168. 248

Burton KD, Lydon JE, D'Alessandro DU, Koestner R. 2006. The differential effects of intrinsic and

identified motivation on well-being and performance: prospective, experimental, and implicit

approaches to self-determination theory. J Pers Soc Psychol 91:750-762.

Bustamante-Marin XM, Ostrowski LE. Cilia, Mucociliary Clearance. 2017. Cold Spring Harb Perspect

Biol 9(4):a028241. doi:10.1101/cshperspect.a028241.

Butler JE. Bovine immunoglobulins: A review J Dairy Sci 52 (1969):1895.

Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP. 2016. DADA2: high-

resolution sample inference from Illumina amplicon data. Nature Methods 13:581. https://doi.org/

10.1038/nmeth.3869.

Callewaert R, Holo H, Devreese B, Van Beeumen J, Nes I, De Vuyst L. 1999. Characterization and

production of amylovorin L471, a bacteriocin purified from Lactobacillus amylovorus DCE 471

by a novel three-step method. Microbiology 145:2559-2568.

Camelo-Castillo A, Henares D, Brotons P, et al. 2019. Nasopharyngeal microbiota in children with

invasive pneumococcal disease: identification of bacteria with potential disease-promoting and

protective effects. Front Microbiol 10:11. doi:10.3389/fmicb.2019.00011.

Cameron A, McAllister TA. 2016. Antimicrobial usage and resistance in beef production. J Anim Sci

Biotechnol 7:68-016-0127-3.

Cameron A, McAllister TA. 2019. Could probiotics be the panacea alternative to the use of

antimicrobials in livestock diets? Beneficial Microbes 0 (0):1-28.

Campana R, Federici S, Ciandrini E, Baffone W. 2012. Antagonistic activity of Lactobacillus

acidophilus ATCC 4356 on the growth and adhesion/invasion characteristics of human

Campylobacter jejuni. Curr Microbiol 64:371-378.

249

Caramia G, Silvi S. 2011. Probiotics: From the Ancient Wisdom to the Actual Therapeutical Perspective.

In: Malago JJ, Koninkx JFJG, Marinsek-Logar R (eds) Probiotic bacteria and enteric infections.

Springer, Dordrecht, pp 3-38. ISBN 978-94-007-0385-8.

Caswell JL, Bateman KG, Cai HY, Castillo-Alcala F. 2010. Mycoplasma bovis in respiratory disease of

feedlot cattle. Vet Clin North Am Food Anim Pract 26:365-79.

Caswell JL. 2014. Failure of respiratory defenses in the pathogenesis of bacterial pneumonia of cattle.

Veterinary pathology 51:393-409. https://doi.org/10.1177/0300985813502821.

Catry B, Decostere A, Schwarz S, Kehrenberg C, de Kruif A, Haesebrouck F. 2006. Detection of

tetracycline-resistant and susceptible pasteurellaceae in the nasopharynx of loose group-housed

calves. Vet. Res. Commun. 30:707-715. https://doi.org/10.1007/s11259-006-3347-8.

Catry B, Dewulf J, Maes D, Pardon B, Callens B, Vanrobaeys M, Opsomer G, de Kruif A, Haesebrouck

F. 2016. Effect of antimicrobial consumption and production type on antibacterial resistance in the

bovine respiratory and digestive tract. PloS one 11, e0146488.

https://doi.org/10.1371/journal.pone.0146488.

Cernicchiaro N, White BJ, Renter DG, Babcock AH. 2013. Evaluation of economic and performance

outcomes associated with the number of treatments after an initial diagnosis of bovine respiratory

disease in commercial feeder cattle. Am J Vet Res 74(2):300-309. doi: 10.2460/ajvr.74.2.300.

Cernicchiaro, N, White BJ, Renter DG, Babcock AH, Kelly L, Slattery R. 2012. Associations between

the distance traveled from sale barns to commercial feedlots in the United States and overall

performance, risk of respiratory disease, and cumulative mortality in feeder cattle during 1997 to

2009. J Anim Sci 90(6):1929-39. doi: 10.2527/jas.2011-4599.

Chapman CM, Gibson GR, Rowland I. 2011. Health benefits of probiotics: are mixtures more effective

than single strains? Eur J Nutr 50(1):1-17. 250

Chaucheyras-Durand F, Durand H. 2010. Probiotics in animal nutrition and health. Beneficial Microbes

1(1):3-9.

Chen L, Reeve J, Zhang L, Huang S, Wang X, Chen J. 2018. GMPR: A robust normalization method

for zero-inflated count data with application to microbiome sequencing data. PeerJ 6: e4600.

doi:10.7717/peerj.4600.

Cho I, Blaser MJ. 2010. The human microbiome: at the interface of health and disease. Nat Rev Genet

13:260-270.

Clawson M L, Murray RW, Sweeney MT, Apley MD, DeDonder KD, Capik SF, Smith TPL. 2016.

Genomic signatures of Mannheimia haemolytica that associate with the lungs of cattle with

respiratory disease, an integrative conjugative element, and antibiotic resistance genes. BMC

Genomics 17:982.

CLSI (Ed.). 2015. Performance standards for antimicrobial disk and dilution susceptibility tests for

bacteria isolated from animals (3rd ed), Clinical and Laboratory Standards Insitute, Wayne, PA

CLSI supplement VET01S.

CLSI. 2016. Methods for antimicrobial dilution and disk susceptibility testing of infrequently isolated

or fastidious bacteria; approved guideline-3rd ed, CLSI document M45. CLSI, Wayne, PA.

Collado MC, Gueimonde M, Sanz Y, Salminen S. 2006. Adhesion properties and competitive pathogen

exclusion ability of bifidobacteria with acquired acid resistance. J Food Prot 69:1675-1679.

Collins FWJ, O'Connor PM, O'Sullivan O, et al. 2017. Bacteriocin Gene-Trait matching across the

complete Lactobacillus Pan-genome. Sci Rep 7(1):3481.

Confer AW. 2009. Update on bacterial pathogenesis in BRD Anim Health Res Rev10:145-148.

Corbeil LB, Woodward W, Ward AC, Mickelsen WD, Paisley L. 1985. Bacterial interactions in bovine

respiratory and reproductive infections. J Clin Microbiol 21:803-807. 251

Corr SC, Li Y, Riedel CU, O'Toole PW, Hill C, Gahan CG. 2007. Bacteriocin production as a mechanism

for the antiinfective activity of Lactobacillus salivarius UCC118. Proc Natl Acad Sci U S A

104(18):7617-7621. doi:10.1073/pnas.0700440104.

Cotter PD. 2014. An ‘Upp’-turn in bacteriocin receptor identification. Mol Microbiol 92:1159-1163.

10.1111/mmi.12645.

Dassanayake RP, Call DR, Sawant A.A, Casavant NC, Weiser GC, Knowles DP, Srikumaran S. 2010.

Bibersteinia trehalosi inhibits the growth of Mannheimia haemolytica by a proximity-dependent

mechanism. Appl Environ Microbiol 76:1008-1013.

Davey ME, O'toole GA. Microbial biofilms: from ecology to molecular genetics. Microbiol Mol Biol

Rev. 2000, 64(4):847-867. doi:10.1128/mmbr.64.4.847-867.2000. de Melo Pereira GV, de Oliveira Coelho B, Magalhaes Júnior AI, Thomaz-Soccol V, Soccol CV. 2018.

How to select a probiotic? A review and update of methods and criteria Biotechnol. Adv 36

(8):2060-2076. de Steenhuijsen Piters WA, Sanders EA, Bogaert D. 2015. The role of the local microbial ecosystem in

respiratory health and disease. Philos Trans R Soc Lond B Biol Sci 370(1675):20140294.

doi:10.1098/rstb.2014.0294.

Dec M, Puchalski A, Urban-Chmiel R, Wernicki Andrzej. 2014. Screening of Lactobacillus strains of

domestic goose origin against bacterial poultry pathogens for use as probiotics. Poultry Sci

93(10):2464-2472.

Deng H, Li Z, Tan Y, et al. 2016. A novel strain of Bacteroides fragilis enhances phagocytosis and

polarises M1 macrophages. Sci Rep 6:29401. doi:10.1038/srep29401.

252

Derrien M, van Passel MW, van de Bovenkamp JH, Schipper RG, de Vos WM, Dekker J. 2010. Mucin-

bacterial interactions in the human oral cavity and digestive tract. Gut Microbes 1(4): 254-268.

doi: 10.4161/gmic.1.4.12778.

Desmolaize B, Rose S, Wilhelm C, Warrass R, Douthwaite S. 2011b. Combinations of macrolide

resistance determinants in field isolates of Mannheimia haemolytica and Pasteurella multocida.

Antimicrob. Agents Chemother 55:4128-4133. 10.1128/AAC.00450-11.

Distinct adhesion of probiotic strain Lactobacillus casei ATCC 393 to rat intestinal mucosa. Anaerobe

18:417-420.

Dobson A, Cotter PD, Ross RP, Hill C. 2012. Bacteriocin production: a probiotic trait? Appl. Environ.

Microbiol 78:1-6.

Doron S, Snydman DR. 2015. Risk and Safety of Probiotics, Clinical Infectious Diseases, 60( Issue

suppl_2):S129-S134. https://doi.org/10.1093/cid/civ085.

Duary RK, Rajput YS, Batish VK, Grover S. 2011. Assessing the adhesion of putative indigenous

probiotic lactobacilli to human colonic epithelial cells. Indian J Med Res 134(5):664-671.

Dubernet S, Desmasures N, Guéguen M. 2002. A PCR‐based method for identification of lactobacilli at

the genus level. FEMS Microbiology Letters 214:271-275. https://doi.org/10.1111/j.1574-

6968.2002.tb11358.x.

Dubrovsky SA, Van Eenennaam AL, Karle BM, Rossitto PV, Lehenbauer TW, Aly SS, 2019.

Epidemiology of bovine respiratory disease (BRD) in preweaned calves on California dairies: The

BRD 10K study. J Dairy Sci 102(8):7306-7319. https://doi.org/10.3168/jds.2018-14774.

Duff GC, Galyean ML. 2007. Board-invited review: recent advances in management of highly stressed,

newly received feedlot cattle. J Anim Sci 85:823-840.

253

Dunne C, O’ Mahony L, Murphy L, Thorton G, Morrissey D, O’ Halloran S, Feeney M, Flynn S,

Fitzgerald G, Daly C. Kiely B. O’ Sullivan GC, Shanahan F, Collins JK. 2001. In vitro selection

criteria for probiotic bacteria of human origin: correlation with in vivo findings. Am J Clin Nutr

73:S386-S392.

EFSA Panel on Additives and Products or Substances used in Animal Feed (FEEDAP), Rychen G,

Aquilina G, Azimonti G, Bampidis V, Bastos ML, Bories G, Chesson A, Cocconcelli PS,

Flachowsky G, Gropp J, Kolar B, Kouba M, Lopez-Alonso M, Lopez Puente S, Mantovani A,

Mayo B, Ramos F, Villa RE, Wallace RJ, Wester P, Brozzi R and Saarela M, 2017. Scientific

Opinion on the safety and efficacy of Lactobacillus buchneri NRRL B-50733 as a silage additive

for all animal species. EFSA Journal 15(7):4934-9

EFSA. 2007. European Food Safety Authority. Introduction of a Qualified Presumption of Safety (QPS)

approach for assessment of selected microorganisms referred to EFSA Opinion of the Scientific

Committee (Question No EFSA-Q-2005-293). EFSA Journal 587:1-16.

EFSA. 2012. Scientific Opinion on Lactiferm® (Enterococcus faecium) as a feed additive for weaned

piglets and calves. EFSA J. 10:2574 10.2903/j.efsa.2012.2574. doi: 10.2903/j.efsa.2012.2574.

Elias S, Banin E. 2012. Multi-species biofilms: living with friendly neighbors. FEMS Microbiol Rev.

http://dx.doi.org/10.1111/ j.1574-6976.2012.00325.x.

Ellis JA, Chamorro MF, Lacoste S, Gow SP, Haines DM. 2018. Bovine respiratory syncytial virus-

specific IgG-1 in nasal secretions of colostrum-fed neonatal calves. Can Vet J 59(5):505-508.

Eloe-Fadrosh EA, Brady A, Crabtree J, Drabek EF, Ma B, Mahurkar A, Ravel J, Haverkamp M, Fiorino

A-M, Botelho C, Andreyeva I, Hibberd PL, Fraser CM. 2015. Functional dynamics of the gut

microbiome in elderly people during probiotic consumption. mBio 6(2):e00231-15.

https://doi.org/10.1128/mBio.00231-15. 254

Elshaghabee FMF, Rokana N, Gulhane RD, Sharma C, Panwar H. 2017. Bacillus as potential probiotics:

status, concerns, and future perspectives. Front Microbiol 8:1490. doi:10.3389/fmicb.2017.01490.

European Commission. 2008.Technical guidance prepared by the Panel on Additives and Products or

Substances used in Animal Feed (FEEDAP) on the update of the criteria used in the assessment of

bacterial resistance to antibiotics of human or veterinary importance. The EFSA J,1-15.

European Food Safety Authority (EFSA). 2012. Guidance on the assessment of bacterial susceptibility

to antimicrobials of human and veterinary importance. EFSA Journal 10: 2740

Ewels P, Magnusson M, Lundin S, Käller M. 2016. MultiQC: Summarize analysis results for multiple

tools and samples in a single report. Bioinformatics 32:3047-3048.

FAO. 2016. Probiotics in animal nutrition - Production, impact and regulation by Yadav S Bajagai, Athol

V Klieve, Peter J Dart, Wayne L Bryden. Editor Harinder P.S. Makkar. FAO Animal Production

and Health Paper No. 179. Rome.

Faust K, Raes J. 2012. Microbial interactions: from networks to models. Nat Rev Microbiol 10:538-

550. https://doi.org/10.1038/nrmicro2832.

Faust K, Sathirapongsasuti JF, Izard J, Segata N, Gevers D, et al. 2012. Microbial co-occurrence

relationships in the human microbiome. PLoS Comput Biol 8(7):e1002606.

doi:10.1371/journal.pcbi.1002606.

Feichtinger M, Mayrhofer S, Kneifel W, Domig KJ. 2016. Tetracycline resistance patterns of

Lactobacillus buchneri group strains. Journal of Food Protection 79:1741-1747.

Fiegna F, Moreno-Letelier A, Bell T, Barraclough TG . 2014. Evolution of species interactions

determines microbial community productivity in new environments. ISME J 9:1235-1245.

Fijan S. 2014. Microorganisms with claimed probiotic properties: an overview of recent literature. Int J

Environ Res Public Health 11(5):4745-4767. doi: 10.3390/ijerph110504745. 255

Flórez AB, Mayo B. 2018. Genome analysis of Lactobacillus plantarum LL441 and genetic

characterisation of the locus for the lantibiotic plantaricin C. Front Microbiol 9: 1916.

Foster KR, Bell T. 2012. Competition, not cooperation, dominates interactions among culturable

microbial species. Curr Biol 22:1845-1850.

Foster KR, Schluter J, Coyte KZ, Rakoff-Nahoum S. 2017. The evolution of the host microbiome as an

ecosystem on a leash. Nature 548:43-51.

Francoz D, Buczinski S, Belanger M, Forte G, Labrecque O, Tremblay D, Wellemans V, Dubuc J. 2015.

Respiratory pathogens in Quebec dairy calves and their relationship with clinical status, lung

consolidation, and average daily gain. Journal of veterinary internal medicine / American College

of Veterinary Internal Medicine 29:381-387.

https://onlinelibrary.wiley.com/doi/full/10.1111/jvim.12531.

Frank GH, Briggs RE, Zehr ES. 1995. Colonization of the tonsils and nasopharynx of calves by

rifampicin-resistant Pasteurella haemolytica and its inhibition by vaccination. American Journal

of Veterinary Research 56:866-869.

Frank GH, Briggs RE. 1992. Colonization of the tonsils of calves with Pasteurella haemolytica. Am J

Vet Res 53(4):481-484.

Furusawa Y, Obata Y, Hase K, Semin Immunopathol. 2015. 37:17. https://doi.org/10.1007/s00281-014-

0455-3.

Gaeta NC, Lima SF, Teixeira AG, Ganda EK, Oikonomou G, Gregory L, Bicalho RC. 2017. Deciphering

upper respiratory tract microbiota complexity in healthy calves and calves that develop respiratory

disease using shotgun metagenomics. J Dairy Sci 100:1445-1458.

https://doi.org/10.3168/jds.2016-11522.

256

Gaggìa F, Mattarelli P, Biavati B. 2010. Probiotics and prebiotics in animal feeding for safe food

production Int J Food Microbiol 141:S15-S28.

Gao W, Howden BP, Stinear TP. 2018. Evolution of virulence in Enterococcus faecium, a hospital-

adapted opportunistic pathogen, Current Opinion in Microbiology 41:76-82.

Gärtner MA, Bondzio A, Braun N, Jung M, Einspanier R, Gabler C. 2015. Detection and

characterisation of Lactobacillus spp. in the bovine uterus and their influence on bovine

endometrial epithelial cells In Vitro. PLoS ONE 10(3):e0119793.

Gaspar C, Donders GG, Palmeira-de-Oliveira R, et al. 2018. Bacteriocin production of the probiotic

Lactobacillus acidophilus KS400. AMB Express 8(1):153. doi:10.1186/s13568-018-0679-z.

Genís S, Cerri RLA, Bach À, et al. 2018. Pre-calving intravaginal administration of lactic acid bacteria

reduces metritis prevalence and regulates blood neutrophil gene expression after calving in dairy

cattle. Front Vet Sci 5:135. doi:10.3389/fvets.2018.00135.

Gershwin LJ, Van Eenennaam AL, Anderson ML, McEligot HA, Shao MX, Toaff-Rosenstein R, et al.

2015. Single pathogen challenge with agents of the bovine respiratory disease complex. PLoS

ONE 10(11):e0142479. https://doi.org/10.1371/journal.pone.0142479.

Giraffa G, Gatti M, Rossetti L, Senini L, Neviani E. 2000. Molecular diversity within Lactobacillus

helveticus as revealed by genotypic characterization. Appl Environ Microbiol 66:1259-1265.

Gorden PJ, Plummer P. 2010. Control, management, and prevention of bovine respiratory disease in

dairy calves and cows. The Veterinary clinics of North America 26:243-259.

https://doi.org/10.1016/j.cvfa.2010.03.004.

Gordon SB, Read RC. 2002. Macrophage defences against respiratory tract infections. Br Med Bull

61:45-61.

257

Griffin D, Chengappa MM, Kuszak J, McVey DS. 2010. Bacterial pathogens of the bovine respiratory

Griffin D, Chengappa MM, Kuszak J, McVey DS. 2010. Bacterial pathogens of the bovine

respiratory disease complex. Vet Clin North Am Food Anim Pract 26:381-394. doi:

10.1016/j.cvfa.2010.04.004.

Grissett GP, White BJ, Larson RL. 2015. Structured literature review of responses of cattle to viral and

bacterial pathogens causing bovine respiratory disease complex. J Vet Intern Med 29(3):770-780.

doi:10.1111/jvim.12597.

Gueimonde M, Sánchez B, G de Los Reyes-Gavilán C, Margolles A. 2013. Antibiotic resistance in

probiotic bacteria. Front Microbiol 4:202. doi:10.3389/fmicb.2013.00202.

Guglielmetti S, Taverniti V, Minuzzo M, Arioli S, Stuknyte M, Karp M, Mora D. 2010. Oral bacteria as

potential probiotics for the pharyngeal mucosa. Appl Environ Microbiol 76:3948-3958.

Guillot L, Medjane S, Le Barillec K, Balloy V, Danel C, Chignard M, Si-Tahar M. 2004. Response of

human pulmonary epithelial cells to lipopolysaccharide involves Toll-like receptor 4 (TLR4)-

dependent signaling pathways: evidence for an intracellular compartmentalization of TLR4. J Biol

Chem 279:2712–2718.

Hall JA, Isaiah A, Estill CT, Pirelli GJ, Suchodolski JS. 2017. Weaned beef calves fed selenium-

biofortified alfalfa hay have an enriched nasal microbiota compared with healthy controls. PLoS

One 12:e0179215. doi: 10.1371/journal.pone.0179215.

Hand TW. 2016. The Role of the Microbiota in Shaping Infectious Immunity. Trends Immunol

37(10):647-658. doi:10.1016/j.it.2016.08.007.

Hanzlicek GA, White BJ, Mosier D, Renter DG, Anderson DE. 2010. Serial evaluation of physiologic,

pathological, and behavioral changes related to disease progression of experimentally induced

258

Mannheimia haemolytica pneumonia in postweaned calves American Journal of Veterinary

Research 71(3):359-369.

Hay KE, Barnes TS, Morton JM, Clements AC, Mahony TJ. 2014. Risk factors for bovine respiratory

disease in Australian feedlot cattle: use of a causal diagram‐informed approach to estimate effects

of animal mixing and movements before feedlot entry. Prev Vet Med 117:160-169.

Hay KE, Morton JM, Mahony TJ, Clements ACA, Barnes TS. 2016. Associations between animal

characteristic and environmental risk factors and bovine respiratory disease in Australian feedlot

cattle. Prev Vet Med 125: 66-74. doi.org/10.1016/j.prevetmed.2016.01.013.

Heilig HG, Zoetendal EG, Vaughan EE, Marteau P, Akkermans AD, de Vos WM. 2002.Molecular

diversity of Lactobacillus spp. and other lactic acid bacteria in the human intestine as determined

by specific amplification of 16S ribosomal DNA. Appl Environ Microbiol 68(1):114-123.

Hemarajata P, Versalovic J. 2013. Effects of probiotics on gut microbiota: Mechanisms of intestinal

immunomodulation and neuromodulation. Therap Adv Gastroenterol 6:39-51.

Henderson G, Cox F, Ganesh S, Jonker A, Young W. Global Rumen Census Collaborators, Janssen PH.

2015. Rumen microbial community composition varies with diet and host, but a core microbiome

is found across a wide geographical range. Sci Rep 5:14567. doi:10.1038/srep14567.

Hendrick SH, Bateman KG, Rosengren LB. 2013. The effect of antimicrobial treatment and preventive

strategies on bovine respiratory disease and genetic relatedness and antimicrobial resistance of

Mycoplasma bovis isolates in a western Canadian feedlot. Can Vet J 54(12):1146-1156.

Hill C, Guarner F, Reid G, Gibson GR, Merenstein DJ, Pot B, Morelli L, Canani RB, Flint HJ, Salminen

S, Calder PC, Sanders ME. 2014. The international scientific association for probiotics and

prebiotics consensus statement on the scope and appropriate use of the term probiotic. Nat. Rev.

Gastroenterol Hepatol 11:506-514. 259

Hillmann Francesca, Vangay P, Knights D, Hutkins RW, Walter J. 2016. Stable engraftment of

Bifidobacterium longum AH1206 in the human gut depends on individualized features of the

resident microbiome. Cell Host & Microbe 20(4):515-526.

https://doi.org/10.1016/j.chom.2016.09.001.

Hilton WM. 2014. BRD in 2014: where have we been, where are we now, and where do we want to go?

Anim Health Res Rev 15:120-122. http://dx.doi.org/10.1017/S1466252314000115.

Hodgson PD, Aich P, Manuja A, et al. 2005. Effect of stress on viral-bacterial synergy in bovine

respiratory disease: novel mechanisms to regulate inflammation. Comp Funct Genomics 6(4):244-

250. doi:10.1002/cfg.474.

Holman DB, McAllister TA, Topp E, Wright AD, Alexander TW. 2015a. The nasopharyngeal

microbiota of feedlot cattle that develop bovine respiratory disease. Vet Microbiol 180:90-95.

https://doi.org/10.1016/j.vetmic.2015.07.031.

Holman DB, Timsit E, Alexander TW, 2015b. The nasopharyngeal microbiota of feedlot cattle. Sci Rep

5:15557. https://doi.org/10.1038/srep15557.

Holman DB, Timsit E, Amat S, Abbott DW, Buret AG, Alexander TW. 2017. The nasopharyngeal

microbiota of beef cattle before and after transport to a feedlot. BMC Microbiol 17(1):70. doi:

10.1186/s12866-017-0978-6.

Holman DB, Timsit E, Amat S, Abbott DW, Buret AG, Alexander TW. 2017.The nasopharyngeal

microbiota of beef cattle before and after transport to a feedlot. BMC Microbiol 17(1):70.

doi:10.1186/s12866-017-0978-6.

Holman DB, Timsit E, Booker CW, Alexander TW. 2018. Injectable antimicrobials in commercial

feedlot cattle and their effect on the nasopharyngeal microbiota and antimicrobial resistance.Vet

Microbiol 214:140-147. 260

Holman DB, Yang W, Alexander TW. 2019. Antibiotic treatment in feedlot cattle: a longitudinal study

of the effect of oxytetracycline and tulathromycin on the fecal and nasopharyngeal microbiota.

Microbiome 7(1):86. doi: 10.1186/s40168-019-0696-4.

Hooper LV, Littman DR, Macpherson AJ. 2012. Interactions between the microbiota and the immune

system. Science 336:1268-1273. doi: 10.1126/science.1223490.

Hummel AS, Hertel C, Holzapfel WH, Franz CMAP. 2007. Antibiotic Resistances of Starter and

Probiotic Strains of Lactic Acid Bacteria. Appl. Environ. Microbiol73:730-739.

Ives SE, Richeson JT. 2015. Use of Antimicrobial Metaphylaxis for the Control of Bovine Respiratory

Disease in High-Risk Cattle, Veterinary Clinics of North America: Food Animal Practice

31(3):341-350. https://doi.org/10.1016/j.cvfa.2015.05.008.

Jacobsen CN, Nielsen VR, Hayford AE, Moller PL, Michaelsen KF, Paerregaard A, Sandström B, Tvede

M, Jakobsen M. 1999. Screening of probiotic activities of forty-seven strains of Lactobacillus spp.

by in vitro techniques and evaluation of the colonization ability of five selected strains in humans.

Appl Environ Microbiol 65:4949-4956.

Jeong H, Park SY, Chung WH, Kim SH, Kim N, Park SH, Kim JF. 2011. Draft genome sequence of the

Paenibacillus polymyxa type strain (ATCC 842T), a plant growth-promoting bacterium. J

Bacteriol 193:5026-5027.

Jericho KWF, Lejeune A, and Tiffin GB. 1986. Bovine herpesvirus-1 and Pasteurella haemolytica

aerobiology in experimentally infected calves. Am.J Vet Res 47:205-2-9.

Jernberg C, Löfmark S, Edlund C, Jansson J K. 2010. Long-term impacts of antibiotic exposure on the

human intestinal microbiota. Microbiology 156:3216-3223.

Johnson KK, Pendell DL. 2017. Market impacts of reducing the prevalence of bovine respiratory disease

in United States beef cattle feedlots. Front Vet Sci 4:189. doi: 10.3389/fvets.2017.00189. 261

Johnston D, Earley B, Cormican P, et al. 2017. Illumina MiSeq 16S amplicon sequence analysis of

bovine respiratory disease associated bacteria in lung and mediastinal lymph node tissue. BMC

Vet Res 13(1):118. doi:10.1186/s12917-017-1035-2.

Joint FAO/WHO Working group report on drafting guidelines for the evaluation of probiotics in food.

London, Ontario, Canada, April 30 and May 1, 2002.

Jundi K, Greene CM. 2015. Transcription of Interleukin-8: How Altered Regulation Can Affect Cystic

Fibrosis Lung Disease. Biomolecules 5(3):1386-98.

Juntunen M, Kirjavainen PV, Ouwehand AC, Salminen SJ, Isolauri E. 2001. Adherence of probiotic

bacteria to human intestinal mucus in healthy infants and during rotavirus infection. Clin Diagn

Lab Immunol 8:293-296.

Kalischuk LD, Inglis GD. 2011. Comparative genotypic and pathogenic examination of Campylobacter

concisus isolates from diarrheic and non-diarrheic humans. BMC Microbiol 11:53-2180-11-53.

Karapetkov N, Georgieva R, Rumyan N, Karaivanova E. 2011. Antibiotic susceptibility of different

lactic acid bacteria strains. Benef. Microbe 2:335-339.

Kato A, Schleimer RP. 2007. Beyond inflammation: airway epithelial cells are at the interface of innate

and adaptive immunity. Curr Opin Immunol 19(6):711-20.

Kemgang TS, Kapila S, Shanmugam VP, Kapila R. 2014. Cross-talk between probiotic lactobacilli and

host immune system. J Appl Microbiol 117:303-319.

Kim M, Qie Y, Park J, Kim CH. 2016. Gut microbial metabolites fuel host antibody responses. Cell Host

Microbe 20(2):202-14.

Klima CL, Alexander TW, Hendrick S, McAllister TA. 2014a. Characterization of Mannheimia

haemolytica isolated from feedlot cattle that were healthy or treated for bovine respiratory disease.

Can J Vet Res 78(1):38-45. 262

Klima CL, Zaheer R, Cook SR, Booker CW, Hendrick S, Alexander TW, McAllister TA. 2014b.

Pathogens of bovine respiratory disease in North American feedlots conferring multidrug

resistance via integrative conjugative elements. J Clin Microbiol 52:438-448.

https://doi.org/10.1128/JCM.02485-13.

Klima CL, Alexander TW, Read RR, Gow SP, Booker CW, Hannon S, Sheedy C, McAllister TA,

Selinger LB. 2011. Genetic characterization and antimicrobial susceptibility of Mannheimia

haemolytica isolated from the nasopharynx of feedlot cattle. Vet Microbiol 149:390-398.

https://doi.org/10.1016/j.vetmic.2010.11.018.

Klima CL, Cook SR, Zaheer R, et al. 2016. Comparative genomic analysis of mannheimia haemolytica

from bovine sources. PLoS One 11:e0149520. doi:10.1371/journal.pone.0149520 [doi].

Klima CL, Holman DB, Ralston BJ, Stanford K, Zaheer R, Alexander TW, McAllister TA. 2019. Lower

Respiratory Tract Microbiome and Resistome of Bovine Respiratory Disease Mortalities. Microb

Ecol Aug 78(2):446-456. doi: 10.1007/s00248-019-01361-3.

Klima CL, Zaheer R, Cook SR, Rasmussen J, Alexander TW, Potter A, Hendrick S, McAllister TA.

2018. In silico identification and high throughput screening of antigenic proteins as candidates for

a Mannheimia haemolytica vaccine. Veterinary Immunology and Immunopathology 195:19-24.

Knowles MR, Boucher RC. 2002. Mucus clearance as a primary innate defense mechanism for

mammalian airways. J Clin Invest 109(5):571-577. doi:10.1172/JCI15217.

Kommineni S, Bretl DJ, Lam V, et al. 2015. Bacteriocin production augments niche competition by

enterococci in the mammalian gastrointestinal tract. Nature 526(7575):719-722.

doi:10.1038/nature15524.

Koo OK, Amalaradjou MA, Bhunia AK. 2012. Recombinant probiotic expressing Listeria adhesion

protein attenuates Listeria monocytogenes virulence in vitro. PLoS One 7:e29277. 263

Kubinak JL, Petersen C, Stephens WZ, et al. 2015. MyD88 signaling in T cells directs IgA-mediated

control of the microbiota to promote health. Cell Host Microbe 17(2):153-163.

doi:10.1016/j.chom.2014.12.009.

Larsen OFA, Claassen E. 2018. The mechanistic link between health and gut microbiota diversity. Sci

Rep 8:6-10. 10.1038/s41598-018-20141-6.

Larson RL. 2005. Effect of cattle disease on carcass trait. J Anim Sci 83:37-47.

Lebeer S, Vanderleyden J, De Keersmaecker SCJ. 2008. Genes and molecules of Lactobacilli supporting

probiotic action. Microbiology and Molecular Biology Reviews : MMBR 72(4):728-764.

Lee SJ, Bose S, Seo J-G, Chung W-S, Lim C-Y, Kim H. 2014. The effects of co-administration of

probiotics with herbal medicine on obesity, metabolic endotoxemia and dysbiosis: a randomized

double-blind controlled clinical trial. Clin. Nutr 33: 973-981.

https://doi.org/10.1016/j.clnu.2013.12.006.

Lee YK, Lim CY, Teng WL, Ouwehand AC, Tuomola EM, Salminen S. 2000. Quantitative approach in

the study of adhesion of lactic acid bacteria to intestinal cells and their competition with

enterobacteria. Appl Environ Microbiol 66(9):3692-7.

Lhermie G, Ferran AA, Assié S, Cassard H,El Garch F,Schneider M, Woerhlé F,Pacalin D, Delverdier

M,Bousquet-Mélou A and Meyer G. 2016. Impact of timing and dosage of a fluoroquinolone

treatment on the microbiological, pathological, and clinical outcomes of calves challenged with

Mannheimia haemolytica.Front Microbiol 7:237. https://doi.org/10.3389/fmicb.2016.00237.

Li C, Chng KR, Kwah JS, Av-Shalom TV, Tucker-Kellogg L. Nagarajan N. 2009. An expectation-

maximization algorithm enables accurate ecological modeling using longitudinal microbiome

sequencing data. Microbiome volume 7, Article number: 118.

264

Li J, Convertino M. 2019. Optimal Microbiome Networks: Macroecology and Criticality. Entropy 2019,

21, 506.

Li W, Yang S, Kim SO, Reid G, Challis JR, Bocking AD. 2014. Lipopolysaccharide-induced profiles of

cytokine, chemokine, and growth factors produced by human decidual cells are altered by

lactobacillus rhamnosus GR-1 supernatant. Reprod Sci 21:939-947.

Li X, Song L, Zhu S, Xiao Y, Huang Y, Hua Y, Chu Q, Ren Z. 2019. Two Strains of Lactobacilli

Effectively Decrease the Colonization of VRE in a Mouse Model. Frontiers in Cellular and

Infection Microbiology. 9: 6. https://doi.org/10.3389/fcimb.2019.00006.

Li Y, Wang Y, Liu X. 2012. The role of airway epithelial cells in response to mycobacteria infection.

Clin Dev Immunol 2012:791392. doi:10.1155/2012/791392.

Li Z, Kang H, You Q, Ossa F, Mead P, Quinton M, Karrow N. 2018. In vitro bioassessment of the

immunomodulatory activity of Saccharomyces cerevisiae components using bovine macrophages

and Mycobacterium avium ssp. paratuberculosis. Journal of Dairy Science 101:6271-6286.

Li Z, You Q, Ossa F, Mead P, Quinton M, Karrow NA. 2016. Assessment of yeast Saccharomyces

cerevisiae component binding to Mycobacterium avium subspecies paratuberculosis using bovine

epithelial cells. BMC Veterinary Research 12:42.

Libertucci J, Young VB. 2019. The role of the microbiota in infectious diseases. Nat Microbiol 4:35-

45.

Lillie LE, Thomson RG. 1972. The pulmonary clearance of bacteria by calves and mice. Can J Comp

Med 36(2):129-137.

Lima SF, Bicalho MLS, Bicalho RC. 2019. The Bos taurus maternal microbiome: Role in determining

the progeny early-life upper respiratory tract microbiome and health. PLoS One 14(3):e0208014.

Published 2019 Mar 6. https://doi.org/10.1371/journal.pone.0208014. 265

Lima SF, Teixeira AG, Higgins CH, Lima FS, Bicalho RC. 2016. The upper respiratory tract microbiome

and its potential role in bovine respiratory disease and otitis media. Sci Rep 6:29050. doi:

10.1038/srep29050.

Lin C, Agnes JT, Behrens N, et al. 2016. Histophilus somni stimulates expression of antiviral proteins

and inhibits BRSV replication in bovine respiratory epithelial cells. PLoS One 11(2):e0148551.

Published 2016 Feb 9. doi:10.1371/journal.pone.0148551.

Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expression data using real-time quantitative

PCR and the 2(-delta delta C(T)) method. Methods 25:402-408.

Loneragan GH, Thomson DU, Montgomery DL, Mason GL, Larson RL. 2005. Prevalence, outcome,

and health consequences associated with persistent infection with bovine viral diarrhea virus in

feedlot cattle. J Am Vet Med Assoc 226:595-601.

López A, Martinson SA. 2017. Respiratory system, mediastinum and pleurae. In: Pathologic Basis of

Veterinary Diseases. 6th Edition.Chapter 9; Zachary JF, Editor. Elsevier. 2017 pp 471-560.

Love MI, Huber W, Anders S. 2014. Moderated estimation of fold change and dispersion for RNA-seq

data with DESeq2. Genome Biology 15, 550. http://doi.org/ 10.1186/s13059-014-0550-8.

Lynch JB, Hsiao EY. 2019. Microbiomes as sources of emergent host phenotypes. Science

365(6460):1405-1409. DOI: 10.1126/science.aay0240.

MacPherson CW, Shastri P, Mathieu O, Tompkins TA, Burguière P. 2017. Genome-wide immune

modulation of TLR3-mediated inflammation in intestinal epithelial cells differs between single

and multi-strain probiotic combination. PLoS ONE 12(1):e0169847.

Mahendran SA, Booth R, Beekhuis L, Manning A, Blackmore T, Vanhoudt A, Bell N. 2017. Assessing

the effects of weekly preweaning health scores on dairy calf mortality and productivity parameters:

cohort study. The Veterinary record 181(8):196. http://dx.doi.org/10.1136/vr.104197. 266

Mainali K, Bewick S, Vecchio-Pagan B, Karig D, Fagan WF. 2019. Detecting interaction networks in

the human microbiome with conditional Granger causality. PLoS Comput Biol 15(5):e1007037.

https://doi.org/10.1371/journal.pcbi.1007037.

Malago JJ, Koninkx JFJG. 2011. Modulation of immune s by probiotics to protect against enteric

disorders. In: Malago JJ, Koninkx JFJG, Marinsek-Logar R (eds) Probiotic bacteria and enteric

infections. Springer, Dordrecht, pp 263-286. ISBN 978-94-007-0385-8.

Man WH, de Steenhuijsen Piters WA, Bogaert D. 2017. The microbiota of the respiratory tract:

gatekeeper to respiratory health. Nat Rev Microbiol 15:259-270.

Marchisio P, Santagati M, Scillato M, Baggi E, Fattizzo M, Rosazza C, et al. 2015. Streptococcus

salivarius 24SMB administered by nasal spray for the prevention of acute otitis media in otitis-

prone children. Eur J Clin.Microbiol Infect Dis 34:2377-2383. doi: 10.1007/s10096-015-2491-x.

Marden JP, Julien C, Monteils V, Auclair E, Moncoulon R, Bayourthe C. 2008. How does live yeast

differ from sodium bicarbonate to stabilize ruminal pH in high-yielding dairy cows? Journal of

Dairy Science 91:3528-3535.

Markowiak P, Śliżewska K. 2018. The role of probiotics, prebiotics and synbiotics in animal nutrition.

Gut Pathog 10:21. doi:10.1186/s13099-018-0250-0.

Martin BD, Witten D, Willis AD. 2019. Modeling microbial abundances and dysbiosis with beta-

binomial regression. arXiv:190202776 [stat]. 2019; Available: http://arxiv.org/abs/1902.02776

Martin M. 2011. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet

j 17: 10.

Mattarelli P, Holzapfel W, Franz CMAP, Endo A, Felis GE, Hammes W, et al. 2014. Recommended

minimal standards for description of new taxa of the genera Bifidobacterium, Lactobacillus and

related genera. Int J Syst Evol Microbiol 64:1434-1451. doi: 10.1099/ijs.0.060046-0. 267

McMullen C, Orsel K, Alexander TW, van der Meer F, Plastow G, Timsit E. 2018. Evolution of the

nasopharyngeal bacterial microbiota of beef calves from spring processing to 40 days after feedlot

arrival. Vet Microbiol 225:139-148.

McMurdie, P.J., Holmes, S., 2013. phyloseq: an R package for reproducible interactive analysis and

graphics of microbiome census data. PloS one 8:e61217.

https://doi.org/10.1371/journal.pone.0061217.

Meeusen ENT, Walker J, Peters A, Pastoret P-P, Jungersen G. 2007. Current Status of Veterinary

Vaccines. Clinical Microbiology Reviews 20(3):489-510. PMID:17630337.

Mokoena MP. 2017. Lactic Acid Bacteria and Their Bacteriocins: Classification, Biosynthesis and

Applications against Uropathogens: A Mini-Review. Molecules. 22(8):1255.

Monteagudo-Mera A, Rastall RA, Gibson GR, et al. 2019. Adhesion mechanisms mediated by probiotics

and prebiotics and their potential impact on human health. Appl Microbiol Biotechnol 103:6463-

6472. https://doi.org/10.1007/s00253-019-09978-7.

Mosier D.2014. Review of BRD pathogenesis: the old and the new. Anim Health Res Rev 15166–

168.10.1017/S1466252314000176.

Murray GM, O'Neill RG, More SJ. McElroy MC, Earley B, Cassidy JP. 2016. Evolving views on bovine

respiratory disease: An appraisal of selected key pathogens – Part 1,The Veterinary Journal

217:95-102.

Nadell CD, Xavier J, Foster KR. 2009. The sociobiology of biofilms. FEMS microbiology reviews

33:206-224.

Naghmouchi K, Baah J, Cudennec B, Drider D. 2013. Required characteristics of Paenibacillus

polymyxa JB-0501 as potential probiotic. Arch Microbiol 195:537-543.

268

Nagpal R, Wang S, Ahmadi S, Hayes J, Gagliano J, Subashchandrabose S, et al. 2018. Human-origin

probiotic cocktail increases short-chain fatty acid production via modulation of mice and human

gut microbiome Sci Rep 8:12649.

Nawrocki EP. 2009. Structural RNA homology search and alignment using covariance models. PhD

Thesis: Washington University School of Medicine. 2009.

Neal-McKinney JM, Lu X, Duong T, Larson CL, Call DR, Shah DH, et al. 2012. Production of Organic

Acids by Probiotic Lactobacilli Can Be Used to Reduce Pathogen Load in Poultry. PLoS ONE

7(9):e43928.

Nelson SL, Frank GH. 1989. Bovine serum and nasal secretion immunoglobulins against Pasteurella

haemolytica serotype 1 antigens Am J Vet Res 50:1244-1249.

Newbold CJ, Wallace RJ, McIntosh FM. 1996. Mode of action of the yeast Saccharomyces cerevisiae

as a feed additive for ruminants. British Journal of Nutrition 76:249-261.

Nickell JS, White BJ. 2010. Metaphylactic antimicrobial therapy for bovine respiratory disease in stocker

and feedlot cattle. Vet Clin North Am Food Anim Pract 26:285-301.

Nickell JS, White BJ. 2010. Metaphylactic Antimicrobial Therapy for Bovine Respiratory Disease in

Stocker and Feedlot Cattle, Veterinary Clinics of North America: Food Animal Practice 26(2):285-

301.

Nicola I, Cerutti F, Grego E, Bertone I, Gianella P, D’Angelo A, Peletto S, Bellino C. 2017.

Characterization of the upper and lower respiratory tract microbiota in Piedmontese calves

Microbiome 5:152.

Nishiyama K, Sugiyama M, Mukai T. 2016. Adhesion Properties of Lactic Acid Bacteria on Intestinal

Mucin. Microorganisms 4(3):34. doi:10.3390/microorganisms4030034.

269

Noyes NR, Benedict KM, Gow SP, Booker CW, Hannon SJ, McAllister TA, Morley PS. 2015.

Mannheimia haemolytica in feedlot cattle: prevalence of recovery and associations with

antimicrobial use, resistance, and health outcomes J Vet Intern. Med 29:705-713.

O’toole PW, Marchesi JR, Hill C. 2017. Next-generation probiotics: the spectrum from probiotics to live

biotherapeutics. Nature Microbiol 2:17057.

Oelschlaeger TA. 2010. Mechanisms of probiotic actions – a review. Int J Med Microbiol 300:57-62.

Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin PR, O’hara R, et al. 2013. Package ‘vegan’.

Community ecology package, version 2.

Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin PR, O’Hara RB, et al. 2017. Vegan: Community

Ecology Package. 2017.

Oksanen J, Kindt R, Legendre P, O’Hara B, Stevens MHH, Oksanen MJ, Suggests M. 2007. The vegan

package. Community ecology package 10, 631-637.

Olfert ED, Cross BM. Mcwilliams AA. (eds.). 1993. Guide to the care and use of experimental animals

. 2nd ed. Canadian Council on Animal Care, Ottawa, Ontario, Canada. Vol. 1.

O'Toole PW, Paoli M. 2017. The contribution of microbial biotechnology to sustainable development

goals: microbiome therapies. Microb Biotechnol 10(5):1066-1069. doi:10.1111/1751-7915.12752.

Panciera RJ, Confer AW. 2010. Pathogenesis and pathology of bovine pneumonia. The Veterinary

clinics of North America 26:191-214. https://doi.org/10.1016/j.cvfa.2010.04.001.

Parada JL, Caron CR, Medeiros ABP, Soccol CR. 2007. Bacteriocins from lactic acid 896 bacteria:

Purification, properties and use as biopreservatives. Braz Arch Biol Technol 897 50:521-542.

http://dx.doi.org/10.1590/S1516-89132007000300018 .

Pastva AM, Wright JR, Williams KL. 2007. Immunomodulatory roles of surfactant proteins A and D:

implications in lung disease. Proc Am Thorac Soc 4(3):252-7. 270

Patel RM, Denning PW. 2013. Therapeutic use of prebiotics, probiotics, and postbiotics to prevent

necrotizing enterocolitis: what is the current evidence? Clin Perinatol 40(1):11-25.

Perea Ve´lez M, De Keersmaecker SCJ, Vanderleyden J. 2007. Adherence factors of Lactobacillus in

the human gastrointestinal tract. FEMS Microbiol Lett 276: 140-148.

Petri R M, Schwaiger T, Penner GB, Beauchemin KA, Forster RJ, McKinnon JJ, et al. 2013.

Characterization of the core rumen microbiome in cattle during transition from forage to

concentrate as well as during and after an acidotic challenge. PLoS ONE 8:e83424. doi:

10.1371/journal.pone.0083424.

Pettigrew MM, Laufer AS, Gent JF, Kong Y, Fennie KP, Metlay JP.2012. Upper respiratory tract

microbial communities, acute otitis media pathogens, and antibiotic use in healthy and sick

children. Appl Environ Microbiol 78(17):6262–6270. doi:10.1128/AEM.01051-12.

Pilette C, Durham SR, Vaerman JP, Sibille Y. 2004. Mucosal immunity in asthma and chronic

obstructive pulmonary disease: a role for immunoglobulin A? Proc Am Thorac Soc 1:125-135.

Popova M, Molimard P, Courau S, Crociani J, Dufour C, Le Vacon F, Carton T. 2012. Beneficial effects

of probiotics in upper respiratory tract infections and their mechanical actions to antagonize

pathogens. J Appl Microbiol 113(6):1305-1318. doi: 10.1111/j.1365-2672.2012.05394.x.

Poppy GD, Rabiee AR, Lean IJ, Sanchez WK, Dorton KL, Morley PS. 2012. A meta-analysis of the

effects of feeding yeast culture produced by anaerobic fermentation of Saccharomyces cerevisiae

on milk production of lactating dairy cows. J Dairy Sci 95:6027-6041.

Portis E, Lindeman C, Johansen L, Stoltman G. 2012. A ten-year (2000–2009) study of antimicrobial

susceptibility of bacteria that cause bovine respiratory disease complex–Mannheimia haemolytica,

Pasteurella multocida, and Histophilus somni-in the United States and Canada. J Vet Diagn Invest

24:932-944. 271

Price MN, Dehal PS, Arkin AP. 2010. FastTree 2 -Likelihood Trees for Large Alignments. PLoS ONE

5: e9490.

Price RL, Bugeon L, Mostowy S, et al. 2019. In vitro and in vivo properties of the bovine antimicrobial

peptide, Bactenecin 5. PLoS One 14(1):e0210508. doi:10.1371/journal.pone.0210508.

Pridmore RD, Pittet AC, Praplan F, Cavadini C. 2008. Hydrogen peroxide production by 893

Lactobacillus johnsonii NCC 533 and its role in anti-Salmonella activity. FEMS 894 Microbiol

Lett 283:210-215. https://doi.org/10.1111/j.1574-6968.2008.01176.x.

Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. 2012. The SILVA ribosomal RNA

gene database project: improved data processing and web-based tools. Nucleic Acids Res

41:D590-D596. doi: 10.1093/nar/gks1219.

R Core Team. 2018. R: A language and environment for statistical computing. Vienna, Austria: R

Foundation for Statistical Computing. 2018.

Reese AT, Dunn RR. 2018. Drivers of microbiome biodiversity: a review of general rules, feces, and

ignorance. mBio 9:e01294-18. https://doi.org/10.1128/mBio.01294-18.

Reid G, Bruce AW, McGroarty JA, Cheng KJ, Costerton JW. 1990. Is there a role for Lactobacilli in

prevention of urogenital and intestinal infections? Clin Microbiol Rev 3:335-344.

Regev-Lehavi MD, Brik RB, et al. 2018. Personalized gut mucosal colonization resistance to empiric

probiotics is associated with unique host and microbiome features. Cell 174:1388-1405.e21.

https://doi.org/10.1016/j.cell.2018.08.041.

Ribble CS, Meek AH, Jim GK, Guichon PT. 1995. The pattern of fatal fibrinous pneumonia (shipping

fever) affecting calves in a large feedlot in Alberta (1985–1988). Can.Vet. J 36:753-757.

Rice JA, Carrasco-Medina L, Hodgins DC, Shewen PE. 2007. Mannheimia haemolytica and bovine

respiratory disease. Anim Health Res Rev 8:117-128. doi: 10.1017/S1466252307001375. 272

Rinninella E, Raoul P, Cintoni M, et al.2019. What is the healthy gut microbiota composition? a

changing ecosystem across age, environment, diet, and diseases. Microorganisms 7(1):14.

doi:10.3390/microorganisms7010014.

Roos K, Håkansson EG, Holm S. 2001. Effect of recolonization with interfering alpha streptococci on

recurrences of acute and secretory otitis media in children: randomised placebo-controlled trial.

BMJ 322:1-4. doi: 10.1136/bmj.322.7280.210.

Roos K, Simark-Mattsson C, Håkansson EG, Larsson L, Sandberg T, Åhrén C. 2011. Can probiotic

Lactobacilli eradicate persistent carriage of meticillin-resistant Staphylococcus aureus? J Hosp

Infect 78:77-78. doi: 10.1016/j.jhin.2011.01.010.

Salminen S, E Isolauri, E Salminen. 1996. Clinical uses of probiotics for stabilizing the gut mucosal

barrier: succesful strains for future challenges. Antonie van Leeuwenhoek 70:347-358.

Sanchez B, Delgado S, Blanco-Miguez A, Lourenco A, Gueimonde M, Margolles A. 2017. Probiotics,

gut microbiota, and their influence on host health and disease. Mol Nutr Food Res

61:10.1002/mnfr.201600240. Epub 2016 Oct 10.

Sanderson MW, Dargatz DA, Wagner BA. 2008. Risk factors for initial respiratory disease in United

States’ feedlots based on producer-collected daily morbidity counts. Can Vet J 49:373-378.

Santagati M, Scillato M, Muscaridola Metoldo V, La Mantia I, Stefani S. 2015. Colonization, safety,

and tolerability study of the Streptococcus salivarius 24SMBc nasal spray for its application in

upper respiratory tract infections. Eur J Clin Microbiol Infect Dis 34:2075-2080.

Santagati M, Scillato M, Patane F, Aiello C, Stefani S. 2012. Bacteriocin-producing oral streptococci

and inhibition of respiratory pathogens. FEMS Immunol Med Microbiol 65:23-31.

273

Santee CA, Nagalingam NA, Faruqi AA, DeMuri GP, Gern JE, Wald ER, Lynch SV. 2016.

Nasopharyngeal microbiota composition of children is related to the frequency of upper respiratory

infection and acute sinusitis Microbiome 4:34.

Saroj SD, Maudsdotter L, Tavares R, Jonsson AB. 2016. Lactobacilli interfere with Streptococcus

pyogenes hemolytic activity and adherence to host epithelial cells. Front Microbiol 7:1176.

https://doi.org/10.3389/ fmicb.2016.01176.

Sassone-Corsi M, Raffatellu M. 2015. No vacancy: how beneficial microbes cooperate with immunity

to provide colonization resistance to pathogens. J Immunol 194(9):4081-4087.

doi:10.4049/jimmunol.1403169.

Saxami G, Ypsilantis P, Sidira M, Simopoulos C, Kourkoutas Y, Galanis A. 2012.

Schleifer KH, Kraus J, Dvorak C, Klipper-Balz R, Collins MD, Fischer W. 1985. Transfer of

Streptococcus lactis and related streptococci to the genus lactococcus gen. nov. Syst Appl

Microbial 6:183-195.

Schwinghamer T, R Backer, D Smith, P Dutilleul. 2017. Block-recursive path models for rooting-

medium and plant-growth variables measured in greenhouse experiments. Agron. J 109:870-882.

https://doi.org/10.2134/agronj2016.08.0473.

Sharon Goh HM, Adeline Yong M H, Kelvin Kian LC, Kimberly AK. 2017. Model systems for the study

of Enterococcal colonization and infection, Virulence 8 (8):1525-1562, DOI:

10.1080/21505594.2017.1279766.

Shewale RN, Sawale PD, Khedkar CD, Singh A. 2014. Selection criteria for probiotics: a review. Int. J.

Probiotics Prebiotics 9:17-22.

Shreiner AB, Kao JY, Young VB. 2015. The gut microbiome in health and in disease. Curr Opin

Gastroenterol 31(1):69-75. doi:10.1097/MOG.0000000000000139. 274

Singh K, Ritchey JW, Confer AW. 2011. Mannheimia haemolytica: bacterial-host interactions in bovine

pneumonia. Vet Pathol 48(2):338-348.

Snyder E, Credille B, Berghaus R, Giguere S. 2017. Prevalence of multi drug antimicrobial resistance

in isolated from high-risk stocker cattle at arrival and two weeks after processing. J Anim Sci

95:1124-1131. doi: 10.2527/jas.2016.1110.

Sokol H, Pigneur B, Watterlot L, et al. 2008. Faecalibacterium prausnitzii is an anti-inflammatory

commensal bacterium identified by gut microbiota analysis of Crohn disease patients. Proc Natl

Acad Sci USA 105(43):16731–16736. doi:10.1073/pnas.0804812105.

Sotolongo J, Ruiz J, Fukata M. 2012. The role of innate immunity in the host defense against intestinal

bacterial pathogens. Curr Infect Dis Rep 14(1):15-23. doi:10.1007/s11908-011-0234-4.

Srikumaran S, Kelling CL, Ambagala A. 2007. Immune evasion by pathogens of bovine respiratory

disease complex. Anim. Health Res. Rev 8:215-229.

Step DL, Krehbiel CR, DePra HA, et al. 2008. Effects of commingling beef calves from different

sources and weaning protocols during a forty-two-day receiving period on performance and

bovine respiratory disease. J Anim Sci 86:3146-3158. doi: 10.2527/jas.2008-0883.

Stephens TP, Loneragan GH, Karunasena E, Brashears MM. 2007. Reduction of Escherichia coli O157

and Salmonella in feces and on hides of feedlot cattle using various doses of a direct-fed microbial.

Journal of Food Protection 70:1346-1349.

Stroebel C, Alexander TW, Workentine ML, Timsit T. 2018. Effects of transportation to and co-mingling

at an auction market on nasopharyngeal and tracheal bacterial communities of recently weaned

beef cattle. Vet Microbiol 223:126-133. doi: 10.1016/j.vetmic.2018.08.007.

275

Suez, Zmora N, Zilberman-Schapira G, Mor U, Dori-Bachash M, Bashiardes S, et al. 2018. Post-

antibiotic gut mucosal microbiome reconstitution is impaired by probiotics and improved by

autologous FMT. Cell 174, 1406–1423.e16. https://doi.org/10.1016/j.cell.2018.08.047.

Surachat K, Sangket U, Deachamag P, Chotigeat W. 2017. In silico analysis of protein toxin and

bacteriocins from Lactobacillus paracasei SD1 genome and available online databases. PLoS One

12(8):e0183548.

Surendran Nair M, Amalaradjou MA, Venkitanarayanan K. 2017. Antivirulence properties of probiotics

in combating microbial pathogenesis. New York, NY: Elsevier Ltd.

10.1016/bs.aambs.2016.12.001.

Tabe E S, Olya J, Doetkott D K, Bauer M L, Gibbs P S, Khaitsa M L.2008. Comparative effect of direct-

fed microbials on fecal shedding of Escherichia coli O157:H7 and Salmonella in naturally infected

feedlot cattle. J Food Protect 71:539-544.

Tam A, Wadsworth S, Dorscheid D, Man SFP, Sin DD. 2011. The airway epithelium: more than just a

structural barrier. Therapeutic Advances in Respiratory Disease, 255-273.

https://doi.org/10.1177/1753465810396539.

Taylor JD, Fulton RW, Lehenbauer TW, Step DL, Confer AW. 2010. The epidemiology of bovine

respiratory disease: What is the evidence for preventive measures? Can Vet J 51(12):1351-1359.

Teixeira AGV, McArt JAA, Bicalho RC. 2017. Efficacy of tildipirosin metaphylaxis for the prevention

of respiratory disease, otitis and mortality in pre-weaned Holstein calves. Vet J 219:44-48.

https://doi.org/10.1016/j.tvjl.2016.12.004.

Theurer ME, DE Amrine, BJ White. 2012. Remote noninvasive assessment of pain and health status in

cattle. Vet Clin North Am Food Anim Pract 29:59-74.

276

Thrift FA, Thrift TA. 2011. Review: Update on preconditioning beef calves prior to sale by cow-calf

producers Prof Anim Sci 27:73.

Timsit E, DB Holman, J Hallewell, TW Alexander. 2016a. The nasopharyngeal microbiota in feedlot

cattle and its role in respiratory health. Animal Frontiers 6:44-50. doi:10.2527/af.2016-0022.

Timsit E, Workentine M, Schryvers AB, Holman DB, van der Meer F, Alexander TW. 2016b. Evolution

of the nasopharyngeal microbiota of beef cattle from weaning to 40 days after arrival at a feedlot.

Vet Microbiol 187:75-81. doi: 10.1016/j.vetmic.2016.03.020.

Timsit E, Hallewell J, Booker C, Tison N, Amat S, Alexander TW. 2017. Prevalence and antimicrobial

susceptibility of mannheimia haemolytica, pasteurella multocida, and histophilus somni isolated

from the lower respiratory tract of healthy feedlot cattle and those diagnosed with bovine

respiratory disease. Vet Microbiol. 208:118-125. https://doi.org/10.1016/j.vetmic.2017.07.013.

Timsit E, Workentine M, van der Meer F, Alexander T. 2018. Distinct bacterial metacommunities inhabit

the upper and lower respiratory tracts of healthy feedlot cattle and those diagnosed with

bronchopneumonia. Vet Microbiol 221:105-113. doi: 10.1016/j.vetmic.2018.06.007.

Townsend HG, Meek AH, Lesnick TG, Janzen ED. 1989. Factors associated with average daily gain,

fever and lameness in beef bulls at the Saskatchewan Central Feed Test Station. Can J Vet Res

53:349-354.

Travis SM, Conway BA, Zabner J, Smith JJ, Anderson NN, Singh PK, Greenberg EP, Welsh MJ. 1999.

Activity of abundant antimicrobials of the human airway. Am J Respir Cell Mol Biol 20:872-9.

Tuomola EM, Ouwehand AC, Salminen SJ. 2000. Chemical, physical and enzymatic pre-treatments of

probiotic lactobacilli alter their adhesion to human intestinal mucus glycoproteins. Int. J. Food

Microbiol 60:75-81. 277

USDA.2013.. Feedlot 2011. Part IV: Health and Health Management on U.S. Feedlots with a Capacity

of 1,000 or More Head. Available online at:

http://www.aphis.usda.gov/animal_health/nahms/feedlot/downloads/feedlot2011/Feed11_dr_Part

IV.pdf.

USDA, 2012. Dairy Heifer Raiser, 2011.

https://www.aphis.usda.gov/animal_health/nahms/dairy/downloads/dairyheifer11/HeiferRaiser.p

df.

USDA–APHIS–VS: National Animal Health Monitoring System Beef Feedlot Study 2011. Part IV:

Health and Health Management on U.S. Feedlots with a Capacity of 1,000 or More Head. (2013).

Available from:

https://www.aphis.usda.gov/animal_health/nahms/feedlot/downloads/feedlot2011/Feed11_dr_Pa

rtIV.pdf van Heel AJ, de Jong A, Song C, Viel JH, Kok J, Kuipers OP. 2018. BAGEL4: a user-friendly web

server to thoroughly mine RiPPs and bacteriocins. Nucleic Acids Res. 46(W1):W278–W281.

doi:10.1093/nar/gky383. van Putten JPM, Strijbis K. 2017. Transmembrane Mucins: Signaling Receptors at the Intersection of

Inflammation and Cancer. J Innate Immun 9(3):281-299.

Van Tassell ML, Miller MJ .2011. Lactobacillus adhesion to mucus. Nutrients 3(5):613-36.

Vanbelle M, Teller E, Focant M. 1989. Probiotics in animal nutrition: a review Arch Anim Nutr 7:543-

567.

Vanderpool C, Yan F, Polk DB. 2008. Mechanisms of probiotic action: Implications for therapeutic

applications in inflammatory bowel diseases. Inflamm Bowel Dis 14:1585-1596.

278

Venturelli OS, Carr AC, Fisher G, et al. 2018. Deciphering microbial interactions in synthetic human

gut microbiome communities. Mol Syst Biol 14(6):e8157. doi:10.15252/msb.20178157.

Verna EC, Lucak S. 2010. Use of probiotics in gastrointestinal disorders: what to recommend? Therap

Adv Gastroenterol 3(5):307-19.

Versalovic J, Schneider M, de Bruijn FJ, Lupski JR. 1994. Genomic fingerprinting of bacteria using

repetitive sequence-based polymerase chain reaction. Methods Mol Cell Biol 5:25-40.

Wall R, Hussey SG, Ryan CA, O'Neill M, Fitzgerald G, Stanton C, Ross RP. 2008. Presence of two

Lactobacillus and Bifidobacterium probiotic strains in the neonatal ileum. ISME J 2:83-91.

Wang J, Tang H, Zhang C, Zhao Y, Derrien M, Rocher E, van-Hylckama Vlieg JE, Strissel K, Zhao L,

Obin M, et al. 2015. Modulation of gut microbiota during probiotic-mediated attenuation of

metabolic syndrome in high fat diet-fed mice. ISME J 9:1-15.

Wang Q, Garrity GM, Tiedje JM, Cole JR. 2007. Naive Bayesian classifier for rapid assignment of rRNA

sequences into the new bacterial taxonomy. J Appl Environ Microbiol 73:5261-5267.

https://doi.org/10.1128/AEM.00062-07.

Wang Y, Berg BP, Barbieri LR, Veira DM, McAllister TA. 2006. Comparison of 815 alfalfa and mixed

alfalfa-sainfoin pastures for grazing cattle : Effects on incidence of 816 bloat, ruminal

fermentation, and feed intake. Can J Anim Sci 86:383-392. 817 https://doi.org/10.4141/A06-009.

Weng FCH, Shaw GTW, Weng CY,Yang YJ and Wang D. 2017. Inferring microbial interactions in the

gut of the Hong Kong Whipping Frog (Polypedates megacephalus) and a validation using

probiotics. Front. Microbiol 8:525. https://doi.org/10.3389/fmicb.2017.00525.

Willis A, Bunge J, Whitman T. 2017. Improved detection of changes in species richness in high diversity

microbial communities. Journal of the Royal Statistical Society: Series C (Applied Statistics)

66:963-977. doi:10.1111/rssc.12206. 279

Willis A, Bunge J. 2015. Estimating diversity via frequency ratios. Biometrics 71:1042-1049.

doi:10.1111/biom.12332.

Willis AD, Martin BD. 2018. DivNet: Estimating diversity in networked communities. bioRxiv. 2018.

305045. doi:10.1101/305045.

Windeyer MC, Timsit E, Barkema H. 2017. Bovine respiratory disease in pre-weaned dairy calves: Are

current preventative strategies good enough? Vet J 224:16-17.

https://doi.org/10.1016/j.tvjl.2017.05.003.

Woo TD, Oka K, Takahashi M, Hojo F, Osaki T, Hanawa T, et al. 2011. Inhibition of the cytotoxic effect

of Clostridium difficile in vitro by Clostridium butyricum MIYAIRI 588 strain. J Med Microbiol

60:1617-25. doi:10.1099/jmm.0.033423-0.

Woolums AR, Karisch BB, Frye JG, Epperson W, Smith DR, Blanton J, Austin F, Kaplan R, Hiott

L, Woodley T, Gupta SK, Jackson CR. McClelland M. 2018. Multidrug resistant Mannheimia

haemolytica isolated from high-risk beef stocker cattle after antimicrobial metaphylaxis and

treatment for bovine respiratory disease. Vet Microbiol 221:143-152.

https://doi.org/10.1016/j.vetmic.2018.06.005.

Word AB, Broadway PR, Burdick Sanchez NC, Roberts SL, Richeson JT, Liang YL, Holland BP,

Cravey MD, Corley JR, Ballou MA, Carroll JA. 2019. Immune and metabolic responses of beef

heifers supplemented with Saccharomyces cerevisiae to a combined viral-bacterial respiratory

disease challenge, Translational Animal Science 3 (1): 135-148.

https://doi.org/10.1093/tas/txy117.

Wu HJ, Wu E. 2012. The role of gut microbiota in immune homeostasis and autoimmunity. Gut

Microbes 3(1):4-14.

280

Wu R, Smith D. 1982. Continuous multiplication of rabbit tracheal epithelial cells in a defined, hormone-

supplemented medium. In vitro 18:800-812.

Yan F, Polk DB. 2011. Probiotics and immune health. Curr Opin Gastroenterol 27(6):496-501.

doi:10.1097/MOG.0b013e32834baa4d.

Yang G, Peng M, Tian X, Dong S. 2017. Molecular ecological network analysis reveals the effects of

probiotics and florfenicol on intestinal microbiota homeostasis: An example of sea cucumber. Sci

Rep 7(1):4778. https://doi.org/10.1038/s41598-017-05312-1.

Zago M, Fornasari ME, Carminati D, Burns P, Suarez V, Vinderola G, Reinheimer J, Giraffa G. 2011.

Characterization and probiotic potential of Lactobacillus plantarum strains isolated from cheeses.

Food Microbiol 28:1033-1040.

Zaheer R, Cook SR, Klima CL, Stanford K, Alexander T, Topp E, Read RR, McAllister TA. 2013. Effect

of subtherapeutic vs. therapeutic administration of macrolides on antimicrobial resistance in

Mannheimia haemolytica and enterococci isolated from beef cattle. Front Microbiol 4:133.

http://dx.doi.org/10.3389/fmicb.2013.00133.

Zecchinon, L., T. Fett, D. Desmecht, 2005. How Mannheimia haemolytica defeats host defense through

a kiss of death mechanism. Veterinary Research 36:133-156.

Zeineldin M, Lowe J, Aldridge B. 2019. Contribution of the mucosal microbiota to bovine respiratory

health. Trends Microbiol. 1:18. https://doi.org/10.1016/j.tim.2019.04.005.

Zeineldin M, Lowe J, de Godoy M, Maradiaga N, Ramirez C, Ghanem M, Abd El-Raof Y, Aldridge B.

2017a. Disparity in the nasopharyngeal microbiota between healthy cattle on feed, at entry

processing and with respiratory disease. Vet Microbiol 208:30-37.

281

Zeineldin M, Lowe JF, Grimmer ED, et al. 2017b. Relationship between nasopharyngeal and

bronchoalveolar microbial communities in clinically healthy feedlot cattle. BMC Microbiol

17:138. doi: 10.1186/s12866-017-1042-2.

Zhang Y, Wang L, Zhang J, Li Y, He Q, Li H, Guo X, Guo J, Zhang H. 2014. Probiotic Lactobacillus

casei Zhang ameliorates high-fructose induced impaired glucose tolerance in hyperinsulinemia

rats. Eur J Nutr 53:221. https://doi.org/10.1007/s00394-013-0519-5.

Zhao Y, Chen F, Wu W, et al. 2018. GPR43 mediates microbiota metabolite SCFA regulation of

antimicrobial peptide expression in intestinal epithelial cells via activation of mTOR and STAT3.

Mucosal Immunol 11(3):752-762. doi:10.1038/mi.2017.118.

Zheng J, Gänzle MG, Lin XB, Ruan L, Sun M. 2015. Diversity and dynamics of bacteriocins from human

microbiome. Environ Microbiol 17:2133-2143. http://dx.doi.org/10.1111/1462-2920.12662.

Zhu YG, Johnson TA, Su JQ, et al. 2013. Diverse and abundant antibiotic resistance genes in Chinese

swine farms. Proc Natl Acad Sci USA 110(9):3435-3440. doi:10.1073/pnas.1222743110.

Zmora N, Zilberman-Schapira G, Suez J, Mor U, Dori-Bachash M, Bashiardes S, Kotler E, Zur M,

Regev-Lehavi D, Brik RB, et al. 2018. Personalized gut Mucosal colonization resistance to empiric

probiotics is associated with unique host and microbiome features. Cell. 174:1388–1405.e21.

https://doi.org/10.1016/j.cell.2018.08.041.

282

Appendix A: Supplementary tables

Supplementary Table S3.1 Differentially abundant OTUs in nasopharyngeal microbiota of feedlot cattle between d 0 and d 14 (N=13)a.

log2 fold Mean change FDR Phylum Family Genus OTU97 8.6 -5.9 0.000771 Actinobacteria Dietziaceae Dietzia OTU230 8.5 -5.9 5.32E-05 Actinobacteria Pseudonocardiaceae Saccharopolyspora OTU1008 8.4 -5.9 0.001134 Firmicutes Bacillaceae Bacillus OTU1631 7.8 -5.8 0.000289 Firmicutes Lachnospiraceae NA OTU1609 7.2 -5.7 0.000341 Firmicutes Lachnospiraceae Marvinbryantia OTU1856 7.2 -5.6 0.000426 Firmicutes Ruminococcaceae Faecalibacterium OTU1102 6.9 -5.6 1.11E-06 Firmicutes Thermoactinomycetaceae Thermoactinomyces OTU1119 6.6 -5.5 0.000464 Firmicutes Carnobacteriaceae Alkalibacterium OTU1634 6.4 -5.5 0.006191 Firmicutes Lachnospiraceae NA OTU930 5.8 -5.3 0.000155 Chloroflexi JG30-KF-CM45 NA OTU1087 5.5 -5.3 0.003588 Firmicutes Staphylococcaceae Nosocomiicoccus OTU2325 5.4 -5.2 1.79E-06 Firmicutes Ruminococcaceae Ruminococcaceae UCG-014 OTU1393 5.4 -5.2 2.49E-08 Firmicutes Family_XII Guggenheimella OTU1992 5.3 -5.2 0.003707 Firmicutes Ruminococcaceae Oscillibacter OTU48 5.2 -5.2 0.003756 Actinobacteria Iamiaceae Iamia OTU1993 4.7 -5.0 0.005813 Firmicutes Ruminococcaceae Oscillibacter OTU239 4.7 -5.0 0.012168 Actinobacteria Nocardiopsaceae Thermobifida OTU383 4.4 -4.9 0.005813 Bacteroidetes Muribaculaceae NA OTU565 4.4 -4.9 0.00193 Bacteroidetes Prevotellaceae Prevotellaceae UCG-003 OTU346 4.4 -4.9 0.014 Bacteroidetes Barnesiellaceae NA OTU2044 4.1 -4.8 0.00724 Firmicutes Ruminococcaceae Ruminococcaceae UCG-002 OTU2842 33.5 -4.8 0.000123 Proteobacteria Moraxellaceae Acinetobacter OTU246 3.8 -4.7 0.007817 Actinobacteria Atopobiaceae Olsenella OTU72 443.6 -4.7 7.52E-07 Actinobacteria Corynebacteriaceae Corynebacterium OTU1149 3.8 -4.7 0.008 Firmicutes Lactobacillaceae Lactobacillus OTU2054 47.0 -4.7 8.41E-06 Firmicutes Ruminococcaceae Ruminococcaceae UCG-005 OTU1554 24.1 -4.7 0.019568 Firmicutes Lachnospiraceae Lachnospiraceae NK3A20 group OTU629 3.6 -4.7 0.010055 Bacteroidetes Rikenellaceae dgA-11 gut group OTU143 39.1 -4.6 7.52E-06 Actinobacteria Intrasporangiaceae Ornithinimicrobium OTU2804 11.6 -4.6 0.000587 Proteobacteria Enterobacteriaceae Escherichia/Shigella OTU2385 21.2 -4.6 1.55E-06 Firmicutes Ruminococcaceae Ruminococcus OTU2453 3.4 -4.6 0.000294 Firmicutes Veillonellaceae Dialister OTU385 3.4 -4.6 0.004582 Bacteroidetes Muribaculaceae NA

283

OTU2092 3.4 -4.5 0.012248 Firmicutes Ruminococcaceae Ruminococcaceae UCG-010 OTU382 3.3 -4.5 0.012168 Bacteroidetes Muribaculaceae NA OTU1544 3.0 -4.4 0.028898 Firmicutes Lachnospiraceae Lachnospiraceae FCS020 group OTU1776 3.0 -4.4 0.025942 Firmicutes NA NA OTU1150 2.7 -4.3 0.03209 Firmicutes Lactobacillaceae Lactobacillus OTU60 2.7 -4.2 0.032498 Actinobacteria Actinomycetaceae Flaviflexus OTU415 2.5 -4.1 0.034278 Bacteroidetes NA NA OTU2055 43.8 -4.0 5.32E-05 Firmicutes Ruminococcaceae Ruminococcaceae UCG-005 OTU73 205.5 -4.0 0.000121 Actinobacteria Corynebacteriaceae Corynebacterium OTU2433 23.7 -4.0 0.005813 Firmicutes Erysipelotrichaceae Turicibacter OTU784 10.6 -3.8 0.041667 Bacteroidetes Flavobacteriaceae Flavobacterium OTU1139 115.6 -3.7 7.51E-06 Firmicutes Carnobacteriaceae Jeotgalibaca OTU1861 1.9 -3.7 0.02911 Firmicutes Ruminococcaceae Fastidiosipila OTU2869 43.4 -3.6 0.002348 Proteobacteria Moraxellaceae Psychrobacter OTU929 16.6 -3.6 0.006537 Chloroflexi JG30-KF-CM45 NA OTU285 12.5 -3.6 0.012519 Bacteroidetes Bacteroidaceae Bacteroides OTU284 32.8 -3.1 0.003874 Bacteroidetes Bacteroidaceae Bacteroides OTU1060 264.3 -2.7 0.005813 Firmicutes Planococcaceae Planococcus OTU473 26.7 -2.6 0.0305 Bacteroidetes Prevotellaceae Alloprevotella OTU2056 27.9 -2.6 0.046387 Firmicutes Ruminococcaceae Ruminococcaceae UCG-005 OTU1046 76.3 -2.2 0.015083 Firmicutes Planococcaceae Caryophanon OTU2841 429.7 -1.6 0.00312 Proteobacteria Moraxellaceae Acinetobacter OTU2673 155.8 1.9 0.000365 Proteobacteria Sphingomonadaceae Sphingomonas OTU2809 208.8 2.6 7.21E-07 Proteobacteria Enterobacteriaceae Proteus OTU103 42.2 2.6 0.012944 Actinobacteria Nocardiaceae Rhodococcus OTU1016 225.6 2.7 1.16E-11 Firmicutes Bacillaceae NA OTU3036 15075.3 2.9 0.015707 Tenericutes Mycoplasmataceae Mycoplasma

OTU1000 19.4 3.0 0.036686 Epsilonbacteraeota Campylobacteraceae Campylobacter OTU75 107.3 3.2 0.021723 Actinobacteria Corynebacteriaceae Corynebacterium OTU159 344.1 3.3 0.018534 Actinobacteria Microbacteriaceae NA OTU2911 147.1 3.5 5.02E-11 Proteobacteria Rhodanobacteraceae NA OTU2839 683.3 4.0 0.002617 Proteobacteria Pasteurellaceae Pasteurella OTU2674 18.1 4.2 0.041667 Proteobacteria Sphingomonadaceae Sphingomonas OTU2857 1174.1 4.2 0.002902 Proteobacteria Moraxellaceae Moraxella OTU2912 35.0 4.4 0.000832 Proteobacteria Rhodanobacteraceae Rhodanobacter OTU1163 8.8 7.2 0.000109 Firmicutes Lactobacillaceae Pediococcus OTU18 16.2 8.1 8.41E-06 Euryarchaeota Methanocorpusculaceae Methanocorpusculum OTU2875 29.4 9.0 8.15E-11 Proteobacteria Pseudomonadaceae Pseudomonas aOnly those OTUs with a log2 fold change of less than +/- 6 are included in this table.

284

Negative fold change values indicate OTUs that were enriched in the d 14 NP samples. FDR = false discovery rate.

285

Supplementary Table S5.1 The mean rectal temperature (± SD) and respiratory rate (± SEM) between dairy calves (n =12) that received intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT + Mh) or only M. haemolytica (Mh).

Days of experiment -1 3 5 7 9 11 13 16 Rectal temperature (°C) 38.4 ± 38.5 ± 38.6 ± 38.6 ± 38.7 ± 38.5 ± 38.7 ± 38.6 ± Mh 0.40 0.27 0.19 0.38 0.54 0.54 0.42 0.48 BT + 38.3 ± 38.3 ± 38.7 ± 38.5 ± 38.5 ± 38.4 ± 38.7 ± 38.5 ± Mh 0.23 0.57 0.44 0.22 0.23 0.21 0.41 0.30 Respiratory rate per min Mh 26 ± 1.5 25 ± 1.0 28 ± 0.9 28 ± 0.9 31 ± 1.6 30 ± 1.1 30 ± 1.1 33 ± 2.5 BT + Mh 24 ± 0.9 23 ± 1.0 26 ± 1.1 29 ± 1.4 32 ± 2.6 28 ± 0.9 31 ± 1.3 36 ± 2.7

286

Supplementary Table S5.2 Comparison of rectal temperature and respiratory rate between dairy calves received intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT +

Mh) or only M. haemolytica (Mh).

a Independent variable and level Estimated mean (SE) P-value Rectal temperature (ºC) Intercept 5.357 (0.2790) 0.0000 Mh Ref. BT + Mh -0.0403 (0.3947) 0.9196

Respiratory rate (Breaths per minute) Intercept 24.95 (1.367) 0.0000 Mh Ref. BT + Mh 0.0476 (0.881) 0.9574 Day of sampling -0.2646 (0.122) 0.0320 b Random effect = animal (n = 24)

287

Supplementary Table S5.3 Isolation rate of intranasal inoculated bacterial therapeutic (BT) strains from the nasopharynx of dairy calves (n =12) received intranasal inoculation of either bacterial therapeutics and M. haemolytica (BT + Mh) or M. haemolytica only

(Mh).

3E 72B 86D 63A 103C (L. 57A Treatment Sampling day (L. amylovorus) (L. buchneri) (L. buchneri) (L. curvatus) paracasei) (L. paracasei) -1 0 0 0 0 0 0 3 0 6 (50%) 6 (50%) 0 4 (33%) 4 (33%) 5 0 0 0 0 1 (8%) 1 (8%) 7 2 (16%) 2 (16%) 0 0 0 0 BT + Mh 9 0 0 0 0 0 0 11 0 0 0 0 0 0 13 2 (16%) 0 0 0 0 0 16 0 0 0 0 0 0 -1 0 0 0 0 0 0 3 0 0 0 0 0 0 5 0 0 0 0 0 0 7 0 0 0 0 0 0 Mh 9 0 0 0 0 0 0 11 0 0 0 0 0 0 13 0 0 0 0 0 0 16 0 0 0 0 0 0 aIdentification of BT strains was completed by two steps: in step one, a Lactobacillus genus-specific PCR was performed to identify the Lactobacillus isolates; in the second step, the (GTG)5 Rep PCR was performed on those Lactobacillus strains.

288

Supplementary Table S5.4 Differentially abundant OTUs at d 16 in the nasopharyngeal microbiota of dairy calves that received intranasal inoculation of only M. haemolytica or bacterial therapeutics and M. haemolytica (BT + Mh)a.

log2 (Fold OTU ID Mean Change) FDRI Phylum Family Genus OTU57 74.3 24.6b 1.6E-15 Firmicutes Lachnospiraceae NA OTU5 66.5 24.4 1.8E-15 Proteobacteria Moraxellaceae Moraxella OTU128 43.9 23.8 7.4E-15 Firmicutes Ruminococcaceae Faecalibacterium OTU62 26.6 23.1 3.6E-14 Bacteroidetes Prevotellaceae Prevotella OTU240 23.9 23.0 4.2E-14 Firmicutes Ruminococcaceae Butyricicoccus OTU164 16.5 22.5 1.3E-13 Firmicutes Lachnospiraceae Blautia OTU34 16.4 22.5 1.3E-13 Firmicutes Streptococcaceae Lactococcus OTU31 24.8 22.2 2.1E-13 Actinobacteria Bifidobacteriaceae Bifidobacterium OTU12 6.9 21.3 2.2E-12 Actinobacteria Bifidobacteriaceae Bifidobacterium OTU140 6.9 21.3 2.2E-12 Bacteroidetes Bacteroidaceae Bacteroides OTU167 3.8 20.4 1.6E-11 Firmicutes Lactobacillaceae Lactobacillus OTU706 3.7 20.4 1.7E-11 Actinobacteria Brevibacteriaceae Brevibacterium OTU6 1273.1 8.9 6.2E-03 Actinobacteria Microbacteriaceae NA OTU14 13.3 -7.5 3.3E-02 Actinobacteria Bifidobacteriaceae Bifidobacterium OTU107 6.2 -22.6 1.2E-13 Firmicutes Streptococcaceae Lactococcus OTU189 9.1 -23.1 3.7E-14 Firmicutes Veillonellaceae Veillonella OTU88 13.9 -23.5 1.5E-14 Bacteroidetes Prevotellaceae Prevotella OTU365 27.7 -24.6 1.6E-15 Bacteroidetes Bacteroidaceae Bacteroides OTU23 57.0 -25.6 2.4E-16 Actinobacteria Atopobiaceae Olsenella aMean abundance values are the mean abundance for each OTU among all Mh and BT + Mh samples. bPositive fold change values indicate OTUs that were enriched in the Mh group and negative values indicate OTUs enriched in the BT

+ Mh calves. 289

Supplementary Table S5.5 Spearman rank-based correlation coefficients of relative abundance of top 10 most abundant genera within the nasal microbiota of calves (n =12) that received an intranasal inoculation of M. haemolyticaa

Time Mannheimi Moraxell Lactococcu Acinetobacte Bifidobacteriu Streptococcu Lactobacillu Prevotell Bacteroide Klebsiell a a s r m s s a s a

Time 1 -0.1583 0.0368 -0.1851 -0.2149 -0.0589 -0.2022 -0.0294 0.2295 0.2828 -0.0733

Mannheimia - 1 -0.0978 -0.309 -0.0997 -0.2027 0.0804 -0.1091 0.141 -0.0093 -0.0625 0.1583

Moraxella 0.0368 -0.0978 1 0.1089 0.3531 0.0624 0.2139 0.0428 0.2024 0.1187 0.3395

Lactococcus - -0.309 0.1089 1 0.4001 0.5465 0.2448 0.3124 -0.1989 0.1383 0.1701 0.1851

Acinetobacter - -0.0997 0.3531 0.4001 1 0.5516 0.3503 0.4798 0.1439 0.0308 0.5915 0.2149

Bifidobacteriu - -0.2027 0.0624 0.5465 0.5516 1 0.2454 0.6726 0.1638 0.2107 0.2911 m 0.0589

Streptococcus - 0.0804 0.2139 0.2448 0.3503 0.2454 1 0.2052 0.1862 0.2082 0.2037 0.2022

Lactobacillus - -0.1091 0.0428 0.3124 0.4798 0.6726 0.2052 1 0.1556 0.119 0.427 0.0294

Prevotella 0.2295 0.141 0.2024 -0.1989 0.1439 0.1638 0.1862 0.1556 1 0.2417 0.1555

Bacteroides 0.2828 -0.0093 0.1187 0.1383 0.0308 0.2107 0.2082 0.119 0.2417 1 0.1509

Klebsiella - -0.0625 0.3395 0.1701 0.5915 0.2911 0.2037 0.427 0.1555 0.1509 1 0.0733 aCorrelation analysis was performed on the relative abundance data obtained from all the nasal swabs (n = 91) collected over the course of study.

290

Supplementary Table S5.6 Spearman rank-based correlation coefficients of relative abundance of top 10 most abundant genera within the nasal microbiota of calves (n =12) received an intranasal inoculation of bacterial therapeutics and M. haemolyticaa.

Time Mannheimi Moraxell Lactococcu Acinetobacte Bifidobacteriu Streptococcu Lactobacillu Prevotell Bacteroide Klebsiell a a s r m s s a s a

Time 1 -0.0108 -0.0267 -0.3009 -0.3347 -0.0345 -0.2279 -0.163 0.2615 0.3979 -0.2904

Mannheimia - 1 -0.0203 -0.2106 -0.1004 -0.169 0.148 -0.1554 0.1795 -0.085 0.033 0.0108

Moraxella - -0.0203 1 0.0077 -0.015 0.0158 0.0597 0.0356 0.0424 0.174 0.0309 0.0267

Lactococcus - -0.2106 0.0077 1 0.3771 0.2681 0.0933 0.1413 -0.3348 0.1138 0.4044 0.3009

Acinetobacter - -0.1004 -0.015 0.3771 1 0.6 0.3774 0.5244 0.0918 0.1516 0.639 0.3347

Bifidobacteriu - -0.169 0.0158 0.2681 0.6 1 0.3232 0.5408 0.2765 0.4676 0.3511 m 0.0345

Streptococcus - 0.148 0.0597 0.0933 0.3774 0.3232 1 0.4223 0.2441 0.1425 0.3163 0.2279

Lactobacillus -0.163 -0.1554 0.0356 0.1413 0.5244 0.5408 0.4223 1 0.2022 0.217 0.3205

Prevotella 0.2615 0.1795 0.0424 -0.3348 0.0918 0.2765 0.2441 0.2022 1 0.3599 0.1123

Bacteroides 0.3979 -0.085 0.174 0.1138 0.1516 0.4676 0.1425 0.217 0.3599 1 0.2134

Klebsiella - 0.033 0.0309 0.4044 0.639 0.3511 0.3163 0.3205 0.1123 0.2134 1 0.2904 aCorrelation analysis was performed on the relative abundance data obtained from all the nasal swabs (n = 88) collected over the course of study. 291

Supplementary Table S5.7 Squared multiple correlations in path models fitted to data from the relative abundance of top 10 genera of nasal microbiota in calves that received an intranasal inoculation of either only M. haemolytica (Model 1), or bacterial therapeutics and M. haemolytica

(Model 2).

Relative abundance 2 2 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 1 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑙𝑙 2 Acinetobacter 𝑅𝑅0.5147 𝑅𝑅0.4083

Bacteroides 0.1752 0.4494

Bifidobacterium 0.4524 0.4786

Lactococcus 0.4100 0.1944

Mannheimia 0.1621 .

Prevotella 0.1257 0.1171

Streptococcus 0.1190 0.2484

Klebsiella . 0.0843

Lactobacillus . 0.1027

292

Supplementary Table S6.1 Animal health records.

Rectal Animal Treatment Treatment Antibiotic # of antibiotic Day of temperature ID group Date treatment treatments experiment (°C ) 60 BT 40.1 19/11/18 Micotil 1 4 2 BT 39.7 20/11/18 Micotil 1 5 1 CTRL 40.6 17/18/11 Micotil 1 2 36 CTRL 39.8 19/11/18 Micotil 1 4 6 CTRL 39.7 4/12/19 Micotil 1 18

293

Supplementary Table S6.2 Standardized total effect (mean ± SE) from predictor variables (listed on the columns) on the dependent variables (listed on the rows) (Model 1, BT group).

Ace Aci Ato Chr Lac Man Pha Pse Psy Rum Day His Pas 0.10 ± 0.32 ± 0.24 ± 0.42 ± 0.68 ± 0.19 ± Ace 0 0.025 0.07 0 0.009 0 0 0.044 0.053 0.042 -0.05 ± 0.051 0 -0.08 ± 0.026 p < .0001 p < .0001 p = 0.0073 p < .0001 p < .0001 p < .0001 p = 0.3016 p = 0.0034 0.24 ± 0.08 ± Aci 0 0 0 0 0.062 0 0 0.026 0 0 -0.02 ± 0.007 0 0 p = 0.0001 p = 0.0038 p = 0.0405 0.20 ± 0.11 ± 0.07 ± 0.03 ± 0.18 ± 0.28 ± 0.36 ± 0.10 ± 0.18 ± All 0.081 0.027 0.029 0 0.009 0 0.072 0.045 0.072 0.029 -0.25 ± 0.068 0.065 -0.04 ± 0.016 p = p = 0.0001 p = 0.0284 p = 0.0076 p = 0.0135 p < .0001 p < .0001 p = 0.0005 p = 0.0001 p = 0.0048 p = 0.0099 0.0116 0.13 ± 0.03 ± 0.54 ± 0.88 ± 0.24 ± Ato 0 0.031 0 0 0.011 0 0 0.042 0.05 0.052 -0.41 ± 0.053 0 -0.10± 0.034 p < .0001 p = 0.0064 p < .0001 p < .0001 p < .0001 p < .0001 p = 0.0034 0.28 ± 0.22 ± 0.10 ± 0.14 ± 0.24 ± 0.34 ± 0.20 ± 0.40 ± Chr 0 0.044 0.075 0 0.025 0.053 0.064 0.043 0.067 0.067 -0.34 ± 0.055 0 -0.02 ± 0.011 p < .0001 p = 0.0029 p = 0.0002 p = 0.0087 p = 0.0002 p < .0001 p = 0.0033 p < .0001 p < .0001 p = 0.0034 0.15 ± 0.06 ± 0.27 ± 0.01 ± -0.12 ± 0.37 ± 0.50 ± 0.23 ± Jeo 0 0.029 0.025 0.066 0.02 0.063 0.10 ± 0.03 0.043 0.058 0.041 -0.01 ± 0.045 0 -0.06 ± 0.020 p < .0001 p = 0.0168 p < .0001 p = 0.5419 p = 0.0632 p = 0.0005 p < .0001 p < .0001 p < .0001 p = 0.904 p = 0.0047 0.32 ± Lac 0 0 0 0 0 0 0 0.072 0 0 -0.06 ± 0.026 0 0 p < .0001 p = 0.0137 -0.08 ± 0.19 ± -0.26 ± -0.02 ± 0.003 ± Man 0 0.03 0 0 0.076 0 0.075 0.039 0 0 0.008 0 0 p = 0.0118 p = 0.0134 p = 0.0004 p = 0.6922 p = 0.6948 Myc 0 0 0 0 0 0 0 0 0 0 0.33 ± 0.071 0 0 p < .0001 0.29 ± 0.07 ± 0.31 ± Pha 0 0.078 0 0 0.026 0 0 0.075 0 0 -0.06 ± 0.025 0 0 p = 0.0002 p = 0.0077 p < .0001 p = 0.0164 Pse 0 0 0 0 0 0 0 0 0 0 -0.20 ± 0.066 0 0 p = 0.0026 0.15 ± 0.04 ± 0.62 ± 0.28 ± Psy 0 0.035 0 0 0.013 0 0 0.043 0 0.056 0.19 ± 0.057 0 -0.11± 0.038 p < .0001 p = 0.0063 p < .0001 p < .0001 p = 0.0008 p = 0.0029

294

0.32 ± 0.06 ± 0.25 ± 0.08 ± 0.04 ± 0.06 ± 0.28 ± 0.05 ± 0.58 ± Rik 0 0.044 0.024 0.069 0.023 0.016 0.023 0.039 0.022 0.054 -0.12 ± 0.032 0 -0.01 ± 0.003 p < .0001 p = 0.021 p = 0.0002 p = 0.0002 p = 0.033 p = 0.0095 p < .0001 p = 0.0216 p < .0001 p = 0.0001 p = 0.0684 0.53 ± 0.13 ± 0.40 ± Rum 0 0.054 0 0 0.035 0 0 0.056 0 0 -0.08 ± 0.029 0 0 p < .0001 p = 0.0004 p < .0001 p = 0.0062 Genera: Ace, Acetitomaculum; Aci, Acinetobacter; All, Alloprevotella; Ato, Atopostipes; Chr, Christensenellaceae_R7_group;His,

Histophilus; Jeo, Jeotgalibaca; Lac, Lactobacillus; Man, Mannheimia; Myc, Mycoplasma; Pas, Pasteurella; Pha, Phascolarctobacterium;

Pse, Pseudomonas; Psy, Psychrobacter; Rik, Rikenellaceae_RC9_gut_group; Rum, Ruminococcaceae_UCG005

295

Supplementary Table S6.2 Continue (Model 2, CTRL group).

Ace All Ato Chr Jeo Lac Pse Psy Rik Rum Day His Ace 0 0.08 ± 0.026 0 0.08 ± 0.032 0.28 ± 0.073 0.30 ± 0.100 0.41 ± 0.070 0.16 ± 0.037 0 0.48 ± 0.055 0.03 ± 0.019 0 p = 0.0020 p = 0.014 p = 0.0001 p = 0021 p <.0001 p <.0001 p <.0001 p = 0.0705

Aci 0 0.09 ± 0.030 0 0 0 0.03 ± 0.011 0 0.33 ± 0.069 0 0.55 ± 0.068 0.12 ± 0.029 0 p = 0.0024 p = 0.0163 p <.0001 p <.0001 p <.0001

All 0 0 0 0 0 0.28 ± 0.069 0 0.52 ± 0.057 0 0 0.18 ± 0.032 0 p <.0001 p <.0001 p <.0001 Ato 0 0.13 ± 0.038 0 0.18 ± 0.49 0 0.04 ± 0.014 0.22 ± 0.052 0.07 ± 0.022 0.96 ± 0.044 0.74 ± 0.037 -0.30 ± 0.050 0 p = 0.0010 p = 0.0002 p = 0.0118 p <.0001 p = 0.0024 p <.0001 p <.0001 p <.0001

Chr 0 0.11 ± 0.035 0 0.11 ± 0.0002 0 0.04 ± 0.012 0 0.06 ± 0.020 0 0.65 ± 0.044 -0.23± 0.057 0 p = 0.0014 p = 0.0131 p = 0.003 p <.0001 p <.0001

Jeo 0 0.07 ± 0.025 0 0 0 0.98 ± 0.230 0 0.44 ± 0.067 0 0.43 ± 0.071 0.15 ± 0.032 0 p = 0.0040 p <.0001 p <.0001 p <.0001 p <.0001

Lac 0 0 0 0 0 0 0 0.41 ± 0.065 0 0 0.14 ± 0.03 p <.0001 p <.0001 Man -0.29 ± 0.072 -0.02 ± 0.010 0 -0.02 ± 0.011 -0.08 ± 0.030 -0.084 ± 0.035 -0.12 ± 0.037 -0.04 ± 0.016 0 -0.13 ± 0.039 (-0.01 ± 0.006) 0 p < .0001 p = 0.016 p = 0.0386 p = 0.0063 p = 0.0162 p = 0.0014 p = 0.0048 p = 0.0005 p = 1017

Myc 0 -0.03 ± 0.015 0 -0.45 ± 0.022 0 -0.01 ± 0.005 -0.24 ± 0.075 -0.02± 0.016 0 -0.21± 0.067 0.01 ± 0.006 0 p = 0.0238 p = 0.0395 p = 0.05 p = 0.0016 p = 0.0303 p = 0.002 p =0.379

Pas 0 -0.03 ± 0.014 -0.25 ± 0.076 -0.45 ± 0.019 0 -0.01 ± 0.004 -0.06 ± 0.021 -0.02 ± 0.007 -0.24 ± 0.074 -0.19 ± 0.058 -0.20 ± 0.075 0 p = 0.02 p = 0.001 p = 0.0144 p = 0.0451 p = 0.001 p = 0.0263 p = 0.0011 p = 0.0012 p = 0.0092 Pha 0 0.10 ± 0.031 0 0.04 ± 0.023 0 0.03 ± 0.011 0.22 ± 0.086 0.05 ± 0.018 0 0.59 ± 0.055 -0.17 ± 0.064 0 p = 0.0017 p = 0.0609 p = 0.0141 p = 0.001 p = 0.0035 1 p <.0001 p = 0.0063 Pse 0 0.15 ± 0.045 0 0.19 ± 0.071 0 0.04 ± 0.016 0 0.08 ± 0.025 0 0.87 ± 0.05 -0.02 ± 0.023 0 p = 0.0011 p = 0.0064 p = 0.0141 p = 0.0024 p <.0001 p = 0.3606 Psy 0 0 0 0 0 0 0 0 0 0 0.35 ± 0.048 0 p <.0001 Rik 0 0.13 ± 0.040 0 0.19 ± 0.051 0 0.04 ±0.015 0.23 ± 0.054 0.07 ± 0.022 0 0.78 ± 0.034 -0.02 ± 0.019 0 p = 0.0009 p = 0.0002 p = 0.0114 p <.0001 p = 0.0021 p <.0001 p = 0.2266 Rum 0 0.17 ± 0.051 0 0 0 0.05 ± 0.019 0 0.09 ± 0.029 0 0 0.03 ± 0.011 0 p = 0.0001 p = 0.0117 p = 0.0021 p = 0.0073 296

Supplementary table S6.2. Continue (Model 3, MP group).

At Ma Pa Ps Ac Al Ch Day Hi Ph Ps Ri Ru Ac 0.75 ± 0.080 0 0 0 0 0 0 0.29 ± 0.063 0 0 0 0 0.6251 ± 0.068 p <.0001 p < .0001 p < .0001

At 0 0 0 0 0 0 0 -0.36 ± 0.051 0 0 0 0 0.83 ± 0.049 p < .0001 p < .0001

Je 0.10 ± 0.037 0 0 0.39 ± 0.074 0 0 0 0.16 ± 0.039 0 0 0 0 0.50 ± 0.060 p= 0.0068 p <.0001 p < .0001 p < .0001

La 0 -0.21 ± 0.073 0 0.18 ± 0.079 0 0.20 ± 0.079 0.045 ± 0.023 0 0 0 0 0 p = 0.0041 p = 0.0208 p = 0.0132 p = 0.0471

Ma 0 0 0 0 0 0 0 -0.19 ± 0.077 0 0 0 0 0 p = 0.0165

My -0.48 ± 0.055 -0.16 ± 0.057 0 0 0 0 0 0.02 ± 0.035 0 0 0 -0.35 ± 0.058 -0.40 ± 0.053 p <.0001 p = 0.0053 p < .0001 p < .0001 p < .0001

Pa 0 0 0 0 0 0 0 -0.22 ± 0.076 0 0 0 0 0 p = 0.005

Ps 0.26 ± 0.081 0 0 0 0 0 0 0.42 ± 0.055 0 0 0 0 0.60 ± 0.051 p = 0.0014 p < .0001 p < .0001

297

Supplementary Table S6.3 Standardized total effects from Lactobacillus to the abundances of other bacteria.The standardized total effects of Lactobacillus are shown in Table 2 for Model 1

(BT group) and Model 2 (CTRL group). Lactobacillus was not a predictor variable in Model 3

(MP group).

Model 1 (BT) Model 2 (CTRL) Acetitomaculum 0.0237 ± 0.008837 0.2949 ± 0.1018 p = 0.007316 p = 0.003762

Acinetobacter 0.2379 ± 0.0615 0.0257 ± 0.0107 p = 0.000108 p = 0.0166

Alloprevotella 0.025 ± 0.009363 0.2794 ± 0.071 p = 0.007598 p <.0001

Atopostipes 0.0305 ± 0.0112, 0.0348 ± 0.0138 p = 0.006419 p = 0.0116

Christensenellaceae_R7_group 0.0946 ± 0.0251 0.0307 ± 0.0124 p 0.00016 p = 0.0136

Jeotgalibaca 0.0119 ± 0.0196 0.9835 ± 0.2532 p = 0.5419 p = 0.000102

Mannheimia 0.1875 ± 0.0758 -0.0844 ± 0.0367 p = 0.0134 p = 0.0216

Mycoplasma 0 -0.00966 ± 0.004982 p = 0.0525

Pasteurella 0 -0.0883 ± 0.004454 p = 0.0474

Phascolarctobacterium 0.0688 ± 0.0258 0.0275 ± 0.0113 p = 0.007703 p = 0.0149

Pseudomonas 0 0.0409 ± 0.0163 p = 0.0123

Psychrobacter 0.0349 ± 0.0128 0 p = 0.006275

Rikenellaceae_RC9_gut_group 0.0833 ± 0.0225 0.0366 ± 0.0145 p = 0.00021 p = 0.0116

Ruminococcaceae_UCG005 0.1251 ± 0.0352 0.047 ± 0.0186 p = 0.00038 p = 0.0115

298

Appendix B: Supplementary Figures

Supplementary Figure S3.1 Stacked bar chart of the 15 most abundant genera in the nasopharyngeal microbiota of cattle (n = 13) by sampling time.

299

Supplementary Figure S3.2 Correlations between LAB families and the Pasteurellaceae family in the nasopharyngeal microbiota of cattle (n = 13) by sampling time.

The Spearman’s correlation coefficient (blue line bar) for the correlations of each LAB family with

Pasteurellaceae was calculated based on the relative abundance of these families observed on each sampling time using Spearman’s rank correlation method. The dotted lines indicate where r

(actually Pearson’s correlation) would be statistically significant with P = 0.05. These are calculated based on t values of approximately 1.96 and n of 13. Therefore the lines are located at

+/- 0.51.

300

Supplementary Figure S5.1 An example photo showing the intranasal inoculation of bacterial therapeutics to a dairy calf.

301

Supplementary Figure S5.2 Lung gross pathology of the dairy calves (n =12) that received an intranasal inoculation of either M. haemolytica and PBS (Mh) or bacterial therapeutics and M. haemolytica (BT + Mh).

A) Proportion of animals had lung lesions; B) The severity of the lung lesions measured by lung lesion scoring; The bar plot represents the mean percent of the lung tissue affected. Error bars

302

indicate ± standard error of the meanC) Example of a healthy lung from a BT + Mh animal (calf#

07-R2) (right) and a lung from a Mh animal (Calf # 03-R2) displaying lesions (left, dark color)

303

Supplementary Figure S5.3 Phylogenetic tree of pulsed-field gel electrophoresis profiles of M. haemolytica isolates. Not all confirmed M. haemolytica isolates were typed.

304

Supplementary Figure S5.4 A representative example of (GTG)5 Rep PCR fingerprints of

Lactobacillus bacteria isolated from the nasal cavity of challenged dairy that calves received intranasal bacterial therapeutics and M. haemolytica.

Fragment sizes are indicated on the left of the image, for control markers (1Kb plus ladder). Strains are listed above each (GTG)5 pattern: 103C, 72B, 63A, 86D, 57A, and 3E were bacterial therapeutics described in methods. For other strains, animal number with strain ID (N) followed by the replicate number (R) and day of isolation (d) are indicated.

305

Supplementary Figure S6.1 Average daily gain of the cattle received either intranasal bacterial

therapeutics (BT), PBS (CTRL) or subcutaneous metaphylaxis (MP) (n = 20 per group) over the course of 42 days.

The bar plot represents the mean average daily gain. Error bars indicate ± standard error of the mean.

306

Supplementary Figure S6.2 The relative abundance of families within the order Lactobacillales in the nasopharyngeal microbiota of calves received either intranasal bacterial therapeutics (BT),

PBS (CTRL) or subcutaneous tulathromycin (MP) (n = 20 per group) over the course of 42 days.

307

Supplementary Figure S6.3 The relative abundance of different Lactobacillus species within

Lactobacillus genus determined by 16S rRNA amplicon sequencing in nasopharyngeal microbiota of calves received either intranasal bacterial therapeutics (BT), PBS (CTRL) or subcutaneous tulathromycin (MP) (n = 20 per group) over the course of 42 days.

308

Supplementary Figure S6.4 Relative abundance of genera associated with bovine respiratory disease nasopharyngeal microbiota of calves received either intranasal bacterial therapeutics (BT),

PBS (CTRL) or subcutaneous tulathromycin (MP) (n = 20 per group) over the course of 42 days.

309

Appendix C: Copyright permissions

1. As the co-authors that contributed to the paper “Amat S, Subramanian S, Timsit E,

Alexander TW. (2017) Probiotic bacteria inhibit the bovine respiratory pathogen

Mannheimia haemolytica serotype 1 in vitro. Letters in Applied Microbiology, 64: 343-

349”, we permit using this paper as Chapter 2 of Samat Amat’s thesis entitled

"Development of Intranasal Bacterial Therapeutics to Mitigate the Bovine Respiratory

Pathogen Mannheimia haemolytica" that will be submitted to the Faculty of Graduate

Studies at the University of Calgary in November 2019.

Co-Author Signature Date

Sangeetha Subramanian

Edouard Timsit

Trevor W. Alexander

310

2. As the co-authors that contributed to the paper “Amat S, Holman DB, Timsit E,

Schwinghamer T, Alexander TW (2019) Evaluation of the nasopharyngeal microbiota in

beef cattle transported to a feedlot, with a focus on lactic acid-producing bacteria. Frontiers

in Microbiology. 10:1988”, we permit using this paper as Chapter 3 of Samat Amat’s thesis

entitled "Development of Intranasal Bacterial Therapeutics to Mitigate the Bovine

Respiratory Pathogen Mannheimia haemolytica" that will be submitted to the Faculty of

Graduate Studies at the University of Calgary in November 2019.

Co-Author Signature Date

Devin B. Holman

Edouard Timsit

Timothy Schwinghamer

Trevor W. Alexander

311

3. As the co-authors that contributed to the paper “Amat S, Timsit E, Baines D, Yanke J,

Alexander TW. (2019). Development of Bacterial Therapeutics Against the Bovine

Respiratory Pathogen Mannheimia haemolytica. Applied and Environmental Microbiology

AEM.01359-19”, we permit using this paper as Chapter 4 of Samat Amat’s thesis entitled

"Development of Intranasal Bacterial Therapeutics to Mitigate the Bovine Respiratory

Pathogen Mannheimia haemolytica" that will be submitted to the Faculty of Graduate

Studies at the University of Calgary in November 2019.

Co-Author Signature Date

Edouard Timsit

Danica Baines

Jay Yanke

Trevor W. Alexander

312

4. As the co-authors that contributed to the paper “Amat S, Alexandera TW, Holman DB,

Schwinghamer T, Timsit E. Intranasal bacterial therapeutics reduce colonization by the

respiratory pathogen Mannheimia haemolytica in dairy calves” submitted to “mSystems”

(manuscript number mSystems00629-19, under minor revision), we permit using this paper

as Chapter 5 of Samat Amat’s thesis entitled "Development of Intranasal Bacterial

Therapeutics to Mitigate the Bovine Respiratory Pathogen Mannheimia haemolytica" that

will be submitted to the Faculty of Graduate Studies at the University of Calgary in

November 2019.

Co-Author Signature Date

Trevor W. Alexander

Devin B. Holman

Timothy Schwinghamer

Edouard Timsit

313