I Haemophilus Pathogenesis During Otitis Media: Influence of Nutritional Immunity on Bacterial Persistence and Intracellular

Haemophilus pathogenesis during otitis media: Influence of nutritional immunity on bacterial persistence and intracellular lifestyles

Dissertation

Presented in partial fulfillment of the Requirements for the Degree of Doctor of Philosophy In the Graduate School of The Ohio State University

Rachael L. Hardison B.S.

Biomedical Sciences Graduate Program

The Ohio State University

2018

Dissertation Committee:

Kevin Mason, PhD, Co-Advisor

Sheryl Justice, PhD, Co-Advisor

Jennifer Edwards, PhD

Stephanie Seveau, PhD, Chair

John Gunn, PhD

Rachael Hardison

2018

ABSTRACT

Nontypeable Haemophilus influenzae (NTHI) is a gram-negative opportunistic pathogen that is a major cause of otitis media (OM) and other diseases of the upper and lower respiratory tract. As a commensal, NTHI resides in the human nasopharynx, an environment with sufficient levels of nutrients to allow colonization. Permissive risk factors, such as a preceding viral infection, can allow NTHI to migrate up the Eustachian tube and into the sterile middle ear, an environment that is initially depleted of nutrients and prone to a robust and rapid inflammatory response. Nutritional immunity occurs when the host tightly sequesters essential nutrients, such as heme-iron, in sterile areas like the middle ear. As NTHI is auxotrophic for heme-iron, the ability for NTHI to sense, respond, and adapt to shifts in heme-iron availability is important for NTHI survival during the transition from a commensal to an opportunistic pathogen and in the progression of OM. We have previously described that transient heme-iron restriction of NTHI promotes diverse phenotypes, including changes in biofilm formation, increased survival of NTHI and altered disease severity in a preclinical model of OM, and the transition to an intracellular lifestyle for NTHI. Specifically, transient heme-iron restriction promotes intracellular bacterial community (IBC) formation by NTHI in epithelial cells in vitro. However, the host and bacterial responses to nutrient restriction that impact intracellular fate and persistence of

NTHI in chronic OM are unknown.

This work demonstrates a role for transient heme-iron restriction in inducing persistence through adaptation to two relatively underexplored NTHI lifestyles: long-term stationary phase persistence and survival in the host cell as a cytosolic pathogen. In this study, we gained mechanistic insight into NTHI persistence through the use of an in vitro model of long-term stationary phase growth. Transient heme-iron restriction potentiates extended survival of NTHI for weeks in vitro. We hypothesized that transient heme-iron restriction induces genetic alterations that contribute to persistence of NTHI. By performing whole genome sequencing of NTHI isolates from long-term culture, we identified microevolution of NTHI in response to transient nutrient limitation through mutation of icc in two independent in vitro experiments. Mutation in icc results in decreased

3',5'-cyclic adenosine monophosphate phosphodiesterase activity that is associated with increased competence for NTHI and survival of NTHI within middle ear mucosal tissue in

IBCs that persist during chronic OM.

This study provides the first evidence for the role of NTHI adaptation to nutrient limitation in promoting the formation of IBCs in vitro and in vivo in a preclinical model of

OM. We hypothesized that transient heme-iron restriction of NTHI resulted in altered uptake into epithelial cells leading to escape or evasion of the endolysosomal pathway.

Using colocalization studies and pharmacological inhibition of endocytosis, we demonstrated a role for entry by macropinocytosis and subsequent escape or evasion of the endolysosomal pathway in NTHI IBC formation in epithelial cells. Further, inhibition of macropinocytosis altered the intracellular fate of transiently restricted NTHI for degradation in the endolysosomal pathway. Blocking macropinocytosis reduced the number of IBCs in cultured middle ear epithelial cells, providing evidence for the feasibility

iii of this approach to reduce OM persistence. Collectively, the results from these studies reveal that transient heme-iron restriction induces genetic and phenotypic changes that promote NTHI competence, a transition to persistence in a long-term stationary phase state, and a novel lifestyle as a cytosolic pathogen in IBCs leading to new mechanisms for survival during disease progression. The new data generated herein will direct future studies toward investigation of a role for NTHI IBCs in driving recurrent OM.

DEDICATION

To my daughter, Evie. You are my little bird and nothing I accomplish or do in life will ever be better or more important than being your mom. I hope this inspires you to challenge yourself, believe in who you are, and to never put limits on what you can do or who you can be. You are strong, and full of life. I love you.

ACKNOWLEDGMENTS

“And whatever you do, whether in word or deed, do it all in the name of the Lord Jesus,

giving thanks to God the Father through him.” – Colossians 3:17 (NIV)

All of the thanks in the world go directly to my husband, Zach. Your support and love has been the strength that I needed many, many times over the past few years.

This has truly been a partnership, and you have been with me for every step. Thank you for moving to Columbus with me, for putting your career on hold at times to support me in this endeavor. Thank you for always challenging me to be better, for being my friend through many struggles and celebrations. Thank you for all the times you went above and beyond as a husband and as a father to support our family. Thanks for making so many toddler dinners and for all of the times you solo parented on weekends so that I could go into the lab for a while. You are my rock and you inspire me every day. I love you always.

Thank you to my daughter, Evie, for being my inspiration and joy every day. I am so incredibly lucky that you came into my life and have been on this journey with me.

Thank you for approaching life with such curiosity and wonder and teaching me to do the

vi same. Thank you for being my reason, and for your sweet love and snuggles at the end of every day. I will always cherish that you got to go on this journey with me, and were my partner during my candidacy exam at 33 weeks pregnant.

I would like to extend my sincere thanks and gratitude to my co-advisors, Dr.

Kevin Mason and Dr. Sheryl Justice. Kevin, your enthusiasm and excitement for science and research is so evident every day, and I have been so fortunate to learn, grow, and be challenged as a member of your lab. Thank you for always challenging me in experimental design, my writing, and my presentation and communication of my science.

Sheryl, I will always look up to you as an excellent example of a woman in science, and I am inspired by you every day. Thank you for challenging and supporting me through all of our conversations, whether it was while doing microscopy, in your office, or while sharing a room when traveling for conferences. Thank you both for your constant support and direction these past years, and, thank you for continuing to give me direction even when it was not something I wanted to hear. I will always be grateful for that. I am fortunate to have been given the opportunity to work on such a fun project and to become a much better scientist thanks to mentorship from both of you.

Thank you to the members of my committee: Dr. Stephanie Seveau, Dr. John

Gunn, and Dr. Jennifer Edwards. Thank you for your guidance over my project and my development as a scientist. You demonstrated true care for me not only as a scientist and graduate student, but also as a person and new mother during my graduate career.

Thank you always keeping my best interest in mind and for guiding me towards the best outcomes. I am so grateful that I had the opportunity to learn from all three of you.

vii

I want to extend special thanks to Dr. Alistair Harrison, for being the best lab mate I could ask for, for always playing good music, and for being my friend. I will always look up to you as a scientist. Your mentorship, guidance, and thoughtful conversation has had a tremendous impact on my own development as a scientist. Thank you for always listening and for never making me feel like I was asking dumb questions. Also, thank you for all of the times you did chinchilla things for me because I was allergic, for all of the times you helped me out on weekends, and for taking care of the plant.

I would like to thank all of the members of the Mason/Justice lab that I was so fortunate to overlap with over the years: Derek, Liz, Rachel, Meghan, your friendship will always be a bright spot in my graduate career. Thank you to the members of the Center for Microbial Pathogenesis for always being willing to listen, help, and learn. Special thanks are due to Ken, Anirudh, and Dr. Sam King for always being so supportive and providing thoughtful insight and feedback on my project.

Special thanks to our collaborators: Dr. Heather Pinkett, the Duke Proteomics and Metabolomics Core, Dr. Wandy Beatty, and the Genomics core here at Nationwide

Children’s. A special shout out also to the NCH morphology core for processing so many samples for me, and to the vivarium staff for taking such great care of our animals. I would also like to thank my funding sources, particularly the RITA Trainee Fellowship and the Infectious Disease Institute T32 Fellowship for providing me with the opportunity to complete these exciting studies and to grow as an independent scientist.

viii

To Vidhya and Abbie, I can’t think of anyone else I’d rather go through graduate school with. Your friendship means the world and I will miss you both dearly. I can’t wait to see the positive impact that you will both make in the biomedical field, and on society as a whole. Cassie, thank you for your constant encouragement and for the sweet scripture texts. You have been such a blessing to me and I am so thankful for you.

A huge amount of thanks and love go to my family. I would not be where I am today without all of the love and support from my family. Mom, thank you for always answering the phone, for your constant presence and love, and for always believing in me. Thank you for modeling what it looks like to succeed as a working mom without compromise, and for showing me what it means to work hard and never give up. Thank you for always being there no matter what else you had going on. Thank you for all of the prayers you said for me over the course of my PhD, for all the times that you listened to everything I was stressed and worried about, and for inspiring me to always keep going. Dad, what an incredible experience it has been to both be getting our PhDs at the same time! Thanks for both commiserating and celebrating with me throughout my graduate career, and for being an amazing example of someone who pursues their dreams at all stages of life. You’re amazing. To Doug, thank you for being a huge support and encouragement to me these past years. Abigail, thank you for being my best friend and always encouraging me, praying for me, and laughing with me over the past few years. Tim and Abigail, I will never forget that time that you specifically prayed over me for my candidacy exam. I love you both. Thank you to Jennifer, Jeff, Nathan,

Kendra, Dawn, and all of my extended family for going on this journey with me and always asking me such fun questions about my project. I love you all.

Special thanks and mentions to: the entire JTT family for taking such wonderful care of Evie while I was pursing my PhD – knowing that she was safely cared for and that she was always having such a great time at daycare was a huge weight off of my shoulders that allowed me to focus on my project and research while in the lab; my One

Church family (especially my connect group) for being a place of encouragement and peace every week; Dr. Cynthia Canan, for providing me with such a great experience at

COSI for my T32 externship; the entire Office of Technology Commercialization at

Nationwide Children’s Hospital, especially Matt, Susannah, Susan, Kyle, Andrew, and

Margaret, for giving me the best internship experience the past few years, for your support, and for teaching me everything you know about tech transfer, and for just being generally so much fun to work with; finally, to Elizabeth Shannon, for being there every time I call and for buoying me so many times – you are the model of friendship and I love you.

VITA

1990…………………………………………………………………………….Born - Salem, VA

2012…………………………...... B.S. Biology, Xavier University

2014 – present……………...... Graduate Research Associate, The Ohio State University

Publications

Hardison RL, Heimlich DR, Harrison A, Beatty WL, Rains S, Moseley MA, Thompson

JW, Justice S, and Mason KM. (2018) Transient nutrient deprivation promotes macropinocytosis-dependent intracellular bacterial community development. mSphere.

3:e00286-18. https://doi.org/10.1128/mSphere.00286-18.

Hardison RL*, Harrison A*, Wallace RM, Heimlich DR, O'Bryan ME, Sebra RP, Pinkett

HW, Justice SS, and Mason KM. (2018) Microevolution in response to transient heme- iron restriction enhances intracellular bacterial community development and persistence.

PloS Pathog 14(10): e1007355. http://doi.org/10.1371/journal.ppat.1007355. *Authors contributed equally to this work

Harrison A, Dubois L, John-Williams L, Mosely MA, Hardison RL, Heimlich DR,

Stoddard A, Kerschner J, Justice SS, and Mason KM. (2015) Comprehensive proteomic and metabolomic signatures of nontypeable Haemophilus influenzae-induced acute otitis

xi media reveal bacterial aerobic respiration in an immunosuppressed environment. Mol

Cell Proteomics 15(3):1117-38.

Fields of Study

Major Field: Biomedical Sciences

Specialization: Microbial Pathogenesis

xii

TABLE OF CONTENTS

Abstract……………………………………………………………...... ii

Dedication………………………………………………………………………………………...v

Acknowledgements……………………………………………………………………………...vi

Vita………………………………………………………………………………………………..xi

Table of Contents……………………………………………………………………………....xiii

List of Tables…………………………………………………………………………………...xvi

List of Figures………………………………………………………………………………….xvii

List of Abbreviations……………………………………………………………………………xix

Chapter 1: Introduction…………………………………………………………………………1

1.1 Haemophilus background…………………………………………………………1

1.2 Otitis Media…………………………………………………………………………3

1.3 NTHI in other diseases…………………………………………………………….7

1.4 NTHI persistence…………………………………………………………………..9

1.5 Nutritional immunity……………………………………………………………….10

1.6 Heme-utilization by Haemophilus………………………………………………..13

1.7 Bacterial signaling…………………………………………………………………16

1.8 Colonization and adherence……………………………………………………..20

1.9 Biofilms……………………………………………………………………………..22

xiii

1.10 Immune response to NTHI……………………………………………………..25

1.11 NTHI immune evasion strategies……………………………………………...28

1.12 Bacterial entry into host cells…………………………………………………..32

1.13 Intracellular trafficking…………………………………………………………..35

1.14 Intracellular bacterial communities……………………………………………38

1.15 Vaccine approaches against NTHI……………………………………………39

1.16 Remaining Questions…………………………………………………………..40

1.17 Project goals and hypotheses…………………………………………………41

Chapter 2: Microevolution in response to transient heme-iron restriction enhances intracellular bacterial community development and persistence………………………….43

2.1 Introduction………………………………………………………………...... 44

2.2 Materials and methods…………………………………………………………...47

2.3 Results……………………………………………………………………………..57

2.4 Discussion…………………………………………………………………………73

Chapter 3: Transient nutrient deprivation promotes macropinocytosis-dependent intracellular bacterial community development……………………………………………..79

3.1 Introduction………………………………………………………………...... 80

3.2 Materials and methods…………………………………………………………...83

3.3 Results……………………………………………………………………………...99

3.4 Discussion………………………………………………………………………..120

Chapter 4: Cytokine responses to transiently heme-iron restricted NTHI-mediated infection………………………………………………………………………………………..125

4.1 Introduction………………………………………………………………...... 125

4.2 Materials and methods………………………………………………………….127

4.3 Results……………………………………………………………………………131

xiv

4.4 Discussion………………………………………………………………………..135

Chapter 5: General Discussion……………………………………………………………...138

5.1 Research findings………………………………………………………………..138

5.2 cAMP and competence: Implications for NTHI persistence…………………141

5.3 Proteomic insights into RM33 persistence……………………………………144

5.4 NTHI adaptation to an intracellular lifestyle…………………………………...151

5.5 NTHI as a cytosolic pathogen…………………………………………………..157

5.6 Microevolution of NTHI in chronic infection…………………………………...162

5.7 Concluding Remarks…………………………………………………………....166

References…………………………………………………………………………………….168

Appendix A. Significantly altered proteins in RM33 biofilms……………………………..211

Appendix B. Mutations acquired by NTHI following passaging through the chinchilla.……………………………………………………………………………………....213

Appendix C. Additional methods…………………………………………………………....214

LIST OF TABLES

Table 2.1. Bacterial strains used in this study……………………………………………….49

Table 2.2. Primers used in this study………………………………………………………...52

Table 3.1. Pharmacological compounds used to inhibit endocytosis in this study………90

Table 3.2. Assessment of differentially expressed proteins of NTHI transiently restricted and continuously exposed to heme-iron……………………………………………………114

Table 4.1. Volumes of recovered chinchilla middle ear fluid during experimental OM……………………………………………………………………………………………...130

Table A.1. List of proteins that are significantly altered in biofilms formed by RM33 compared to the parent 86-028NP at 48 hours……………………………………………211

Table B.1. List of mutations identified in NTHI after four subsequent passages through the chinchilla…………………………………………………………………………………..213

Table C.1. Primers used for validation of mutations acquired by NTHI after passaging in the chinchilla…………………………………………………………………………………..221

xvi

LIST OF FIGURES

Figure 2.1. Transient heme-iron restriction promotes a cyclical pattern of viability and microevolution during stationary phase………………………………………………………59

Figure 2.2. Independent validation of cyclical pattern of viability and microevolution…..61

Figure 2.3. In silico structural modeling of Icc……………………………………………….63

Figure 2.4. Biochemical and functional assays confirm reduced cAMP phosphodiesterase activity of mutant Icc…………………………………………………….65

Figure 2.5. RM33 remains mucosa-associated in the absence of middle ear effusion during experimental OM……………………………………………………………………….67

Figure 2.6. NTHI persists in middle ear mucosae as intracellular bacterial communities (IBCs)...... 70

Figure 2.7. In silico comparison of genetic plasticity of icc across NTHI clinical isolates………………………………………………………………………………………….72

Figure 2.8. Phosphodiesterase activity of paired NTHI clinical isolates…………………74

Figure 3.1. Transient heme-iron restriction of NTHI promotes intracellular bacterial community formation in a preclinical model of otitis media………………………………101

Figure 3.2. Transiently restricted NTHI invades and survives within human epithelial cells in intracellular bacterial communities……………………………………………………….103

Figure 3.3. Nutritionally conditioned NTHI is not cytotoxic to NHBE cells………………105

Figure 3.4. Effect of multiplicity of infection on intracellular survival of NTHI…………..106

Figure 3.5. Intracellular localization of nutritionally conditioned NTHI in the presence and absence of permeabilization…………………………………………………………………107

Figure 3.6. Transient heme-iron conditioning of NTHI alters trafficking to early endosomes…………………………………………………………………………………….109

xvii

Figure 3.7. Transient heme-iron restriction of NTHI alters lysosomal trafficking and biogenesis...... 111

Figure 3.8. Pharmacological inhibition of endocytosis pathways reveals that nutritionally conditioned NTHI is internalized into cells through multiple mechanisms………………115

Figure 3.9. Association of nutritionally conditioned NTHI with NHBE cells is unchanged in the presence of pharmacological inhibitors……………………………………………..116

Figure 3.10. Inhibition of macropinocytosis redirects transiently restricted NTHI to the endolysosomal pathway and decreases intracellular survival……………………………119

Figure 3.11. Proposed model for differential trafficking of transiently restricted NTHI through macropinocytosis resulting in IBC formation……………………………………..122

Figure 4.1. In vitro cytokine production by respiratory epithelial cells stimulated by NTHI transiently restricted or continuously exposed to heme-iron……………………………..132

Figure 4.2. Transiently restricted or continuously exposed NTHI cytokine stimulation in the chinchilla middle ear……………………………………………………………………...134

Figure 5.1. Prior transient heme-iron restriction is not required for induction of long-term survival of RM33 in stationary phase culture………………………………………………142

Figure 5.2. RM33 forms biofilms with distinct architecture that remain segregated from the parent in mixed culture…………………………………………………………………..146

Figure 5.3. Enzymes in the tryptophan biosynthesis pathway are significantly increased in biofilms formed by RM33 compared to biofilms formed by the parent strain………..148

Figure 5.4. NTHI IBC formation in middle ear mucosal tissue significantly increases after four sequential periods of infection in a chinchilla model of OM…………………………164

xviii

LIST OF ABBREVIATIONS

AC Adenylate cyclase

ALS-1 Acid labile surfactant

AOM Acute otitis media

AP Antimicrobial peptide

AUC Area under curve

BEGM Bronchial epithelial growth medium cAMP cyclic AMP

CE Continuously exposed

CF Cystic Fibrosis

CFU Colony forming unit

ChoP Phosphorylcholine

CMEE Chinchilla middle ear epithelia

COPD Chronic Obstructive Pulmonary Disorder

CPZ Chlorpromazine

CRP Catabolite repression protein

CytoD Cytochalasin D

DIS Defined iron source

DMSO Dimethyl sulfoxide

DPBS Dulbecco’s Phosphate Buffered Saline

xix eDNA Extracellular DNA

EE Early endosome

EEA1 Early endosomal antigen 1

EIPA 5-(N-ethyl-N-isopropyl)-amiloride

EPS Extracellular polymeric substance

FDR False discovery rate

GASP Growth advantage in stationary phase

G-CSF Granulocyte-Colony Stimulating Factor

GFP Green fluorescent protein

GM-CSF Granulocyte Macrophage-Colony Stimulating Factor

HMW High molecular weight

HRP Horseradish peroxidase

HRV Human Rhinovirus

HU Histone-like protein

IAV Influenza A virus

IBC Intracellular bacterial community

ICAM1 Intracellular adhesin molecule 1

Ig Immunoglobulin

IHF Integration host factor

IL Interleukin

LAMP1 Lysosomal associated membrane protein 1

LC-MS/MS Liquid chromatography – tandem mass spectrometry

LDH Lactate dehydrogenase

LE Late endosome

LLO Listeriolysin O

LOS Lipooligosaccharide

LPS Lipopolysaccharide

MAPK MAP kinase

MβCD Methyl-β-cyclodextrin

MOI Multiplicity of infection

NCE Normalized collision energy

NET Neutrophil extracellular trap

NF-κB Nuclear Factor-kappa B

NGAL Neutrophil gelatinase-associated lipocalin

NHBE Normal human bronchial epithelia

NTHI Nontypeable Haemophilus influenzae

OM Otitis media

OME Otitis media with effusion

OMP Outer membrane protein

OMV Outer membrane vesicle

PAF Platelet activating factor

PDE Phosphodiesterase

PFA Paraformaldehyde

PMN Polymorphonuclear

RE Recycling endosome

RLU Relative light units

RSV Respiratory syncytial virus sBHI Supplemented Brain Heart Infusion

xxi

SCV Salmonella-containing vacuole

SE Sorting endosome

SEM Standard error of the mean

SNARE Snap receptor

SNP Single nucleotide polymorphism

SNX Sorting nexin

SSR Slip strand repeat

TEM Transmission electron microscopy

TLR Toll-like receptor

TNS Trypsin neutralizing solution

TR Transiently restricted

UPEC Uropathogenic Escherichia coli

URI Upper respiratory infection

WGA Wheat germ agglutinin

xxii

Chapter 1. Introduction

Nontypeable Haemophilus influenzae (NTHI) is a gram-negative opportunistic pathogen and a major causative agent of otitis media (OM). NTHI resides in the nasopharynx as a commensal, but risk factors such as a prior viral infection can allow

NTHI to migrate up the Eustachian tube and into the middle ear to cause disease. The transition from a commensal in the nasopharynx to an opportunistic pathogen in the middle ear requires that NTHI have survival mechanisms for adapting to fluctuations in the host microenvironment and for evasion of the host immune response. The following sections will describe what is currently known about NTHI host-pathogen interactions, from initial colonization to establishment of biofilms or intracellular niches for survival in the middle ear.

1.1. Haemophilus background

The Haemophilus genus includes several gram-negative commensal and pathogenic bacterial species of the Pasteurellaceae family. Haemophilus species inhabit mucosal surfaces and cause disease in the respiratory tract, mouth, middle ear, and genitals of humans (1). Pathogenic members of the Haemophilus genus include

Haemophilus ducreyi, the causative agent of chancroid, Haemophilus parainfluenzae, which can cause bacterial endocarditis or pneumonia, and Haemophilus influenzae, the major pathogen of this genus.

H. influenzae, a coccobacillus that was first identified by Richard Pfeiffer in 1892, can exist both as a commensal in the nasopharynx and an opportunistic pathogen. H. influenzae can be divided into two groups: strains with capsular polysaccharide

(serotypes a, b, c, d, e and f) and nonencapsulated strains. Haemophilus influenzae type b (Hib) is the most virulent of the encapsulated strains and for a period of time was the major causative agent of meningitis in children (2). However, introduction of a Hib vaccine for children in 1987 and for infants in 1990 decreased the incidence of pediatric

Hib-related illness by 92% by 2008 (3). The Hib vaccine is now one of the recommended childhood immunizations in the United States. The near eradication of Hib disease in children from effective vaccinations paved the way for a shift in the contribution of other encapsulated H. influenzae strains to invasive disease. While many of the serotypes of encapsulated Haemophilus still rarely cause disease, infections caused by serotypes a and f are increasing in the post-vaccine era. In some geographical areas the rate of invasive infection caused by these serotypes has risen to a higher rate than that of infections caused by Hib before the introduction of the conjugate vaccine (4, 5).

Strains that do not contain capsular polysaccharide are commonly referred to as nontypeable Haemophilus influenzae (NTHI). NTHI resides as a commensal organism in the nasopharynx of humans, but can become an opportunistic pathogen to cause several diseases, most notably acute and chronic OM, conjunctivitis, sinusitis, and exacerbations of chronic obstructive pulmonary disease (COPD) and cystic fibrosis (CF).

Widespread use of the Hib conjugate vaccine has also resulted in an increase in invasive and non-invasive disease caused by NTHI across multiple age groups (3, 6-13).

NTHI is now the most common cause of both invasive Haemophilus disease as well as

OM (4). Increasing our understanding of how NTHI adapts to and takes advantage of niches in the middle ear and lung environment to cause disease will ultimately lead to 2 the development of more targeted and efficient therapies. Remaining sections in this introduction will discuss what is currently known about NTHI pathogenesis, including colonization, immune evasion, mechanisms of persistence and adaptation, and vaccine development.

1.2 Otitis Media

Humans are the only natural reservoir and host for NTHI, and transmission between hosts occurs primarily via aerosolized respiratory secretions. As a commensal, NTHI colonizes the mucus membranes of the human nasopharynx. Colonization rates have been reported to range from 25-84% (14-16). Colonization with NTHI often occurs early in life; one study found that 70-80% of children are colonized with a bacterial otopathogen by age 12 (15). Daycare centers represent a reservoir for NTHI (17, 18) and earlier colonization is correlated with a greater risk of developing OM (15, 19).

Transition from a commensal to pathogen occurs following migration of NTHI from the nasopharynx into the middle ear where it drives a pro-inflammatory response and the buildup of fluid and mucus in the middle ear, hallmarks of OM (20). Migration of NTHI typically occurs in patients with certain risk factors, such as a preceding viral infection

(21, 22).

OM is an inflammatory disease of the middle ear caused by the migration of pathogenic agents from the nasopharynx to the middle ear. There are 709 million cases of OM globally each year, and despite advances in antibacterial therapies OM remains a significant socioeconomic burden (23). OM is largely a pediatric disease that affects young children with a peak incidence of between 9 and 15 months of age (24). Children younger than 6 months of age are often protected from OM by the presence of maternal antibodies and a diagnosis of OM at this young age is often predictive of frequent 3 episodes in the first few years of life (24). From a financial standpoint, acute and chronic

OM can be costly to manage for both healthcare providers and caregivers. It is estimated that the financial burden of diagnosing and managing OM cases in the U.S. alone exceeds $5 billion annually (25-27). Therefore, the development of therapeutics with increased efficacy is paramount to reducing the burden of OM on society.

1.2.1 Pathology

OM can be classified into different manifestations of disease, including acute otitis media (AOM), otitis media with effusion (OME) and chronic otitis media with or without effusion (28). Clinical diagnosis occurs when a healthcare provider visually observes signs of infection within the middle ear: a reddened and thickened tympanic membrane that is often bulging from a build-up of pressure in the middle ear space (29).

The presence of fluid in the middle ear space determines a diagnosis of OME and OME can persist after an initial diagnosis of AOM. Approximately 50% of children will present with effusion one month following an episode of AOM (28). Children who are diagnosed with at least three episodes of AOM within six months are defined as otitis-prone children. These children often have immature immune systems characterized by poor antibody and/or cellular responses to AOM infections (30, 31).

1.2.2. Etiology

The predominant causative agents of OM are the opportunistic pathogens NTHI,

Streptococcus pneumoniae, and Moraxella catarrhalis (24). These commensal residents of the nasopharynx are thought to induce OM by either passive aspiration or active growth and ascension up the Eustachian tube and into the middle ear space. Once in the middle ear, a robust inflammatory response is generated leading to symptoms of the disease. Risk factors for OM can include prior viral infection or Eustachian tube dysfunction (22). Also, children are especially susceptible to OM due to an immature 4 immune system as well as difference in Eustachian tube anatomy. The anatomy of the

Eustachian tube changes over time from the short, horizontal anatomy of the Eustachian tube in a child’s ear to a longer and more vertical Eustachian tube by adulthood. These anatomical differences mean children are prone to OM because the short, horizontal

Eustachian tube allows easier transit of bacteria from the nasopharynx to the middle ear.

After the development and use of the 7-valent pneumococcal conjugate vaccine

(PCV7), the number of cases of AOM attributed to S. pneumoniae decreased by a reported 34% (32). However, in the following years the number of cases of AOM caused by M. catarrhalis and H. influenzae increased (33). While the causative agents of OM can vary by geographical region, approximately 56% of cases are caused by H. influenzae, 22% of cases are caused by M. catarrhalis, 12% of cases are caused by S. pneumoniae and 10% of cases are caused by unidentified bacterial species (28).

Interestingly, H. influenzae is more likely to be isolated from one or both ears in bilateral cases of AOM compared to other otopathogens (34). A recent study examined 105 acute OM episodes in Spanish children and found that NTHI was the most common otopathogen isolated from recurrent or recalcitrant AOM (35).

1.2.3 Role of Viral Infection

OM is often not solely bacterial in nature, but can manifest as a polymicrobial disease caused by both bacteria and viruses. In most children, AOM develops as a coinfection with bacterial and upper respiratory pathogens. At least 90% of AOM and OME cases develop after a viral upper respiratory infection (URI) (28). Respiratory viruses can be identified in the middle ear fluid of up to 70% of children with AOM cases (36). The most common viruses associated with OM are the influenza A virus (IAV), respiratory syncytial virus (RSV), and human rhinovirus (HRV) (37).

Respiratory viral infections play a significant role in elevating a child’s risk for developing OM. A prior viral URI can activate inflammatory processes and compromise the innate and adaptive immune responses, providing an opportunity for nasopharyngeal bacteria to enter the middle ear space and cause OM. The presence of a viral pathogen has been found to increase mucous production in the nasopharynx (38). Although the mucin response is generally a defense mechanism against invading pathogens, overexpression of mucin and the resulting goblet cell hyperplasia can slow the beating of ciliated epithelium of the bronchial wall and, thus, impair the ability of cilia to clear pathogens (39). As a result, inflammation occurs in the nasopharynx, and the

Eustachian tube begins to constrict. Building negative pressure promotes retraction of the tympanic membrane and the build-up of mucus and fluid behind the middle ear.

Because the middle ear is normally sterile, it lacks innate immune factors at the onset of disease, and as a result, pathogens that gain access to this site are initially able to survive and replicate. Subsequent activation of the immune response in the middle ear to clear the pathogen results in middle ear inflammation and clinical symptoms of AOM such as edema and a bulging tympanic membrane, often accompanied by effusion.

1.2.4 Current Treatments for Otitis Media

Following a clinical diagnosis of acute OM, the major treatment regimen is a course of antibiotics. Specifically, the first-line choice of antibiotic treatment is a high dose of amoxicillin (40). As a β-lactam antibiotic, amoxicillin is effective against both penicillin-resistant S. pneumoniae and H. influenzae. However, many strains of H. influenzae and M. catarrhalis produce β-lactamase, which can inactivate amoxicillin and confer resistance. To combat this, the β-lactamase neutralizing product clavulanate is often added to amoxicillin to increase efficacy. Other antibiotics regularly prescribed to treat OM include oral cephalosporins, such as cefuroxime and cefdinir. If both first and 6 second-line antibiotics fail, clindamycin can be used; however, this antibiotic is also ineffective against β-lactamase-positive bacteria (40)

Even though a course of antibiotics will resolve most cases of OM, it is evident that new and innovative treatments are needed to combat this disease. OM is the number one reason for antibiotic use in children, and cyclical exposure to antibiotics, as occurs in children with persistent or recurrent infection, can potentiate antibiotic resistance. Repeated high doses of antibiotics can have detrimental effects on the overall microbial flora of children. Additionally, for children who are otitis prone, courses of antibiotics do not effectively clear the pathogen. Whether this is because of ineffective antibiotics (due to production of β-lactamases, for example) or bacterial strains that have more broadly adapted to survive and persist despite antibiotic and immune pressures remains underexplored. To curb unnecessary antibiotic exposure, in 2004 the American

Academy of Pediatrics introduced a recommendation for the process of “watchful waiting” (delayed therapy with active surveillance) for managing OM. A 2010 study found that prescribing antibiotics for OM showed only minor benefit compared to watchful waiting (41). Therefore, updated treatment guidelines in 2013 recommended that children with a fever of greater than 102.2 degrees Fahrenheit, children younger than two years old, or children with a ruptured ear drum with drainage should receive antibiotics right away; all others should be monitored under the watchful waiting approach and only given antibiotics if there is no improvement within 72 hours (40).

1.3 NTHI in other diseases

As an opportunistic pathogen of the human respiratory tract, NTHI plays a significant role in not only the pathogenesis of OM but other respiratory diseases as well.

NTHI is a causative agent of both rhinosinusitis and conjunctivitis, as well as can cause serious and persistent exacerbations of CF and COPD.

1.3.1 NTHI in Upper Respiratory Diseases

Acute conjunctivitis is the most common pediatric ocular disease, and NTHI is a major causative agent of conjunctivitis in children (42). Conjunctivitis can frequently occur in parallel or as a complication of acute OM (known as conjunctivitis-otitis media syndrome) (42). However, NTHI can also cause conjunctivitis in adults, particularly the elderly, albeit more rarely. In 2005, a Veterans Affairs nursing home reported an outbreak of NTHI-mediated conjunctivitis (43). NTHI has also been implicated as a causative agent of rhinosinusitis. A recent study found that in a cohort of 50 patients with rhinosinusitis, 56% had NTHI associated with mucosal changes at early disease time points, and 40% of patients still harbored NTHI within the maxillary sinus cavity out to at least 10 days, underscoring the persistent nature of NTHI (44).

1.3.2 NTHI in Lower Respiratory Tract Diseases

COPD is a progressive obstructive lung disease that is characterized by difficulty breathing, excess production of mucus, and frequent coughing or wheezing. It is often caused by long-term exposure to pollutants or irritants, such as tobacco or chemical fumes; however, some cases of COPD are caused by a genetic disease called alpha-1- antitrypsin deficiency (45). CF, a chronic, genetic lung disease, is also characterized by a buildup of mucus in the lungs or other organs.

Both COPD and CF patients struggle with repeated exacerbations of disease often caused by viral or bacterial infection (46). NTHI was found to be the most common organism isolated from the airway of COPD patients during such exacerbations (47-50), whereas Staphylococcus aureus and NTHI have been shown to colonize the respiratory tract of infants with CF (51, 52). NTHI can often persist in the lungs of COPD or CF 8 patients for years at a time. In a 5 year study, persistent colonization with the same NTHI clone was found in 37% of patients (53), further demonstrating the ability of NTHI to adapt to the environment for long-term survival.

1.4 NTHI persistence

The role of persistence in NTHI disease is becoming increasingly evident.

Longitudinal studies have revealed that NTHI is able to persist for months to years in patients with COPD (54). Alarmingly, NTHI can remain viable even in patients whose expectorated sputum cultures tested clinically negative for the presence of NTHI (55).

These data suggest that NTHI can adapt to changing environments and, perhaps, is able to gain a beneficial protective niche within the host to enhance survival despite both immune pressures and antibiotic treatments. Support for this hypothesis came from the observation that, in a cohort of patients, a population of NTHI remained viable in the pharynges despite adenotonsillectomy and subsequent antibiotic treatment (56).

There are multiple mechanisms that NTHI can use to persist for extended periods within the host. The formation of recalcitrant biofilms within the middle ear or the lung protects bacteria from antibiotics, antimicrobials, and phagocytosis by host immune cells (57). Invasion into host epithelial cells and the formation of intracellular bacterial communities (IBCs) (58-60) provide a temporary safe haven. Additionally, type II toxin- antitoxin systems generate phenotypic NTHI variants that form a “persister population”, which can survive various pressures (61-66). Finally, adaptation of the bacteria to nutrient stress, for example heme-iron limitation, promotes multiple phenotypes beneficial for survival in a preclinical model of OM (60).

Genetic and phenotypic diversity of NTHI strains contributes to NTHI adaptability and persistence in the host. The genetic content of NTHI can be modified by 9 polymorphisms, deletions, insertions, or by horizontal gene transfer, as NTHI is a naturally competent organism (67). Other mechanisms for varying gene content include hypermutation and phase variation, which occurs by slipped-strand mispairing of by simple sequence repeats within the coding or promoter regions of genes (68, 69). Phase variation results in rapid and reversible switching between on and off states of gene expression (70). Several recent studies have started to characterize the genetic evolution of persistent NTHI strains over time, and a variety of genetic features have been found associated with persistent NTHI strains. For example, the gene sequence of the outer membrane protein (OMP) P2 changes over time (71, 72). Additionally, phase variation of hmw1A and hmw2A, which leads to decreased production of the encoded adhesins, was observed in persistent isolates of NTHI characterized in COPD patients

(73). The role of phase variation in genetic diversity and in NTHI evasion of the immune response is discussed further in section 1.11.2.

1.5 Nutritional immunity

Transition metals are important for a variety of cellular processes. Over 30% of all enzymes require an interaction with a metal cofactor for their function (74, 75).

However, the human body tightly sequesters transition metals, to help maintain homeostasis. In the absence of tight regulation, metals can generate toxicity through the non-specific catalysis of redox reactions via the Fenton reaction (76). Importantly, microbes also require transition metals for a variety of processes including enzymatic and cytochrome functions (75, 77). Therefore, the sequestration of essential nutrients by the human body provides a secondary line of innate immune defense. For this reason, this process has been termed “nutritional immunity”.

The term nutritional immunity first arose out of studies in iron limitation. It was found that a person’s level of susceptibility to infections, and ultimately the disease outcome, could be influenced by how much iron was consumed in their diet (78, 79).

Similarly, a link between iron levels and the frequency of OM was established in a study that revealed a higher prevalence of OM in children with previously diagnosed anemia

(80). The remainder of this section will discuss how nutritional immunity regulates and restricts microbial access to not only iron, but also zinc, manganese, and copper.

Iron is used by the human body for many essential processes, including nucleic acid and protein synthesis, electron transport, and cellular respiration (81, 82). Most of the iron in the body is stored intracellularly, with 65-75% of all iron complexed to heme and located within erythrocytes (83). Any extracellular iron or heme is kept tightly sequestered by iron and heme-binding proteins such as transferrin, lactoferrin, hemoglobin, and haptoglobin. In addition to binding free iron and heme, other proteins such as NRAMP1 or ferroportin can transport iron and other metals out of phagosomes, further restricting access to microbes (84, 85).

To combat iron sequestration by the host, bacteria have evolved complex mechanisms to acquire iron and heme. Many gram-negative bacteria secrete siderophores to bind and subsequently transport iron through TonB-mediated receptors

(86). When siderophores reach the bacterial cytoplasm, the bound ferric iron is released by either degradation of the siderophore or reduction to ferrous iron, which is then used as a nutrient source (87). Further highlighting the push-pull nature of host-pathogen interactions, host proteins have evolved that can bind bacterial siderophores to prevent iron transport into bacteria. For example, neutrophil gelatinase-associated lipocalin

(NGAL) and lipocalin-2, are host proteins released by neutrophils and epithelial cells that sequester siderophores of both Salmonella typhimurium and Klebsiella pneumoniae (88- 11

90). Bacteria can also use heme uptake systems, transferrin or lactoferrin receptors, and ferric or ferrous iron transporters to import iron and/or heme. Alternatively, some bacteria have also evolved to substitute alternative transition metals to iron as cofactors in certain enzymatic processes (91).

Because pathogens require manganese or zinc as cofactors for many enzymes, sequestration of these transition metals, in addition to iron, also impacts bacterial survival and pathogenesis. Many eukaryotic proteins that function to sequester other transition metals also have the ability to bind and sequester manganese or zinc. For example, natural resistance-associated macrophage protein 1 (NRAMP1) and ferroportin have both been implicated in the transport of manganese in addition to iron (92-94).

Calprotectin is a member of the S100 family of proteins that is expressed by neutrophils to sequester zinc and manganese using dual metal binding sites (95, 96). The Zip family of proteins function as zinc transporters, with Zip8 specifically acting to transport zinc away from the lysosomes and into the cell cytosol to restrict access to microbes being trafficked through the endolysosomal pathway (97).

Nutritional immunity is not only accomplished through sequestration of essential nutrients, in some cases, an increased presence of transition metals functions as an antibacterial strategy through toxicity. Copper is a transition metal which has a role as an antibacterial when present in levels toxic to bacteria. Copper toxicity is multifactorial and includes oxidative damage through Fenton chemistry, nitrosative stress, and disruption of iron-sulfur clusters in bacteria (98). Copper resistance is necessary for the virulence of some bacterial species, such as Mycobacterium tuberculosis (99). Pathogens possess a variety of mechanisms for copper resistance, including copper transport proteins, the presence of copper exporters, a low requirement for copper, and localization of copper dependent proteins outside of the cytoplasm (98). 12

Considering the many host strategies for sequestering nutrients, it is clear that nutritional immunity is multi-faceted in combating infection by pathogens which require transition metals from the host for survival and growth. Organisms with effective strategies for overcoming the challenge of nutritional immunity through de novo synthesis, or through acquisition of nutrients from host-proteins, therefore, have the best chance for survival and persistence in the host.

1.6 Heme-iron utilization by Haemophilus

1.6.1 Heme-iron acquisition

NTHI is a heme-iron auxotroph, unable to convert 5-aminolevulinic acid to protoporphyrin IX, the immediate precursor of heme. NTHI must, therefore, acquire heme-iron from the environment to grow and survive within its host (100, 101). Currently, no strains of H. influenzae have been identified that encode genes involved in the synthesis of siderophores. Recently, though, a siderophore utilization locus was identified in a subset of NTHI strains, suggesting that NTHI may be able to hijack the siderophores of other organisms for iron and heme acquisition (102).

Without the ability to synthesize heme or produce siderophores, NTHI combats nutritional immunity by acquiring heme-iron using a variety of heme- and iron-binding proteins and transporters. Annotation of the NTHI strain 86-028NP genome revealed two iron transport systems: hitABC, which encodes an ABC transporter involved in ferric iron uptake and hfeABCD, encoding a second ABC iron transport system (103). An additional putative ferric iron transport protein is encoded by NTHI2035 (103). Hup and the Hxu family of proteins function as heme and heme-hemopexin-binding proteins that are essential for heme acquisition at the outer membrane (104, 105). The Hgp family of proteins similarly function to bind hemoglobin; however, deleting all Hgp proteins does 13 not abolish hemoglobin utilization, suggesting that NTHI possesses overlapping mechanisms for binding hemoglobin (106-108). NTHI also uses the Tbp1/Tbp2 proteins to bind transferrin on its outer membrane (109). Interestingly, an adhesin, protein E, has recently also been shown to be able to bind and to store hemin, later distributing it to nearby NTHI under iron-starved conditions (110).

The transport of heme across the NTHI inner membrane is accomplished through two major transporters, Dpp and Sap (111, 112). In Escherichia coli, Dpp is a peptide- heme permease encoded by the dppABCDF genes (111). The NTHI genome lacks the heme-binding protein DppA, but encodes a dppBCDF gene cluster (111). Recent studies demonstrated a role for Dpp in heme-utilization by NTHI and proposed a role for the independently expressed heme-binding protein HbpA as a homolog of E. coli DppA, which delivers heme to the Dpp permeases (111). However, deletion of the dppC gene

(encoding a permease of the Dpp transport system) in NTHI does not completely abrogate heme utilization in NTHI, further supporting overlapping heme-uptake mechanisms (111).

NTHI does possess a second ABC transporter that functions in heme-acquisition, which is encoded by the sap operon, comprising genes sapABCDFZ. The Sap transporter consists of two permeases (sapBC) and two ATPases (sapDF), with SapA functioning as a periplasmic heme-binding protein (112-116). Importantly, deletion of the sap genes results in a heme-starved phenotype for NTHI that is characterized by changes in bacterial morphology as well as virulence (59, 112, 116). Environmental transient heme-iron limitation of NTHI is shown to induce new and different phenotypes that include a filamentous morphology, the formation of biofilm towers, and persistence both in long-term stationary phase in vitro and as IBCs in vivo in a preclinical model of otitis media (60). The ability for NTHI to adapt to changes in heme-iron availability, both 14 because of genetic or environmental heme restriction, demonstrates the complex relationship between NTHI iron and heme uptake and subsequent regulation of gene expression.

1.6.2 Fur regulation

NTHI uses iron to regulate gene expression through the ferric uptake regulator

(Fur). Fur acts as an iron responsive repressor of the transcription of several genes, a subset of which are involved in iron import (77). Classically, under conditions where iron levels are high, Fur and its co-repressor Fe(II) bind to a conserved DNA sequence, the

Fur box, upstream of the regulated gene. Binding of Fur blocks RNA polymerase binding, and prevents gene transcription. When iron levels are low, Fe(II) releases from

Fur, and Fur disengages from the promoter, allowing RNA polymerase to now bind and initiate transcription (117). Genes with roles in iron and heme import or utilization that are known to be Fur-regulated in NTHI include hitABC, hemR, tonB, hxuBC, hfeABC, hgpB, hbpA, tbp1, tbp2, and ftnAB (77).

In addition to acting as a negative regulator of genes involved in iron acquisition,

Fur can also act as a positive regulator either by directly binding to DNA distal to the regulated gene’s promoter region (77) or indirectly through small regulatory RNAs

(sRNAs). To date, only one Fur-regulated sRNA, known as HrrF (Haemophilus regulatory RNA responsive to iron Fe) has been identified in NTHI (118). HrrF regulatory targets include genes involved in amino acid and deoxyribonucleic acid synthesis as well as molybdate uptake (118).

1.7 Bacterial signaling

1.7.1 Roles for cAMP: catabolite repression and competence

Cyclic adenosine monophosphate (cAMP) is a major secondary messenger found in all kingdoms of life. cAMP is synthesized from ATP by adenylyl cyclases (ACs), of which there are six classes. The bacterial ACs are members of classes I, II, and IV.

ACs have roles both in virulence and catabolite repression in multiple bacteria (119,

120). Some pathogens secrete AC as a toxin that can alter intracellular cAMP concentrations in host cells; known AC toxins include Bordatella pertussis CyaA, Bacillus anthrasis edema factor, Pseudomonas aeruginosa ExoY, and Yersinia pestis AC (121-

125). More commonly, AC is involved in initiating catabolite repression as discussed later in this section. Transcription of ACs can be either positively or negatively regulated in prokaryotes. For example, E. coli cya encodes AC, and its expression is repressed by catabolite repression protein (CRP) in complex with cAMP. cAMP levels are, therefore, regulated via a negative feedback loop (126).

Degradation of cAMP occurs through hydrolysis of cAMP to AMP by phosphodiesterases (PDEs). There are three classes of PDEs based on the structure of their catalytic sites (127). The majority of bacterial PDEs are members of Class III and have dual specificity for cAMP and cyclic guanosine monophosphate (cGMP). The catalytic activity of class III PDEs is dependent on the binding of two divalent cations in the active site (128). Specifically, the presence of iron (ferrous or ferric) is essential for the function of most class III PDEs, as demonstrated by the observation that iron chelation significantly reduced the catalytic activity of the P. aeruginosa PDE CpdA

(129).

In prokaryotes, cAMP acts as a second messenger acting on multiple cAMP- dependent signaling pathways. The basic requirements for a cAMP-dependent signaling 16 pathway includes production of cAMP, a binding partner for cAMP for signal transduction, and a method for signal termination (for example, PDE activity) (130). The most well studied cAMP signaling pathway is the catabolite repression system, which allows bacteria to adapt to their preferred carbon source. This occurs by a regulatory process that reduces the expression of genes involved in the use of a secondary carbon source when the preferred carbon source is present. The catabolite repression pathway requires the cyclic AMP-binding protein, CRP, in addition to cAMP as a signaling molecule. The presence of a preferred carbon source triggers AC to synthesize cAMP, and high concentrations of cAMP induce the binding of cAMP and CRP (131). The resulting cAMP-CRP complex then interacts with over 300 intergenic-specific DNA sequences, in most cases, activating transcription of the regulated genes (132). PDE activity decreases the concentration of cAMP available for complexing with CRP to terminate the signal and allows adaptation to other carbon sources or regulation of gene expression (131).

The biological roles for bacterial cAMP PDEs have expanded in recent years. In

E. coli, PDEs have the dual function of regulating carbon metabolism and protecting against oxidative damage (133). The PDEs of P. aeruginosa and M. tuberculosis regulate expression of genes involved in virulence and cell wall permeability (129, 134-

137), whereas in Serratia marcescens, PDEs help regulate biofilm formation (138).

In Haemophilus, cAMP regulates competence or the ability to take up extracellular DNA from the environment through transformation. The ability to take up

DNA from the environment is important for survival and virulence in many bacteria.

Competence allows the exchange of genetic material and generates genetic diversity within a species (139). Further, bacterial competence, with the uptake and subsequent degradation of DNA, provides a nutritional source of nucleotides (140, 141). In bacteria, 17 competence is generally induced in response to a set of environmental cues such as

DNA damage, quorum sensing, or nutritional starvation (142). Haemophilus is naturally transformable in stationary phase (143), but nutritional signals are also important for regulating competence in NTHI. When levels of phosphotransferase sugars are low,

NTHI increases intracellular cAMP levels. The resulting cAMP-CRP complex then induces transcription of sugar utilization genes as well as the competence activator known as sxy (143-145). Sxy and CRP induce the transcription of the CRP-S regulon, which includes 25 genes in Haemophilus and induces DNA uptake (143, 146, 147). The levels of intracellular cAMP directly influence NTHI competence, as is evidenced by the fact that the addition of exogenous cAMP to NTHI enhances competence (148). The icc gene in Haemophilus encodes an ortholog of the E. coli PDE CpdA (143, 149). Icc regulates intracellular cAMP levels and competence in Haemophilus and acts to prevent toxicity by cAMP buildup (150). As CpdA in E. coli is activated by iron (151), the interplay between nutritional immunity and regulation of cAMP levels in NTHI is an underexplored area of study with regard to NTHI persistence and survival during disease.

1.7.2 Indole signaling in bacteria

Indole is a small, aromatic molecule that plays an integral role in both interspecies and interkingdom signaling (152). As a metabolic byproduct of over 85 gram-negative and gram-positive microbes, indole modulates many aspects of bacterial physiology, including biofilm and spore formation, antibiotic resistance, and the generation of persister cells (153, 154). In indole-producing bacteria, tryptophanase

(encoded by tnaA) converts tryptophan into indole, pyruvate, and ammonia. Tryptophan may either be synthesized by the bacteria or imported from the environment. The genome of NTHI strain 86-028NP includes the entire tryptophan biosynthesis operon

(trpABCDGE) as well as genes tnaAB involved in tryptophan catabolism and import. 18

However, our knowledge of the specific role of tryptophan synthesis and indole export in

NTHI pathogenesis is lacking.

While indole can be generated from tryptophan catabolism, indole may also be transported into the cell from the environment via the permeases Mtr, TnaB, or AroP

(155). Once produced or imported, cell-associated indole can be secreted, with extracellular concentrations of indole in E. coli ranging from 0.5 – 5mM, a concentration that is proportional to the amount of exogenous tryptophan in the environment (156).

Indole may be exported from the cell through the AcrEF-TolC pump E. coli (155, 157).

However, indole is freely diffusible across membranes, suggesting that any bacterium in a polymicrobial environment may be subject to the effect of indole (158).

Bacterial responses to the presence of indole include increased biofilm formation by P. aeruginosa and increased antibiotic resistance, via upregulation of multidrug exporters, in E. coli and Salmonella (159-161). Further, indole can directly induce quiescence and the generation of persister populations of E. coli, suggesting a role for indole in recalcitrant infection (162, 163). Strengthening this idea, the ability to produce indole is more frequent among disease-causing isolates of Haemophilus compared to commensal strains (164).

The interaction of bacteria with host cells is also modulated by indole signaling.

Indole inhibits attachment of enterohemorrhagic E. coli (EHEC) and Candida albicans to host cells, as well as decreasing the invasion of Salmonella species in vivo (165-167).

The ability to synthesize tryptophan promotes the intracellular survival of both

Francisella and Chlamydia species; however, host cell production of the tryptophan degrading enzyme, indoleamine 2,3-dioxygenase, normally limits intracellular bacterial growth (168, 169).

1.8 Colonization and Adherence

To circumvent clearance by the mucosal epithelium, NTHI uses strategies to adhere to host tissues, evade the host immune response, and persist in the host through the formation of biofilms or invasion of host cells. The remaining sections will discuss mechanisms for these various NTHI lifestyles during disease progression and persistence.

The first step in infection is adherence to host mucosal tissue and subsequent colonization of host disease sites. NTHI preferentially adheres to non-ciliated epithelium or damaged epithelia with non-functional cilia (170, 171). NTHI expresses several adhesins that aid in adherence to the mucosal surface by binding to host molecules.

Outer membrane proteins P2 and P5 bind mucin (172); lipoprotein E (P4) can bind laminin, vitronectin, or fibronectin (173); whereas protein E binds laminin and vitronectin

(174). Protein F binds laminin alone (175). PilA, the major subunit of the type IV pilus of

NTHI, mediates adherence to intracellular adhesin molecule I (ICAM1) on epithelial cells

(176).

NTHI expresses several autotransporter proteins that also play roles in adherence and colonization. Autotransporters are a family of proteins that transverse the outer membrane and share three common structural domains: a N-terminal signal sequence, a C-terminal β-barrel (the translocator), and a passenger domain (177).

Although it was originally accepted that autotransporters do not require accessory proteins for secretion (hence the term ‘autotransporter’), recent studies have revealed that the BamA outer membrane protein, involved in the insertion of β-barrel proteins into the outer membrane, is, indeed, integral for localization of the autotransporter within the outer membrane (178-180).

Haemophilus strains contain several autotransporters, including the high molecular weight (HMW) adhesins HMW1 and HMW2, Hap, Hia, and Hif. The prototypical clinical strain NTHI 86-028NP contains only the HMW proteins and Hap

(103). HMW1 and HMW2 are encoded in two operons that each contains three genes: hmw1A,hmw1B,hmw1C and hmw2A,hmw2B,hmw2C (181, 182). These two HMW proteins are similar, with the amino acid sequences displaying 71% identity (181). Both

HWM1A and HMW2A are similarly processed and exported. Using HMW1A as the example, the adhesin contains an N-terminal signal sequence that interacts with the Sec system for translocation across the inner membrane (183). HMW1B then facilitates transport across the outer membrane, and the C-terminal domain of HMW1B anchors the adhesin in the outer membrane for presentation on the bacterial surface (184). The

HMW1C accessory protein is required for post-translational glycosylation (185). HMW1 and HMW2 have different binding partners and affinity for specific cell types. HMW1 binds sialyl-α2,3 hexose during early infection and during commensalism in the nasopharynx (186). Sialyl-α2,3 hexose is not present in the respiratory tract, suggesting that NTHI must use other binding partners for adherence during disease progression

(187). HMW2 binds 2,6 linked N-acetylneuraminic acid with high affinity (188). The HMW adhesins are present in approximately 40-75% of NTHI isolates (189-194). Interestingly, strains that do not encode the HMW genes often possess the genes for an alternative adhesin, Hia (186).

The Hap autotransporter was first identified as a factor involved in adherence of clinical Haemophilus isolates to epithelial cells (195). Hap (Haemophilus adherence and penetration) contains an N-terminal globular serine protease domain and a C-terminal β- barrel (196). Interestingly, the Hap passenger domain contains a self-associating autotransporter domain. As a result, Hap can adhere to human epithelial cells, proteins 21 of the extracellular matrix, and other bacteria expressing Hap, which promotes bacterial aggregation (197-199). The host binding partners for Hap include fibronectin and collagen IV (200). Further, Hap can coordinate with protein E to bind laminin (174). Hap is downregulated in the early stages of infection, suggesting a more prominent role for

Hap in the later stages of disease and persistence (201). In support of this idea, the self- aggregation functionality indicates that Hap may contribute to the formation of biofilm communities by NTHI in host niches (197).

1.9 Biofilms

One mechanism NTHI uses to persist in the host, and so cause chronic and recurring disease, is the formation of biofilms. Biofilms are heterogeneous, multicellular communities organized within an extracellular polymeric substance (EPS). The EPS functions as a scaffold for the biofilm structure and connects the bacteria to each other as well as to the surface to which the biofilm is adhered. The structure and architecture of biofilms can vary both between species and in response to environmental stimuli. For example, Pseudomonas forms flat biofilms when grown in a flow chamber in a citrate medium, but biofilms grown in the presence of glucose are mushroom-shaped (202).

Similarly, NTHI alters the structure and organization of biofilms based on heme-iron availability. When continuously exposed to heme-iron, NTHI forms biofilms with a mat- like architecture, but after transient heme-iron restriction, NTHI transitions to a biofilm characterized by tower-like structures (60).

Despite variations in architecture, all biofilms require an EPS to form organized structures. The EPS is composed of polysaccharides, proteins, and extracellular DNA

(eDNA) (57). Some bacteria use multiple distinct polymers in the formation of the biofilm matrix, such as Psl, Pel, and alginate produced by P. aeruginosa (203). In contrast, 22

NTHI does not have any identified polysaccharides that are biofilm-specific; the major polysaccharide that contributes to the EPS of NTHI biofilms is lipooligosaccharide (LOS).

LOS is the major antigenic component of the NTHI outer membrane and is analogous to lipopolysaccharide (LPS) of enteric gram negatives. Modification of LOS can alter bacterial aggregation and biofilm formation (204). The addition of phosphorylcholine

(ChoP) to some LOS isoforms promotes the formation of stable biofilm communities by

NTHI (205). Similarly, sialyation of LOS by addition of N-acetylneuraminic acid results in increased biofilm formation by NTHI and increased persistence in vivo (206).

eDNA is the major matrix component of NTHI biofilms. The length and sequence of eDNA is variable and can be either bacterial or host in origin. Bacterial DNA is derived either from bacterial lysis or secretion of DNA through the comE pore of the type IV pilus, whereas host DNA is believed to be derived from neutrophil extracellular traps

(NETs) (57, 207). In most cases, the eDNA remains in long strands, which provides a support network for the biofilm structure as well as provides protection from host immune effectors. For example, eDNA binds the antimicrobial peptide β-defensin 3 and prevent its access to bacteria within the biofilm (208).

The third component of the EPS is protein. NTHI produces at least 18 biofilm- specific proteins including major outer membrane proteins P1, P2, and P5 (209), and the type IV pilin protein PilA (210). PilA is one of the most well-studied NTHI-derived EPS proteins; it is observed throughout the NTHI biofilm along double stranded DNA and provides structural stability to the biofilm (210). Additionally, a key component of the

NTHI biofilm matrix is a family of DNA-binding proteins, the DNABII proteins. The

DNABII family includes histone-like protein (HU) and integration host factor (IHF). Both

HU and IHF localize to bent eDNA within NTHI biofilms, and, importantly, antibodies directed against IHF destabilize and disperse NTHI biofilms both in vitro and in vivo 23

(211-214). DNABII proteins and DNA are transported to the periplasm via a Tra inner- membrane pore complex that has homology to type IV secretion-like systems (207).

Export of DNA and DNABII proteins out of the bacterial cell then occurs through the outer membrane via the ComE pore of the type IV pilus (207). Other nucleoid-associated proteins; such as H-NS, CbpA, HfQ, and Dps; are present in the biofilm matrix, but, from this class of proteins, only the DNABII proteins are necessary for stability of NTHI biofilms (215).

For biofilm formation to occur, a phenotypic change stimulates planktonic bacteria to adhere to either abiotic or biotic surfaces and to produce components of the

EPS. In many species, biofilm formation is initiated in response to second messengers, such as cyclic diguanylate (c-di-GMP) and small RNAs (sRNAs) (216-218). Further, cAMP acts as a signal to regulate biofilm formation in many species. cAMP and CRP signal the biofilm master regulator protein, CsgD, to upregulate production of cellulose and curli, EPS components in uropathogenic E. coli biofilms (UPEC) (219). In some

NTHI strains, biofilm formation is regulated by the ModA2 phasevarion; phasevarion switching contributes to biofilm formation in vivo with subsequent changes in disease severity (220). Biofilms, therefore, contribute to the chronicity of bacterial-mediated diseases, including in NTHI-mediated otitis media.

All three major otopathogens form biofilms both in vitro and in vivo on middle ear mucosae (221-226). These latter biofilms are often polymicrobial in nature (221). The persistence of biofilms is due in part to inherent resistance to immune effectors and antibiotics. Bacteria in a biofilm are up to 1000-times more resistant to antibiotics than planktonic bacteria (226). The reduced growth rate, and altered transcriptomes of biofilm cells, may contribute to this antibiotic resistance (227). Interestingly, treating an NTHI biofilm with sub-lethal concentrations of β-lactam antibiotics releases eDNA and 24 promotes biofilm formation (228). Strategies for designing more effective therapeutics for biofilm-mediated diseases are, therefore, focused on breaking down, or destabilizing, the biofilm before, or during, antibiotic treatment (57, 211-213).

1.10 Immune Response to NTHI

The survival and persistence of an invading pathogen depends in part on the microbe’s ability to resist host innate and adaptive immune defenses. Immune cell infiltrate, the presence of antimicrobial peptides and complement, and professional phagocytes at the site of infection act to drive inflammation and disease progression.

The following section will focus on different host responses that are initiated in NTHI- mediated diseases, specifically OM. Section 1.11 will focus on NTHI strategies for evading the host immune response.

1.10.1 Recruitment of Immune Cells

The cellular immune response to NTHI is dominated by an infiltrate of macrophages and neutrophils at the site of infection (229, 230). Although macrophages are first-line defenders, neutrophils release NETs in response to NTHI (231). Following the initial onset of OM, additional immune cells migrate to the site of infection, and include dendritic cells, T- and B-cells, and natural killer cells (232-234). The recruitment of immune cells is driven by activation of intracellular signaling pathways through NTHI interaction with toll-like receptors (TLRs) (reviewed in (235)).

1.10.2 Cytokine Response

The release of cytokines by host cells serves to attract leukocytes and other immune cells to sites of infection. The p38 mitogen-activated protein kinase (MAPK) and nuclear factor- kappa B (NF-κB) signaling pathways have emerged as key signaling cascades that regulate transcription of genes encoding cytokines involved in the immune 25 response to NTHI infections (236, 237). Infection with NTHI stimulates the robust production of the cytokines interleukin (IL)-8 and IL-6 by epithelial cells in vitro (238). A more recent study found that the most upregulated genes by epithelial cells in vitro following NTHI infection were those encoding chemokine (C-X-C motif) ligand 5

(CXCL5), CXCL10, CXCL11, C-C motif ligand 5 (CCL5), IL-1a, IL-8 and IL-23 (201).

1.10.3 Mucins

Goblet cells are abundant in the inflamed middle ear during OM (239). Mucins, produced by these goblet cells, form a gel-like substance that gives middle ear effusions their characteristic viscous properties. Overall, 19 mucin genes have been identified in the human middle ear (240). MUC2, MUC5AC, and MUC5B are important contributors to

OM disease progression, and their upregulation leads to mucosal metaplasia (241-243).

The transcription of mucin genes is activated in response to NTHI binding of TLRs. For example, NTHI outer membrane protein P6 binds TLR2 to initiate MUC5AC mucin transcription in the middle ear (244). Similarly, NTHI activates the transforming growth factor (TGF)-β-Smad signaling pathway to mediate NFκB-mediated transcription of

MUC2 (245). Polymicrobial infections may alter mucin production further; in fact, NTHI synergizes with S. pneumoniae to induce MUC2 transcription (246). The role of mucins in innate immunity against invading pathogens is mainly to function as a mechanical barrier. However, mucin production must be tightly regulated to strike a balance between an overwhelming host response and successful protection from pathogens. This is underscored by the observation that overproduction of mucins in the middle ear contributes to dysfunctional mucocilliary clearance, yet it also inhibits attachment of

NTHI to epithelial cells to prevent colonization (247).

1.10.4 Antimicrobial Peptides

Antimicrobial peptides (APs) are small molecules produced by both animals and plants (248). APs disrupt bacterial membranes by altering transmembrane potential and inducing lysis (248, 249). Some APs can cross the cell membrane and target intracellular nucleic acids or proteins, for example, Buforin II, which binds DNA and RNA

(250). Others specifically inhibit protein or RNA synthesis, such as the bactencins of bovines or the human defensin human neutrophilic peptide 1 (HPN-1) (251, 252). APs are divided into classes by charge and structure, and include linear cationic alpha-helical peptides (LL-37 in humans) and peptides containing cysteine that form disulfide bonds

(e.g., human α and β defensins), among others. β-defensins are important as they keep opportunistic pathogens at bay in addition to contributing to the immune response during

OM and other respiratory diseases (253). A lower copy number of the gene encoding β- defensins is correlated with an increased risk for nasopharyngeal colonization with OM pathogens, including NTHI (254).

Bacterial resistance to APs occurs by altered surface charge and/or outer membrane production. Addition of ChoP to the NTHI outer membrane, for example, alters hydrophobicity and increases resistance to LL-37 (255). In addition, many bacteria use ATP transporters to import APs for degradation, such as the Sap transporter in

NTHI, which mediates intake of APs in addition to heme-iron (115).

1.10.5 Complement

The complement system includes over 30 proteins that function to opsonize bacteria, recruit immune cells, and induce inflammation (256, 257). There are three major complement pathways: classical, alternative, and lectin. The classical pathway is activated by antigen-recognition: the antigen-bound immunoglobulin (Ig) G or IgM binds and activates the C1 complex (256, 257). The alternative or lectin pathways are initiated 27 when the host recognizes components of the bacterial envelope or membrane, such as mannose residues (258). The pathways converge at the formation of the complement

C3-convertase. C3 is cleaved into C3a and C3b (opsonin) during a cascade that ultimately leads to the formation of the cytolytic membrane attack complex, which causes bacterial cell lysis (259).

In general, the primary mechanism by which NTHI evades the complement system is through outer membrane modifications and binding of complement system factors by NTHI outer membrane proteins. NTHI proteins P5 and P4 bind human factor

H and vitronectin, respectively, to prevent downstream signaling of the complement cascade after deposition of complement C3 on the bacterial surface (173, 260, 261).

Binding vitronectin also protects NTHI from deposition of C5b-C9 and the subsequent formation of the membrane attack complex (173, 262, 263). Additionally, LOS is an important mediator of complement resistance. Deletion of lgtC, a gene involved in LOS biosynthesis in NTHI, increases the susceptibility of NTHI to C4b binding and complement recognition (264).

1.11 NTHI Immune Evasion Strategies

1.11.1 Outer Membrane Vesicles

Outer membrane vesicles (OMVs) are spherical, self-contained membrane blebs

(rounded outgrowths on the surface of the cell) that have diverse roles in bacterial pathogenesis. All gram-negative bacteria produce and actively secrete OMVs that deliver cargo such as toxins, virulence factors, or DNA. OMVs also play roles in biofilm formation and nutrient acquisition. OMVs elicit a broad immune response, and many promising vaccine strategies are capitalizing on the immunogenic nature of OMVs (265-

269). 28

NTHI produce OMVs that contain the OMPs P4 and P6 (270). NTHI OMVs stimulate the production of LL-37, as well as IL-8 and, additionally, activate B-cells in a

T-cell dependent manner (271, 272). Because of this, there is speculation that NTHI

OMVs may act as decoys to divert the host immune response and protect live NTHI cells from detriment.

1.11.2 Phase variation of surface epitopes

Phase variation enables bacteria to adapt to changing environments through generation of phenotypic diversity. Phase variation is a high-frequency, interchangeable switching between on and off gene expression states. This process occurs randomly and, in most cases, arises from a gain or loss of short tandem repeats in either an open reading frame or in the promoter of the phase variable gene (69). Surface antigens, such as LOS and other outer membrane proteins, can be phase variable in gram-negative bacteria such as Neisseria, Salmonella, and Haemophilus (273-275).

The stochastic generation of phenotypic subpopulations is useful in promoting bacterial survival in the face of selective pressure, while also allowing evasion of the host immune response during transition from a commensal to an opportunistic pathogen.

For example, gain or loss of simple sequence repeats in the hmwA promoter region is directly related to expression of hmwA in NTHI – a higher number of repeats leads to decreased HMW-adhesin protein levels (276). This switching in hmw expression is critical as the amount of HMW-adhesin protein is correlated with adherence and immune evasion; higher amounts of the HMW adhesin protein leads to increased NTHI adherence, whereas subpopulations of NTHI expressing lower levels of the HMW adhesin can better evade antibody-mediated killing (277).

The gene encoding OafA in NTHI similarly phase varies. OafA is a member of the LOS biosynthesis pathway and is used to acetylate LOS (278). Phase variation of 29 oafA was recently shown to occur in paired NTHI isolates, expression transitioning from being ON in the nasopharynx to being OFF in the middle ear (279). As acetylated LOS is immunogenic in other gram-negative bacteria, such as Salmonella and Neisseria, phase variation of this gene to the OFF state in the middle ear may aid in NTHI evasion of the immune response during OM (279-281).

1.11.3 IgA Protease

Immunoglobulin A (IgA) is abundant at the mucosal interface and is protective to the host by both neutralizing bacterial pathogens and preventing microbial association with epithelial cells by agglutination, while also maintaining homeostasis of commensal bacterial communities (282). Bacteria secrete IgA proteases which cleave the hinge region of human IgA1, inactivating the immunogobulin. IgA proteases are unique to human mucosal bacteria and were first identified in Neisseria and Streptococcus species in the 1970’s (283, 284). A homolog of IgA1 was subsequently discovered in H. influenzae (285, 286).

H. influenzae can encode two IgA proteases, IgaA and IgaB, and each protease has two variants with differing proteolytic activities and structural specificities (287, 288).

All H. influenzae strains produce the IgaA protease, but only 40% of strains produce

IgaB. This suggests differing roles for these two proteases in NTHI survival in the host

(287). The NTHI IgaA protease is important for persistent nasopharyngeal colonization, and has additionally been shown to be strongly associated with disease, particularly exacerbations of COPD (289, 290). In addition to cleaving IgA, the IgaB protease can also cleave lysosomal-associated membrane protein 1 (LAMP1) to facilitate intracellular survival of NTHI within mucosal epithelium (55).

1.11.4 Sap Transporter

The sensitivity to antimicrobial peptide (Sap) transporter was originally identified in Salmonella, wherein mutations in the Sap transporter conferred hypersensitivity to

APs (291, 292). The Sap transporter is a member of the ABC family of transporters and is encoded in NTHI strain 86-028NP by the sapABCDFZ operon (113, 291, 293). SapB and SapC are inner membrane permeases whereas SapD and SapF function as

ATPases. SapA is a periplasmic binding protein that has been recently discovered to have dual specificity for both heme-iron and antimicrobial peptides (112). The function of

SapZ is currently unknown.

The Sap transporter aids in NTHI survival and evasion of the host immune system through both heme-iron acquisition as well as the import of APs that are subsequently degraded in the bacterial cytoplasm (115). In an in vivo chinchilla model of

OM, the NTHI sap promoter is upregulated during disease progression, illustrating the importance of the Sap family of proteins in survival in a host (294). Indeed, a role for

SapA in binding antimicrobial peptides for mediating the host immune response was demonstrated by a study which found that neutralizing host APs restores the virulence of an NTHI sapA mutant during in vivo experimental OM (115).

In addition to nutrient acquisition and AP degradation, the Sap transporter may also be critical for the ability of NTHI to sense, and directly adapt to, host mucosal environments. An NTHI mutant lacking sapA displays significantly increased invasion into host epithelial cells, coupled with a decrease in the host pro-inflammatory response and cytokine release (59).

1.12 Bacterial Entry into Host Cells

A key feature of NTHI persistence in disease is the ability of the bacterium to exist in diverse lifestyles, including as a commensal of the nasopharynx and in a biofilm state in the middle ear and lungs. Recent studies have expanded these chiefly extracellular lifestyles to include an intracellular facet of NTHI pathogenesis. NTHI has now been observed within, and between, multiple cell types, including respiratory and middle ear epithelium, and macrophages (58). Invasion of NTHI into epithelial cells is largely strain- and cell-line dependent (295).

Some professional intracellular pathogens, such as Shigella or Listeria, directly manipulate host cell membranes and trafficking processes through the secretion of virulence factors. However, NTHI does not encode Type III or Type IV secretion systems, needed to secrete effectors (103). Therefore, NTHI must rely on the host endocytic pathways for entry into the cell or use yet unknown mechanisms to actively affect host cell processes. Both phagocytosis and endocytosis are natural cell processes that internalize nutrients, proteins, or foreign bodies, such as bacterial pathogens, for degradation or recycling. While phagocytosis occurs in specialized cells, such as macrophages, endocytosis is universal in mammalian cells. The remainder of this section will discuss bacterial interactions with host endocytic pathways.

1.12.1 Host cell cytoskeleton

Endocytosis can be divided into clathrin- or receptor-mediated endocytosis, caveolae/lipid raft-mediated endocytosis, and macropinocytosis. A major component of each of these uptake pathways is the fluid reorganization of the host cell cytoskeleton.

The cytoskeleton is comprised of actin, microtubules, and intermediate filaments that all function in intracellular transport, cellular morphology, and cell signaling, among other roles. Intracellular bacteria are often able to subvert actin polymerization. For example, 32

Shigella directly hijacks the actin filament branching protein Arp2/3 to use actin polymerization for bacterial motility within the cell (296, 297). NTHI infection has been shown to induce actin polymerization in epithelial cells, and NTHI invasion requires actin filaments and the Arp2/3 protein (171, 298-300).

In addition to manipulating actin reorganization, many intracellular pathogens secrete effectors that can manipulate components of host cell microtubules. Shigella and enteropathogenic E. coli hijack α and β tubulin, whereas Campylobacter jejuni manipulates dynein to promote movement towards the interior of the cell (301, 302).

NTHI invasion into epithelial cells requires microtubules, and, interestingly, NTHI may actively modulate microtubule activity to promote entry (299, 303, 304). The interaction of NTHI with host cell integrins initiates a Rak1-Pak1 signaling cascade that leads to inactivation of the microtubule destabilizer, Op18/stathmin (304). Further, NTHI entry into monocytic cells also relies on reorganization and polymerization of microtubules

(305).

1.12.2 Lipid-raft mediated endocytosis

Lipid rafts are glycosphingolipid and cholesterol-rich membrane domains that are required for caveolae-dependent endocytosis (306). In addition to endocytosis, lipid rafts initiate cell signaling pathways and membrane trafficking (307). Bacterial entry by raft- dependent pathways may be beneficial, as endosomes derived from lipid-raft-mediated endocytosis are not always trafficked to the endolysosomal pathway for degradation

(308). Lipid raft domains can be inactivated by pharmacological depletion of cholesterol; for example, the use of methyl-β cyclodextrin demonstrated that NTHI invasion requires lipid raft integrity (303). However, the exact nature and importance of the role of lipid- rafts in NTHI invasion may be cell type- or strain-specific, as additional studies found that NTHI uptake into bronchial epithelium was raft-independent (55). 33

1.12.3 Clathrin- or receptor-mediated endocytosis

Clathrin-mediated endocytosis is the major mechanism for receptor-mediated uptake in most eukaryotic cells (309). Receptor-mediated endocytosis is induced by the binding of extracellular molecules (host or bacterial) to cell surface receptors such as the

β-glucan or platelet activating factor (PAF) receptors. Intracellular adaptor proteins then facilitate membrane invagination and the formation of clathrin-coated vesicles that direct cargo to the endolysosomal pathway (310, 311). NTHI interacts with a number of host cell receptors to initiate clathrin-mediated endocytosis. NTHI strains expressing ChoP bind the PAF receptor to mediate entry into epithelial cells (312, 313). The β-glucan receptor also facilitates the non-opsonic entry of NTHI into epithelial cells and macrophages; however, the NTHI binding partner for this receptor remains uncharacterized (305).

1.12.4 Macropinocytosis

In contrast to receptor-mediated uptake, bacterial entry into host cells can also occur non-specifically through macropinocytosis. Often called “cell drinking”, macropinocytosis is the large-scale uptake of extracellular fluid and nutrients. Depending on cell type, macropinocytosis can be constitutively active or regulated by growth factor activation of receptor tyrosine kinases (314). In all cell types, the formation of macropinosomes requires membrane ruffling through actin polymerization at the cell surface (314). Like vesicles formed by raft-dependent endocytosis, the intracellular fate of the macropinosome is mercurial; it may remain isolated from the endolysosomal pathway or fuse back into the plasma membrane depending on cell type (315, 316).

Colocalization studies provide evidence that NTHI can enter epithelial cells by macropinocytosis and reside within macropinosomes after initial uptake (171). However,

34 the role of macropinocytosis in NTHI trafficking within the cell is not well understood, and, thus, represents an important avenue for further investigation.

1.13 Intracellular Trafficking

1.13.1 The endolysosomal pathway

Following invasion into the host cell, bacteria may be shuttled through the endolysosomal pathway and degraded. The endolysosomal pathway is critical for cellular homeostasis and functions in the transport of host molecules, recycling, signaling, and cytokinesis (317, 318). The first step in endolysosomal trafficking is the transport of cargo from endocytic vesicles – which arise from the clathrin-dependent or - independent internalization pathways mentioned in the above section – by fusion with the early endosomes (EEs). The EEs can be thought of as highly dynamic sorting stations with a specialized structure. EEs are typically located at the cell periphery (319) and consist of a central vacuole surrounded by multiple tubules that extend outward.

There are many EE-specific proteins, which aid in early endocytic events such as tethering, fusion, and motility, including the Rab4 and Rab5 GTPases and the early endosomal antigen 1 (EEA1) protein (320). The vacuolar structure of the EE is referred to as the sorting endosome (SE) because it directs cargo for either degradation or recycling. A portion of the SE is covered by a cytoplasmic clathrin coat which aids in cargo sorting (321); molecules that are destined for transportation to the lysosome are localized within this clathrin coat, while proteins to be recycled are localized within the tubules extending from the EE, also known as the recycling endosome (RE) (321-324).

The next step in the endolysosomal pathway is the transition from an EE to a late endosome (LE). This can occur through maturation of the SE/EE, and the exchange of

Rab5 for Rab7 is a signal that this process has occurred (325). LEs are oval shaped 35 organelles that retain some of the clathrin coat of the SE in an irregular distribution that may indicate that LEs retain some ability to sort cargo (321). LEs then migrate to a perinuclear area to fuse with lysosomes for the final transport step in the endolysosomal pathway (326).

Endolysosomal fusion is a process that has been well studied. It is now widely accepted that fusion requires a physical interaction between LEs and lysosomes (327), and that the fusion event can be permanent or transient (termed a “kiss and run” event)

(327). The fusion process itself has three steps: tethering, the formation of a trans-snap receptor (SNARE) bridge, and membrane fusion (328). Tethering refers to the extension of links spanning >25 nm from the membranes of each organelle (328). The trans-

SNARE bridge is then formed when the 16-turn helix of one SNARE combines with three additional SNAREs to generate a SNAREpin, or a four-helix bundle which is essential for membrane fusion (329). The trans-SNARE bridge, coupled with Ca2+ release, facilitates the phospholipid bilayer fusion of the LE and lysosomal membranes (330).

Lysosomes are the primary degradative components of the host cell, and degradation of bacteria, glycoproteins, oligosaccharides, and other components is driven by a variety of hydrolytic lysosomal enzymes. These enzymes include proteases, lipases, nucleases, phospholipases, and phosphatases, among others (331). Optimal activity of lysosomal enzymes occurs under acidic conditions, and vacuolar ATPases maintain a pH of >5 in the lysosome (332). In addition to the hydrolytic enzymes, lysosomal membrane proteins LAMP1, LAMP2, and LAMP3 are integral to lysosomal activity by protecting the membrane from damage, regulating movement, and directing lysosomal distribution (332-334).

1.13.2 Bacterial Escape of the Endolysosomal System

To survive intracellularly, bacterial pathogens must find ways to evade, escape from, or modify endolysosomal organelles. Several professional intracellular pathogens have elegant strategies for modulating vacuolar transport to evade delivery to the lysosomes. Chlamydia trachomatis uses the effector protein CT147 as a mimic of EEA1 to tether endosomes together and block normal endolysosomal fusion (335). Legionella pneumophilia produces multiple proteins that can subvert and control Rab protein activity (336). Salmonella enterica uses secreted effector proteins to regulate actin polymerization and hijack Rab proteins for optimal perinuclear positioning of the

Salmonella-containing vacuole (337, 338). Intracellular pathogens also employ diverse mechanisms for escaping the endolysosomal pathway, including lysis of the host cell, protrusion and engulfment by neighboring cells, and induction of apoptotic cell death

(301, 339, 340).

A third mechanism of intracellular survival is the escape of endolysosomal trafficking, without inducing host cell death, which allows the bacteria to ultimately reside inside the host cell cytoplasm. Cytosolic pathogens can actively escape from vacuoles, replicate in the cytoplasm, and modulate host immune responses from the cytosolic niche (341). Listeria monocytogenes is a classic example of a cytosolic pathogen. L. monocytogenes produces the pore-forming toxin listeriolysin O (LLO) which binds cholesterol in the vacuolar membrane. This binding allows LLO to insert and form pores in the vacuole. In addition to tight regulation of LLO production by L. monocytogenes itself, LLO also requires activation by the host factor γ-interferon-inducible lysosomal thiol reductase within the vacuole (342). A second secreted factor, phosphatidylinositol- specific phospholipase C, also participates in L. monocytogenes escape from the vacuole into the cytoplasm (343). Another classic example of a cytosolic pathogen is 37

Francisella tularensis which escapes from an acidified vacuole before fusion with the lysosome (344, 345). Escape by F. tularensis requires a vacuolar ATPase pump and multiple Francisella effectors with an unknown role in vacuolar lysis (346-348). A unique facet of the Francisella intracellular lifestyle is that following replication in the cytosol, the microbe re-enters a membrane-enclosed vacuole, termed the Francisella-containing vacuole, to evade the host immune factors (349).

NTHI has not historically been considered an intracellular pathogen. NTHI has been observed within acidic subcellular compartments of the endolysosomal pathway in mouse alveolar macrophages, human alveolar epithelial cells, and normal human bronchial epithelial cells (55, 303, 350). Endolysosomal trafficking of NTHI terminates in the active killing of NTHI within lysosomes in epithelial cells (55). In addition to the endolysosomal pathway, bacteria within host cells have also been shown to interact with components of the autophagy pathway. Autophagy is a cellular recycling pathway which sequesters intracellular components in double-membrane vesicles that will ultimately fuse with the lysosomes to form autolysosomes. NTHI does not colocalize with markers of the autophagy pathway, even under conditions where autophagy has been exogenously induced in the host cell (303).

1.14 Intracellular Bacterial Communities

Recent studies have revealed the influence of environmental factors on the intracellular lifestyle of NTHI. Following a period of transient heme-iron restriction, NTHI can exist in a densely-packed population that appears to be free in the cytoplasm of chinchilla middle ear epithelial cells in vitro up to 72 hours post-inoculation. This suggests NTHI has a mechanism to escape or evade the endolysosomal pathway and survive intracellularly (60). These populations of intracellular NTHI resemble the IBCs 38 that are a hallmark of UPEC pathogenesis (351-357). IBCs provide bacteria access to nutrients, protect them from antibiotics and innate immune effectors, and can serve as a reservoir for recurrent infection (353, 358-362). Although first described for UPEC, IBC formation has now been observed in Klebsiella, Proteus, and Helicobacter pylori (363-

366).

The stages of IBC formation in UPEC have been extensively characterized and include early IBC formation, intracellular differentiation, a mid-IBC stage, egress, and second-generation IBC formation (361). The process begins with the formation of an early IBC following UPEC invasion into the cytoplasm of bladder epithelial cells. Within the cytoplasm, UPEC replicates in both coccoid and filamentous morphologies, and the increasing number of bacteria eventually causes the epithelial cell to become engorged.

Next, growth of intracellular UPEC slows, and the bacteria organize into a biofilm-like structure that fills the volume of the cell, referred to as a mid-IBC or a pod. Eventually the pod lyses the host cell, and the resident UPEC are then able to invade neighboring host cells to promote persistence through establishing second-generation IBCs or quiescent intracellular reservoirs that can later initiate reinfection (353, 354, 357, 367).

The IBC process is an efficient way for UPEC to survive and cause disease, as only a single UPEC bacterium is necessary to initiate early-IBC formation, and the resulting cascade of events leading to egress and reinfection (368).

1.15 Vaccine Approaches Against NTHI

Despite the routine use of antibiotics, NTHI continues to cause morbidity and mortality in populations of patients at risk for NTHI infection, such as otitis-prone children or patients with frequent exacerbations of COPD or CF. Development of a vaccine that is efficacious against the majority of NTHI strains, and acts to protect against a spectrum of 39

NTHI infections, is the ultimate goal. Although promising vaccine candidates have been identified, NTHI strains have high antigenic heterogeneity which, in addition to the lack of polysaccharide capsule for use as a target as in vaccines against Hib, has stalled vaccine development. Current approaches to NTHI vaccine development are focused on identifying NTHI surface proteins that are highly conserved among strains, are not subject to phase variation, and are immunogenic in both infants and adult patients while inducing a protective response (369). Ongoing studies have identified several possible

NTHI vaccine antigens, including P2, P5, P6, Protein D, and the PilA subunit of the type

IV pilus (210, 370-379). Vaccine development for diseases caused by NTHI is an active area of research, but the challenge presented by the heterogeneity of NTHI surface antigens in identifying a candidate that reliably induces the necessary protective response has slowed progression of candidate vaccines to clinical trials. There is also a need for the development of alternative therapeutic approaches to antibiotics for treatment and prevention of NTHI-mediated diseases in the absence of a vaccine.

1.16 Remaining Questions

Many facets of NTHI pathogenesis are still not fully understood, and, thus, provide exciting areas for continued scientific study. An overarching question is how

NTHI uses nutritional cues to navigate diverse lifestyles including biofilm formation and life inside host epithelial cells, and how adaptation of NTHI to these different lifestyles promotes persistence, which is critical to the development of chronic or recurring infections. It is increasingly evident that there is a clear intracellular niche for NTHI that has not been fully appreciated as an integral part of NTHI pathogenesis, and must be considered when designing therapeutics or vaccine candidates. It will be important to determine how transient heme-iron restriction influences NTHI invasion into host cells, 40 how and at what step intracellular NTHI escapes or evades the endolysosomal pathway, and whether cytosolic NTHI can survive long-term in vivo during chronic infections such as OM or recurrent COPD exacerbations. Further, determining which host cellular signaling pathways are co-opted by intracellular NTHI, and by what mechanisms, will increase our ability to target intracellular NTHI.

Finally, it will also be critical to understand the bacterial side of this specific host- pathogen interaction, including the identification of bacterial factors that promote (i) NTHI in vitro and in vivo persistence following heme-restriction, (ii) bacterial escape from the endolysosomal pathway, and (iii) nutrient acquisition and growth in the host cell cytosol.

1.17 Project Goals and Hypotheses

Coaction of host microenvironmental cues and microbial pathogenesis is a critical contributor to the development of chronic and recurring disease. The ability of bacteria to sense and respond to fluctuations in the level of available nutrients in host niches aids in the transition from a commensal to opportunistic pathogen. As a heme- iron auxotroph, NTHI must tightly control acquisition of this essential nutrient from the host and, in concert, regulate metabolism and virulence accordingly for survival.

Ascension of the Eustachian tube and the initial stages of OM in the sterile middle ear are periods during which NTHI experiences heme-iron restriction due to nutritional immunity. As OM progresses, inflammation and immune cell infiltrate naturally expand the availability of heme-iron in the middle ear environment. This transient heme-iron restriction promotes a shift in NTHI pathogenesis that is characterized by changes in biofilm architecture, a tempered host immune response, and the formation of IBCs in epithelial cells (60). We hypothesize that the ability to modulate interactions with host cells in dynamic lifestyles, including extracellular biofilms or invasive populations, drives 41

NTHI persistence and impacts disease severity. Moreover, we hypothesize that transient heme-iron restriction promotes a phenotype for NTHI that is like that of a professional intracellular pathogen, characterized by the ability to escape host degradative pathways and persist in the host in intracellular communities.

The major goals of the studies included in this dissertation were as follows: (i)

Identify bacterial factors that promote the long-term survival of NTHI under transient heme-iron restricted conditions in vitro, (ii) Determine the effect of heme-iron restriction on NTHI internalization and intracellular trafficking in epithelial cells, (iii) Elucidate the mechanism underlying the formation of IBCs by NTHI, and (iv) Evaluate the contribution of IBC formation to NTHI persistence and long-term survival during preclinical OM.

Chapter 2. Microevolution in response to transient heme-iron restriction enhances

intracellular bacterial community development and persistence*.

Abstract

Bacterial pathogens must sense, respond and adapt to a myriad of dynamic microenvironmental stressors to survive. Adaptation is key for colonization and the long- term ability to endure fluctuations in nutrient availability and inflammatory processes. We hypothesize that strains adapted to survive nutrient deprivation are more adept for colonization and establishment of chronic infection. In this study, we detected microevolution in response to transient nutrient limitation through mutation of icc. This mutation results in decreased 3',5'-cyclic adenosine monophosphate phosphodiesterase activity in nontypeable Haemophilus influenzae (NTHI). In a preclinical model of NTHI- induced otitis media (OM), we observed a significant decrease in the recovery of effusion from ears infected with the icc mutant strain. Clinically, resolution of OM coincides with the clearance of middle ear fluid. In contrast to this clinical paradigm, we observed that the icc mutant strain formed significantly more intracellular bacterial communities (IBCs) than the parental strain early during experimental OM. Although the number of IBCs formed by the parental strain was low at early stages of OM, we observed a significant increase at later stages that coincided with absence of

* A version of this chapter was submitted for publication in PloS Pathogens as follows: Hardison RL, Harrison A, Wallace RM, Heimlich DR, O’Bryan ME, Sebra RP, Pinkett HW, Justice SS and Mason KM. Microevolution in response to transient heme-iron restriction enhances intracellular bacterial community development and persistence. PloS Pathog 14(10): e1007355. 43 recoverable effusion, suggesting the presence of a mucosal reservoir following resolution of clinical disease. These data provide the first insight into NTHI microevolution during nutritional limitation and provide the first demonstration of IBCs in a preclinical model of chronic OM.

Significance

Nontypeable Haemophilus influenzae (NTHI) inhabits diverse niches in the host.

The ability to adapt to new microenvironments is consistent with the predominance of

NTHI as a causative agent of otitis media (OM) in children. We evaluated the microevolution of NTHI as it relates to adaptation and persistence in response to nutrient limitation. We identified a naturally-occurring mutation that enhances NTHI persistence and formation of intracellular bacterial communities (IBCs) in a pre-clinical model of chronic OM. The presence of IBCs during OM provides the first opportunity to evaluate the role of intracellular populations in chronicity and quiescence as a new paradigm for recurrent OM. This model provides a new platform to identify novel therapeutics for this highly prevalent and costly infectious disease.

2.1 Introduction

The host sequesters essential micronutrients (i.e. metals) as a mechanism to limit bacterial infection (380). Bacteria that require exogenous nutrients for growth are subjected to stresses that necessitate adaptation for survival. Bacteria can adapt to these stresses through transcriptional, epigenetic and genetic mechanisms (380).

Identification of genes subject to microevolution provides insight into the nutritional and

44 environmental stresses incurred during infection as well as new therapeutic targets to quell infection.

Nontypeable Haemophilus influenzae (NTHI) is a heme auxotroph and must obtain this essential nutrient from the environment (381). The presence of heme- containing compounds is sufficient to support polymicrobial growth in the nasopharynx

(382). During ascent from the nasopharynx to the middle ear, NTHI experiences fluctuations in nutrient availability, particularly heme-iron. The lack of sufficient heme-iron in the middle ear is indicated by the increase in the number of iron-uptake genes in NTHI strains isolated from children with otitis media [(OM); (383)]. Furthermore, iron-uptake genes are transcribed in NTHI early during experimental OM (294, 384). We previously demonstrated that transient heme-iron restriction of NTHI potentiates changes in bacterial morphology, biofilm architecture, disease severity and provides a survival advantage in the preclinical model of OM (60).

NTHI inhabits diverse niches in the host including the nasopharynx, middle ear, sinus, eye and lung. At these sites, NTHI can be planktonic, in a biofilm or associated with mucosa (385). NTHI has also been observed within intracellular niches in vitro in cultured lung and middle ear epithelial cells, as well as in clinical biopsies of adenoid and middle ear tissues from children with a history of OM (59, 60, 303, 386, 387).

New paradigms for recurrent infections were revealed through the identification of intracellular bacterial communities (IBCs) in the cytoplasm of epithelial cells. This complex intracellular lifestyle, first described for UPEC (388), provides access to nutrients and protection from antibiotic therapies and innate immune effectors (358,

361). The IBC pathway is critical for acute urinary tract infection and culminates in a

45 latent intracellular population that serves as a bona fide reservoir for recurrent infection

(361). Although first described in preclinical models, IBCs are present in bladder biopsies and in the urine of patients with urinary tract infections caused by UPEC (388).

IBCs of Klebsiella and Proteus have also been observed in the urinary tract (363, 364) and H. pylori in the stomach (366), in preclinical models of urinary tract infection and gastritis, respectively. We recently observed the formation of NTHI IBCs in cultured chinchilla middle ear epithelial cells in response to transient heme-iron limitation (60).

This observation suggests a potential role for intracellular populations of NTHI during

OM and reveals new potential mechanisms for disease chronicity, persistence and recurrence that remain underexplored.

The ability to adapt to new microenvironments is consistent with the predominance of NTHI as a causative agent of OM, the most common bacterial infection in children (389-391). Longitudinal sampling of patients with COPD revealed the same

NTHI strains, despite intermittent periods of negative culture (54). In this study, we evaluate the microevolution of NTHI associated with adaptation and persistence in response to nutrient limitation. In long-term stationary phase following transient heme- iron restriction, we observed a single nucleotide polymorphism in icc that abolishes 3’,5’- cyclic adenosine monophosphate (cAMP) phosphodiesterase activity. In the preclinical model for NTHI-induced OM, the icc mutant significantly shifts infection kinetics, with decreased middle ear fluid, a clinical indicator of disease resolution. However, the icc mutant was still associated with middle ear tissues at levels similar to the parental strain, suggesting that the presence of fluid and bacterial burden of the middle ear can be uncoupled. Although the bacterial burdens were similar, there was a significant increase in IBC formation by the icc mutant strain early during disease while formation of IBCs by

46 the parental strain occurred gradually, indicating that NTHI adaptation to an intracellular niche occurs during active infection. Using UPEC as the paradigm, this observation suggests that NTHI intracellular populations may serve as a reservoir and contribute to both chronic and recurrent disease associated with OM.

2.2 Materials and methods

Strains and media

Bacterial strains used in this study are listed in Table 2.1. NTHI strain 86-028NP is a minimally passaged clinical isolate which has been sequenced and extensively characterized in the chinchilla model of human otitis media (59, 103, 113, 392, 393).

NTHI strain RM33 was obtained via heme-iron restriction and isolated on day 33 of long- term culture of NTHI strain 86-028NP, as described below.

The 86-028NPrpsL icc63A>T mutant strain was generated using a recombineering method developed for Haemophilus (394). The icc gene in 86-

028NPrpsL was deleted and replaced by a cassette containing a spectinomycin resistance gene and a copy of rpsL from Neisseria gonorrhoeae (rpsLNg). The icc gene including approximately 1kb of DNA flanking sequences were cloned into pGEM-T Easy

(Promega, Madison, WI). The icc63A>T mutation was introduced into icc using

QuikChange (Agilent, Santa Clara, CA). This construct was linearized by SapI digestion and introduced into 86-028NPrpsL ∆icc::spec-rpsLNg using the MIV method (395). After transformation, clones containing chromosomal icc63A>T were selected on chocolate agar supplemented with 1000µg/ml streptomycin. Loss of the cassette was confirmed by screening clones on chocolate plates containing 200µg/ml spectinomycin and the presence of icc63A>T was confirmed by sequencing.

Strain Description Reference or source

86-028NP Nontypeable H. influenzae strain from a Harrison et al (2005) child with chronic otitis media 86-028NPrpsL Streptomycin-resistant derivative of 86- Carruthers et al (2012) 028NP RM12 Derivative of 86-028NP isolated after 12 This study days in static culture RM13 Derivative of 86-028NP isolated after 13 This study days in static culture RM32 Derivative of 86-028NP isolated after 32 This study days in static culture RM33 Derivative of 86-028NP isolated after 33 This study days in static culture AR12 Derivative of 86-028NP isolated after 12 This study days in static culture AR15 Derivative of 86-028NP isolated after 15 This study days in static culture AR19 Derivative of 86-028NP isolated after 19 This study days in static culture AR20 Derivative of 86-028NP isolated after 20 This study days in static culture 86-028NP(pSPEC1) 86-028NP transformed with pSPEC1 This study RM33(pGZRS-39A) RM33 transformed with pGZRS-39A This study 86-028NPrpsL icc63A>T Derivative of 86-028NPrpsL with nt 63 of This study icc changed from an A to a T (63A>T) 86-02NPrpsL∆icc Derivative of 86-028NPrpsL with a deletion This study mutation of icc E. coli BL21 E. coli BL21 (DE3) transformed with This study (DE3)(pET15b-icc) pET15b-icc E. coli BL21 E. coli BL21 (DE3) transformed with This study (DE3)(pET15b-icc63A>T) pET15b-icc63A>T NTHI1521MEE Nontypeable H. influenzae strain from a Gift from L.O. Bakaletz child with chronic otitis media NTHI1521NP Nontypeable H. influenzae strain from a Gift from L.O. Bakaletz child with chronic otitis media

Table 2.1. Bacterial strains used in this study.

Our first approach was to initially determine the role of icc during experimental

OM. To this end, parallel cohorts of chinchillas received either a single infection or a coinfection of the parent and RM33. To distinguish the parent from RM33 in the competition model, we engineered 86-028NP and RM33 to carry antibiotic resistance markers as previously described (77). From this first series of studies, we observed differences in host responses in the ears infected with both strains as compared with the ears infected with either of the strains independently, suggesting that there were dominant phenotypes that resulted in premature clearance of the bacteria in the coinfection model. Therefore, only single infections were used to determine the kinetics of infection of RM33, with either marked or unmarked strains.

For routine culturing on agar plates, strains were grown on chocolate agar

(Becton Dickinson, Sparks, MD). Marked strains were grown on chocolate agar plates supplemented with 200 µg/mL of spectinomycin/ml or 20 µg/mL of kanamycin. All growth was at 37°C in 5% CO2, unless stated otherwise. For routine liquid culture, cells were grown in BHI supplemented with 2 µg/mL of NAD and 2 µg/mL of heme (sBHI). For persistence assays bacteria were grown in defined iron source (DIS) medium in the presence or absence of 2 µg/mL heme (112).

For expression of recombinant Icc, E. coli BL21 (DE3) cells (Thermo Fisher

Scientific) containing plasmid vectors were grown in L broth supplemented with 50

µg/mL ampicillin.

Environmental heme-iron restriction and long-term survival

Environmental heme-iron restriction was performed as previously described (60).

Briefly, 86-028NP was grown overnight on chocolate agar (Becton Dickinson, Sparks,

MD). Individual colonies were suspended in chelated DIS medium to an OD490 of 0.65, then diluted 10-fold into 15 mL round bottom glass tubes (nitric acid-washed to remove all metals) containing DIS medium with either 0, or 2 µg/mL heme (Millipore Sigma,

Billerica, MA). Following 24 hours incubation, cultures were adjusted to an OD490 of 0.37 in DIS medium and diluted 10-fold into 5 ml DIS medium containing 2 µg/mL heme. For this study, to evaluate long-term persistence, cultures were incubated statically for up to

39 days with no fresh medium added. Every 24 hours, cultures were vortexed at 300 rpm for 2 seconds. Cultures were serially diluted and plated on chocolate agar to assess viability. Each day, bacteria were collected from plates and cryopreserved by resuspending in 1 mL 20% glycerol, 0.8% skim milk and stored at -80°C to generate a series of isolates (RM series). RM33 (isolate collected on day 33) was used in subsequent experiments. A second, parallel and independent long-term survival assay was completed using the same methodology and a second series of isolates (AR Series) was collected.

Whole genome sequencing of long-term survival isolates

At each time point the colonies from the enumeration plates were pooled and frozen for long-term storage. To purify genomic DNA, a loop full of cells from the frozen stock was grown on chocolate agar overnight. The resultant colonies were then pooled, lysed, and genomic DNA was purified using a Gentra Puregene Yeast/Bact. Kit

(QIAGEN, Germantown, MD). Sequencing libraries were prepared and sequenced using paired end 300bp chemistry on the MiSeq platform to high coverage on a single flow cell

(Illumina, San Diego, CA). Nucleotide variants were identified using the Churchill algorithm (396) and the percentage of discordant reads as compared to the 86-028NP reference genome was calculated. Single nucleotide variants were independently

50 validated by Sanger sequencing. icc was amplified by PCR using primers AH0422 and

AH0423 (Table 2.2). The amplicons were then purified using a QIAquick PCR

Purification Kit (QIAGEN) and sequenced by Eurofins Genomics (Louisville KY), using the amplification primers as sequencing primers. Sequence data was assembled using

SeqMan Pro (DNASTAR, Madison, WI) and nucleotide variants identified through comparison with the parental gene sequence.

Tertiary structure prediction and validation of 3',5'-cyclic adenosine monophosphate phosphodiesterase Icc

SWISS- MODEL was used to model the 3',5'-cAMP phosphodiesterase Icc from 86-

028NP using the automated modeling mode (397). Templates were populated based on a search for evolutionarily related structures matching the target sequence, with the model built based on the target-template alignment using ProMod3 Version 1.1.0 and quality assessed. The quality of the predicted model was further evaluated using the structure assessment program SAVES (https://services.mbi.ucla.edu/SAVES/),

PROCHECK (398), ERRAT (399), VERIFY_3D (400), PROVE (401), and

Ramachandran Plot. Computation models were prepared and rendered with the PyMOL

Molecular Graphics System (DeLano Scientific, San Carlos, CA, 2002).

Purification of recombinant Icc

The coding sequence of icc from 86-028NP was PCR amplified with a primer that introduced an NdeI restriction site at the 5’ end of the gene (AH0440, Table 2.1) and a second primer that introduced the BamHI restriction site at the 3’ end of the gene

(AH0441, Table 2.1). The latter primer also removed the stop codon to produce Icc with a C-terminal His-tag. Using NdeI and BamHI, the amplicon was cloned as a first codon fusion in pET15b (Millipore Sigma) and transformed into E. coli BL21 (DE3) cells 51

Table 2.2. Primers used in this study. The nucleotides in bold depict a Ndel (AH0440) and a BamHI (AH0441) restriction site.

(Thermo Fisher Scientific). The point mutation at nucleotide 63 in icc from RM33 was introduced into pET15b-icc by QuikChange. This construct was also transformed into E. coli BL21 (DE3). To express and overproduce recombinant Icc (rIcc), bacteria were grown in L broth at 37˚C with shaking at 200 rpm. When bacteria were in mid- exponential phase, IPTG (Thermo Fisher Scientific) was added to a final concentration of 1mM and bacteria were incubated at 37˚C with shaking at 200 rpm for an additional four hours. Bacteria were then harvested by centrifugation at 6,000xg for 15 minutes, resuspended in lysis buffer (50 mM NaH2PO4, 300 mM NaCl, pH 8.0) and lysed by one passage through a high-pressure cell (25,000 psi; One Shot Model, Constant Systems

Ltd., Kennesaw, GA). Insoluble material was removed by centrifugation at 16,000xg for

20 minutes and rICC purified on High Affinity Ni-Charged Resin following the manufacturer’s protocol (GenScript, Piscataway, NJ). Protein concentration was assessed using Coomassie Plus (Bradford) Assay Reagent (Thermo Fisher Scientific) then resolved on a 4-20% Mini-PROTEAN TGX Precast Gel (Bio-Rad, Hercules, CA) and silver stained (Pierce Silver Stain Kit, Thermo Fisher Scientific) to assess protein purity.

Isolation of soluble proteins

Bacteria were grown in sBHI at 37˚C with shaking at 180 rpm. When bacterial cells were in mid-exponential phase, they were harvested by centrifugation at 6,000xg for 15 minutes then resuspended in 10mM HEPES, pH 7.4 and lysed by one passage through a high-pressure cell (25,000 psi; One Shot Model, Constant Systems Ltd).

Insoluble material was removed by centrifugation at 16,000xg for 20 minutes, the clarified supernatant collected and protein concentration assessed using Coomassie

Plus (Bradford) Assay Reagent (Thermo Fisher Scientific).

Analysis of Icc function by phosphodiesterase activity

To measure cAMP dependent phosphodiesterase activity, purified parent and mutant recombinant Icc proteins were buffer exchanged and concentrated into 50mM

Tris pH7.6, 0.1mM DTT, 10µM FeCl3 (129) using an Amicon Ultra 4 10K Ultracel

Centrifugal Filter (Millipore Sigma) after which protein concentration was assessed using

Coomassie Plus (Bradford) Assay Reagent. cAMP phosphodiesterase activity was measured using a PDELight HTS cAMP phosphodiesterase Kit (Lonza, Morristown, NJ).

Thirty-nine microliters of 10µM cAMP was mixed with 2 ng of recombinant Icc, to give a final volume of 40 µL. Samples were incubated for 30 minutes at room temperature then

20 μL of AMP detection reagent was added. After a further incubation for 10 minutes at room temperature, luminescence was quantified using a Synergy Hybrid H1 Reader

(BioTek, Winooski, VT; default luminescence setting with a 0.1 second integration time).

Three biological replicates were assessed in technical triplicate and statistical significance was determined using a one–tailed Student’s t test (Graph Pad Prism, La

Jolla, CA).

Analysis of phosphodiesterase activity in cell lysates

cAMP-dependent phosphodiesterase activity was quantified in lysates from 86-

028NP and RM33 that contained both cytoplasmic and periplasmic proteins. Using the protocol outlined for rIcc, the activity in 100 ng of lysate was assessed. Four biological replicates were tested in technical triplicate and statistical significance was determined using a one–tailed Student’s t test (Graph Pad Prism).

Transformation efficiency experiments

To assess the transformation efficiency of the parent and RM33 isolates, we used recombineering (394) to generate a mutation in a putative sRNA–encoding gene,

54 sRNA121, located in the intergenic region between NTHI0351 and NTHI0353. sRNA 121 and approximately 1kb of flanking DNA was cloned into pGEM-T Easy. sRNA121 was then deleted and replaced by a cassette comprising a spectinomycin resistance gene and an rpsLNg gene. This construct, when linearized by NcoI digestion, was used to determine strain transformability using a modified MIV method in the presence or absence of cAMP (395). After transformation, bacteria were grown on chocolate agar to enumerate the total viable population, as well as on chocolate agar that contained

200µg/ml spectinomycin, to enumerate the transformed population. The transformation efficiency is reported as the percentage of transformed bacteria from the total number of viable bacteria. Seven biological replicates were assayed, and statistical significance was determined using a two–tailed Student’s t test (Graph Pad Prism, La Jolla, CA).

Preclinical models of Otitis Media

Ethics Statement. Animal experiments were completed in adherence to the accredited conditions in the Guide for the Care and Use of Laboratory Animals of the National

Institutes of Health. The protocol was approved by the Institutional Animal Care and Use

Committee at the Research Institute at Nationwide Children’s Hospital (Welfare

Assurance Number A3544-01), AR13-00026. All experimental procedures were performed under anesthesia (xylazine and ketamine administration) and all efforts made to minimize suffering.

Time course to determine kinetics of single strain infections. NTHI strains 86-

028NP(pSPEC1) and RM33(pGZRS-39A) were grown on chocolate agar overnight.

NTHI were resuspended in 0.9% (w/v) sodium chloride in non-pyrogenic sterile water

(Pfizer, New York NY) to an optical density of 0.65 measured at 490 nm and diluted for inoculation. Two cohorts of five chinchillas each were transbullarly inoculated in each ear

55 with 300µL containing either 934 colony forming units (CFUs) of 86-028NP(pSPEC1) or

1070 CFUs of RM33(pGZRS-39A). On days 7, 10, and 15 post-inoculation, middle ear effusions were collected by epitympanic tap from the inferior bullae. Middle ear effusions were serially diluted and plated on chocolate agar as well as on chocolate agar supplemented with 200 µg/mL spectinomycin or 20 µg/mL kanamycin. On day 17 post- inoculation, chinchillas were sacrificed and the middle ear mucosa tissues were removed, weighed, homogenized and serially diluted and plated on antibiotic- supplemented chocolate agar plates. Viable CFUs from middle ear effusions and tissue homogenates were counted to determine burden of infection. The bacterial burden of tissue homogenates was normalized for total tissue weight.

Visualization of intracellular bacterial communities in middle ear mucosal tissue.

To quantitate the formation of IBCs as a result of infection with 86-028NP or RM33, two cohorts of 9 chinchillas each were transbullarly inoculated with 640 CFUs of 86-028NP or 386 CFUs RM33. On days 7, 14, and 28, chinchillas were sacrificed and the middle ears removed, fixed in 4% paraformaldehyde in Dulbecco’s Phosphate Buffered Saline

(DPBS), decalcified in 0.35 M Tris/EDTA solution and embedded in paraffin. Thin sections from fixed middle ears were deparaffinized in xylene and antigen-retrieval was performed as previously described (298) followed by treatment with 0.01% sodium borohydride (Thermo Fisher Scientific) in DPBS for 10 minutes to inhibit inherent fluorescence due to aldehyde production. Slides were then incubated for 10 minutes in

CAS Block (Thermo Fisher Scientific) to prevent non-specific staining followed by a 30- minute incubation with Image-iT FX Signal Enhancer (Thermo Fisher Scientific). NTHI were labeled with a chinchilla primary antibody raised against total outer membrane proteins [OMP; (59)], diluted 1:25 in CAS Block overnight at 4°C. NTHI was visualized

56 using Protein A conjugated with Alexa Fluor 488 (Thermo Fisher Scientific) diluted 1:100 in CAS Block and incubated at room temperature for one hour. Host cell membranes were visualized with wheat germ agglutinin conjugated with Alexa Fluor 594 (Thermo

Fisher Scientific) diluted 1:200 in DPBS. DNA was counterstained with Hoescht 33342

(Thermo Fisher Scientific) diluted 1:300 in DPBS. Photobleaching was reduced by the addition of ProLong Gold Antifade (Thermo Fisher Scientific) prior to addition of the coverslip and tissues were imaged with an Axiovert 200M inverted epifluorescence microscope equipped with the Apotome attachment for improved fluorescence resolution

(Carl Zeiss, Inc., Thornwood, NY). IBCs were quantified by manually counting the total number of infected cells in each of three thin sections for each cohort. Statistics were performed using a two-tailed Student’s t test (GraphPad Prism, La Jolla, CA).

2.3 Results

Transient heme-iron restriction promotes a cyclical pattern of viability and microevolution during stationary phase

Bacterial subpopulations that arise stochastically contribute to persistence during infection. Formation of these populations can be enhanced by antibiotic exposure, oxidative stress, reduced indole signaling and nutrient limitation [as reviewed in (402)].

Based upon previous observations, we hypothesized that transient heme-iron restriction would invoke survival advantages to NTHI (60). As a first approach to elucidate the physiological changes that contribute to persistence following transient restriction of heme-iron, we evaluated the longevity of NTHI in stationary phase. The prototypical

NTHI strain 86-028NP was cultured in DIS medium in the presence or absence of 2

µg/mL heme-iron for 24 hours, then subcultured into DIS containing 2 µg/mL heme-iron,

57 resulting in two parallel cultures, one continuously exposed to heme-iron and one transiently restricted of heme-iron (Figure 2.1A). The continuously exposed and transiently restricted NTHI cultures were incubated statically at 37C, without the addition of fresh medium, and viability was assessed each day. The colonies obtained each day were pooled and cryopreserved to create a series of isolates (termed the “RM

Series”) for further investigations. The NTHI cultures that were transiently restricted of heme-iron were viable for the entire duration of the experiment and exhibited repeated cycles of increasing and decreasing viability with a 6-7-day periodicity (Figure 2.1B). In marked contrast, the NTHI cultures continuously exposed to heme-iron were not viable beyond 17 days of incubation.

The cycling of growth during stationary phase can include the sequential acquisition and loss of single mutants defined as growth advantage in stationary phase

[GASP; (403)]. To determine if mutations contributed to the observed phenotype, the whole genome sequence of RM33 isolated during the last cycle, day 33, was compared to the input strain, 86-028NP. Within the entire RM33 genome, only one single nucleotide variant was identified at position 63 of icc, which encodes a 3’,5’-cAMP phosphodiesterase. This mutation, an adenine to thymine nucleotide change, converts the 21st amino acid in Icc from an aspartic acid to a valine. This base change was confirmed by Sanger sequencing (Figure 2.1D). To determine if other mutations were associated with the observed cycling, bacteria isolated on days 12, 13 and 32 (RM12,

RM13 and RM32, respectively) were also subjected to whole genome sequencing. No additional mutations were identified at these earlier time points. We further observed that the sequences were derived from a mixed population of cells; the parental and the D21V mutation, with an enrichment for the icc mutant strain throughout the experiment (Figure

Figure 2.1. Transient heme-iron restriction promotes a cyclical pattern of viability and microevolution during stationary phase. (A) Schematic representation of environmental heme-iron restriction and extended incubation of stationary phase cultures. (B) Viability was determined every 24 hours for 39 days. Arrow indicates the culture selected for further investigation, termed RM33. Asterisks denote cultures selected for whole genome sequencing. (C) The sequence reads that exhibit the single nucleotide polymorphism (SNP) are reported as percentage of the total reads for each of the days. (D) Sanger sequencing electropherograms validate the mutation [Adenine (A) → Thymine (T)].

2.1C). The presence of the D21V mutation and the increase in mutant genotypes was also confirmed by Sanger sequencing (Figure 2.1D). On day 12, a single peak representing the parental adenine residue could be resolved. In contrast, although the adenine peak predominated in strains RM13 and RM32, a smaller peak representing a thymine could be resolved. Finally, in strain RM33 only a peak representing a thymine could be resolved.

A second independent experiment was performed to confirm that transient heme-iron restriction promotes long term survival during stationary phase. As with the previous study, viability of NTHI transiently restricted of heme-iron was observed for the duration of the experiment. In contrast, NTHI continuously exposed to heme-iron demonstrated reduced longevity in stationary phase (Figure 2.2A). In addition to the increased longevity observed for the transiently restricted NTHI, the 6-7-day periodicity of the cyclical increases and decreases in viability were also observed. To determine whether mutations also contributed to the phenotype in this series, whole genome sequencing was performed on the cultures from four different days. In this case, we observed three mutations that were confirmed by Sanger sequencing (Figure 2.2C). The first mutation resulted in a change in thymine to a guanine at position 434 in icc. This mutation generated a premature stop codon leading to a truncation at amino acid 144, resulting in a loss of the terminal 130 amino acids (Figure 2.2C). As observed with the

RM series, the sequences from the AR series also showed mixed populations of the parental and the icc mutant with the proportion of the mutant increasing over time

(Figure 2.2B, C). The second mutation was a deletion of one of the repeats in the phase variable region of licA and was observed in all samples analyzed. This change is consistent with phase variation observed for this gene (404). The third mutation was a single nucleotide change that converts an alanine to a threonine at position 217 of rne, 60

Figure 2.2. Independent validation of cyclical pattern of viability and microevolution. (A) Viability was determined every 24 hours for 27 days. Asterisks denote cultures selected for whole genome sequencing in the AR series. (B) The sequence reads that exhibit the SNP are reported as percentage of the total reads for each of the days. (C) Sanger sequencing electropherograms validate the mutation [Thymine (T) → Guanine (G)].

61 which encodes an endoribonuclease and the frequency of mutation parallels that of the icc mutant. Although the mechanism of cycling viability is not understood, the absence of additional mutations within each cycle excludes the contribution of GASP, suggesting additional mechanisms of cyclical growth during stationary phase.

Naturally occurring icc mutants demonstrate reduced cAMP phosphodiesterase activity

NTHI icc is an ortholog of E. coli cpdA, which encodes a class III 3’, 5’-cAMP phosphodiesterase that degrades cyclic AMP (128). Class III phosphodiesterases are characterized by a highly conserved active site with the sequence motif D-(X)n-GC-(X)n-

GNH[E/D]-(X)n-GHXH (127). This motif, which is also present in orthologs from NTHI and Enterobacter aerogenes, binds metal ions at the active site and is critical for the enzymatic function of class III phosphodiesterases (127). The three-dimensional structure of the CpdA ortholog GpdQ from E. aerogenes, which has a perfect copy of the active site motif, has been solved (405). To determine if NTHI Icc contained the conserved active site motif, SWISS-MODEL (https://swissmodel.expasy.org) was employed to construct a three-dimensional (3D) structure using crystal structures of homologs. The generated 3D model of NTHI Icc when aligned with the GpdQ structure reveals a similar overall topology (Figure 2.3A). The NTHI Icc model predicts a conservation of the active site where Asp21 contributes to metal coordination (Figure

2.3D). Mutations of this conserved residue, as observed in RM33, may result in a disruption of enzymatic activity [(127);Figure 2.3B, E]. The introduction of a premature stop codon at residue 145 in Icc in the AR series would result in a truncation of important

Figure 2.3. In silico structural modeling of Icc. (A) Computational model of NTHI Icc (colored cyan) aligned with the known three-dimensional structure of GpdQ [PDB ID:2ZO9, (colored green)] from Enterobacter aerogenes. The two orange spheres indicate the position of iron co-factors in the Icc active site and aspartic acid 21 is indicated in red. (B) The computational model of Icc-RM with the mutation at position 21 from an aspartic acid to a valine shown in stick formation (indicated in orange). (C) The computational model of NTHI Icc-AR (colored cyan) demonstrating the truncated protein was overlaid on the full length Icc (colored white). For reference amino acid 21 in the active site is shown in red. (D) Magnification of the NTHI Icc active site and iron co-factors (orange spheres) coordinated by the aspartic acid at position 21, histidine 23, aspartic acid 63, histidine 204, histidine 202, histidine163 and asparagine 93. (E) Magnification of the Icc-RM active site with the aspartic acid to valine mutation depicted in orange. (F) Magnification of NTHI Icc-AR active site depicts the premature stop codon that disrupts the active site and iron coordination due to the lack of histidine 204, 202 and 163.

active site residues that disrupt the two C-terminal metal-binding domains (Figure 2.3C,

F). Thus, we hypothesize that both naturally occurring mutations in icc would affect cAMP phosphodiesterase activity.

Icc phosphodiesterase activity was quantified using two different experimental approaches. Purified recombinant proteins (Figure 2.4A) were incubated with cAMP to quantify the amount of cAMP degraded by Icc and IccD21V using a modified assay for phosphodiesterase activity (129). As expected, we observed a statistically significant

200-fold decrease in recombinant IccD21V activity as compared to the recombinant parental Icc (Figure 2.4B). Icc is the only cAMP-dependent phosphodiesterase ortholog identified in 86-028NP. Therefore, cytoplasmic extracts from RM33 would be expected to be devoid of any cAMP activity. To this end, cytoplasmic extracts were prepared from exponentially grown cells and the phosphodiesterase activity was quantified as described above. The cytoplasmic extract from RM33 exhibited a statistically significant decrease in phosphodiesterase activity (Figure 2.4C). The mutations in icc that arise during long-term stationary phase culture result in loss of cAMP phosphodiesterase activity and likely increase cAMP levels in the cytoplasm.

RM33 displays enhanced transformation efficiency

Prior studies demonstrated that cAMP levels regulate competence in

Haemophilus with increased cAMP resulting in increased transformation efficiency (148,

150). Our data suggest that the increased cAMP levels in the cytoplasm of RM33 would correlate to an increased rate of uptake of exogenous DNA. Consistent with previous reports, we observed that the addition of exogenous cAMP resulted in a significant increase in the transformation efficiency of the parental strain 86-028NP (Figure 2.4D). 64

Figure 2.4. Biochemical and functional assays confirm reduced cAMP phosphodiesterase activity of mutant Icc. (A) Purified recombinant parental and mutant Icc proteins were visualized on a silver stained SDS polyacrylamide gel. Icc migrates with an apparent molecular weight of 32kDa. Each lane represents purification of the protein from an independent culture. (B) Phosphodiesterase activity (PDE) of purified recombinant proteins was assessed using the PDELight HTS cAMP phosphodiesterase kit and activity reported as relative light units (RLU). Three biological replicates were assayed in technical triplicate and statistical significance was determined using a one–tailed Student’s t test. Error bars indicate standard error of the mean (SEM). (C) Bacterial lysates of the parent (86-028NP) and mutant (RM33) strains were assessed for phosphodiesterase activity as described in panel B. Four biological replicates were assayed in technical triplicate and statistical significance was determined using a one–tailed Student’s t test. Error bars indicate standard error of the mean (SEM). (D) Parent (86-028NP) and mutant (RM33) strains were assessed for transformation efficiency in the presence (+) and absence (-) of cAMP as described in Materials and Methods. The transformation efficiency is reported as the percentage of transformed bacteria from the total number of viable bacteria. Statistical significance was determined using a two–tailed Student’s t test on seven biological replicates. Error bars represent standard error of the mean (SEM). (E) Parent (86-028NPrpsL) and mutant (86-028NPrpsL icc63A>T) strains were assessed for transformation efficiency in the absence (-) of cAMP as described in Materials and Methods. Statistical significance was determined using a two–tailed Student’s t test on seven biological replicates. Error bars represent standard error of the mean (SEM).

As predicted, in the absence of exogenous cAMP, the transformation efficiency of RM33 was indistinguishable from that of the parent in the presence of exogenous cAMP. To validate these findings, we repeated the assay with a strain in which we genetically reconstructed the icc mutation identified in strain RM33. Again, we observed a significant increase in the transformation efficiency of the 86-028NP rpsL icc63A>T strain in the absence of exogenous cAMP (Figure 2.4E). While prior studies demonstrate that transformation is increased as cAMP levels are elevated, saturation of cAMP will suppress transformation (150). As predicted, the addition of exogenous cAMP to RM33 resulted in a decreased transformation efficiency (Figure 2.4D). Collectively, these data suggest that RM33 exhibits elevated endogenous levels of cAMP due to the mutation in icc and the increased ability of RM33 to uptake DNA demonstrates biological significance for the mutation observed in icc.

The absence of effusion is not an indicator of clearance of RM33 from middle ear tissue

Our preliminary observations indicate that transient heme-iron restriction promotes survival during long-term stationary phase in vitro. We hypothesize that strains adapted to nutrient limitation are more adept for colonization and evasion of host immune responses resulting in persistence. Diagnosis of OM is clinically confirmed by the presence of middle ear fluid (effusion). Clearance of effusion is typically associated with resolution of disease. We initially investigated whether RM33 would alter the disease course as measured by the presence of effusion in the preclinical model of OM.

For the first week of infection, we observed no difference in either the presence of effusion or the bacterial burden suggesting that RM33 has similar fitness and kinetics as the parent in the early stages of infection (Figure 2.5A, B). In contrast, 10 days following 66

Figure 2.5. RM33 remains mucosa-associated in the absence of middle ear effusion during experimental OM. (A) Middle ear effusions were retrieved on the days indicated on the x-axis from chinchillas (n=10 ears per cohort) experimentally infected with the parent (dark gray circles) or RM33 (light gray squares). Bacterial burden for each effusion was determined by serial dilution and plating and reported as CFU/mL. The absence of data points represents the inability to retrieve middle ear effusion at that time point. There was no statistically significant difference in the bacterial burden for each time point as determined using Mann-Whitney U test. Bar represents geometric mean, dashed line indicates the limit of detection. (B) Kaplan Meier curve demonstrating time to clearance of middle ear effusions from the parent cohort (solid line) and RM33 cohort (dotted line). The number of ears with recoverable effusion are reported as percent of the total ears. Statistical significance was determined using a chi square test. (C) The weight of middle ear mucosal tissue including biofilm retrieved from animal cohorts infected with parent (dark gray circles) or RM33 (light gray squares). There was no statistically significant difference in mucosal weight as determined using a Mann-Whitney U test. (D) Middle ear mucosal tissues retrieved from animal cohorts infected with parent (dark gray circles) or RM33 (light gray squares) and homogenized to determine bacterial burden reported as CFU/mg tissue. There was no statistically significant difference in bacterial burden as determined using a Mann-Whitney U test. (E) The effusion retrieved was correlated with bacteria recovered from middle ear mucosae. Data are reported as the number of cohort ears infected with parent (dark gray bars) or RM33 (light gray bars) as a percentage of the total number of ears in that cohort. Statistical significance was determined with a two tailed binomial test.

infection we observed a significant decrease in the number of effusions recovered from animals infected with RM33 (Figure 2.5B). This decrease is accentuated 15 days following infection when a single effusion was recovered from animals infected with

RM33 (Figure 2.5B). These data suggest that RM33 does not promote chronic disease as determined by the presence of effusion.

The observation that effusion is not recovered from the majority of animals infected with RM33 suggests the bacteria are cleared from the middle ear. To determine the bacterial burden associated with the mucosal tissue and associated extracellular biofilm, we harvested this tissue upon sacrifice (Day 17). Despite the absence of effusion in the ears infected with RM33 (Day 15), we observed similar tissue biomass to that of ears infected with the parent, suggesting that the disease was not resolved (Figure

2.5C). Homogenates of middle ear mucosae also revealed similar bacterial burdens between the tissues of ears infected with RM33 and the parent (Figure 2.5D).

Clinically, absence of effusion is typically associated with resolution of disease, suggesting the absence of bacteria (406). As expected for a resolved infection, 40% of ears infected with the parent had no recoverable effusion and no bacteria recovered from middle ear mucosal homogenates (Figure 2.5E). In contrast, only 14% of ears infected with RM33 exhibited this phenotype (Figure 2.5E). A recoverable effusion is indicative of an active infection. Effusions could be recovered from 50% of ears infected with the parent, whereas none of the ears infected with RM33 had recoverable effusion.

Taken together, 90% of the ears infected with the parent paralleled clinical presentation and resolution of disease, whereas this was observed in only 14% of RM33 infected ears. Remarkably, 86% of ears infected with RM33 had bacteria associated with the middle ear mucosal tissue, yet no effusion was recovered. These observations suggest

68 that the absence of effusion may not be an accurate clinical indicator of resolution of middle ear disease. Furthermore, unlike the parent, RM33 remains associated with the middle ear tissue following resolution of effusion, suggesting that the microevolution of bacteria in response to nutrient limitation may promote subclinical or latent phenotypes.

NTHI persists in middle ear mucosae as intracellular bacterial communities

Previous studies demonstrated that NTHI are internalized into epithelial cells and then typically traffic to degradative pathways (55, 303). Although intracellular bacteria have been observed in biopsy samples of children with OM (407), the potential contribution of viable intracellular NTHI populations to the chronicity and recurrence of

OM is unclear. Previously, we showed that transient heme-iron restriction promotes

NTHI invasion and formation of IBCs in cultured chinchilla middle ear epithelial cells

(60). Given that RM33 arose from a culture transiently restricted of heme-iron and the association of RM33 with mucosal tissues (Figure 2.5D, E), the potential formation of

IBCs in the preclinical model of OM was investigated. IBCs were visualized in thin sections using antisera directed against NTHI outer membrane proteins (green).

Counterstains to visualize the host cell membranes (red) and DNA (blue) were also included. We observed communities of NTHI within host cell membranes. The individual bacteria within each community were readily identifiable by the surface staining of the outer membrane proteins outlining and surrounding the diffuse bacterial nucleoid (Figure

2.6A). The communities observed filled the entire cell cytoplasm as evidenced by 3D orthogonal views from rendered sequential optical sections (Figure 2.6B).

To determine the kinetics of IBC formation during experimental OM, the number of IBCs within thin sections from ears infected for 7, 14 or 28 days were counted.

Figure 2.6. NTHI persists in middle ear mucosae as intracellular bacterial communities (IBCs). Thin sections of middle ears from parent or RM33 cohorts were processed for immunofluorescence microscopy. (A) Immunofluorescence microscopy gallery depicting the surface staining of NTHI by anti- OMP labelling (green), the host cell membrane by wheat germ agglutinin (red), and host and bacterial DNA by Hoescht (blue). Scale bar = 10 µm. (B) 3-dimensional rendering of a series of optical sections to depict the orthogonal views of IBCs formed by RM33 filling the entirety of the epithelial cell and visualized using the fluorophores depicted in panel A. In this image, six IBCs are observed in the 4-micron section. Scale bar = 10 microns. (C) The presence of IBCs was enumerated in three thin sections of middle ears infected with the parent (dark gray circles) or RM33 (light gray squares). IBCs were enumerated from three sections from at least three independent animals. Each data point represents the number of IBCs in one thin section. Statistical significance was determined using a Mann-Whitney U test. The mean for each data set is indicated and error bars represent standard error of the mean (SEM).

Numerous IBCs were readily observed in middle ears infected with RM33 as early as Day 7. Although IBCs were observed in middle ears infected with the parent on

Day 7, there was a statistically significant increase in the number of IBCs formed by

RM33 (Figure 2.6C). We observed similar numbers of IBCs formed by RM33 throughout the time course evaluated. Interestingly, we observed a significant 6-fold increase in the number of IBCs in the parent over the 28 days evaluated. This observation suggests that

NTHI adaptation during active OM promotes the formation of IBCs.

ICC mutations affect phosphodiesterase activity in paired clinical isolates

To determine whether microevolution of icc occurs during disease, we performed an in silico comparison of 63 NTHI sequences deposited in Genbank. As expected, we observed regions of Icc that exhibit marked variation in sequence compared with the consensus (Figure 2.7A), indicating that the Haemophilus phosphodiesterase family is heterogeneous in overall protein sequence. We examined the diversity in Icc sequence in NTHI clinical isolate sequences obtained in Genbank and determined that Icc plasticity is greater than that observed in Gyrase, a known housekeeping protein (Figure

2.7B). These data suggest increased genetic plasticity of proteins under host selective pressures. In contrast to the mutations observed in vitro that abolished cAMP phosphodiesterase activity, the amino acids involved in enzyme activity are conserved in all sequences examined, suggesting evolutionary conservation of cAMP phosphodiesterase activity is under selective pressures of the host. It is possible that levels of cAMP phosphodiesterase activity are affected by some of the sequence variations observed which would suggest that adaptation during disease requires modulation of cAMP levels. In order to investigate this, we sequenced icc from ten clinical isolate pairs and identified two pairs of strains that contained genetic variations.

Figure 2.7. In silico comparison of genetic plasticity of icc across NTHI clinical isolates. (A) Sequences of icc from NTHI were obtained from published sequences in Genbank and aligned using CLUSTAL W. The consensus sequence was determined for all sequences. Residues identical to the majority consensus are shaded in black. Residues that do not match the majority consensus are indicated in white. The NTHI strains for which the genome is complete are highlighted in red. (B) Sequences of Icc and GyrA from NTHI were obtained from published sequences in Genbank. Phylogenetic trees were generated from the alignments using BIONJ (Gascuel 1997) and distances calculated using the uncorrected pairwise distance method.

One of these clinical isolate pairs, NTHI1521MEE and NTHI1521NP, had changes in five amino acids that significantly altered phosphodiesterase activity (Figure 2.8A, B). These data indicate that mutations acquired outside the conserved active site of the cAMP phosphodiesterase can modulate enzyme activity and subsequent cAMP concentrations.

Further, site specific microenvironments may drive microevolutionary changes in bacterial strains for adaptation at these unique host sites during disease.

2.4 Discussion

The diversity of NTHI lifestyles expands beyond commensalism in the nasopharynx and biofilm formation in the middle ear, to now include intracellular populations of bacteria, a phenotype known to contribute to persistence and recurrence in UPEC, Proteus, Klebsiella and Helicobacter (361). In vitro models have revealed the potential importance of intracellular NTHI, but in these models NTHI typically traffics through the endolysosomal pathway (55, 303, 350). We have previously demonstrated that transient heme-iron restriction promotes IBC formation in cultured middle ear epithelial cells in vitro (60). Here, we now provide evidence that NTHI completely fills the cytoplasm of middle ear epithelial cells during experimental OM. Thus far, IBCs have primarily been attributed to pathogens of the urinary tract, but our studies now reveal

IBCs formed by a pathogen of the respiratory tract. Thus, our studies expand upon previous in vitro models and use a preclinical model of OM to provide mechanistic insight into an understudied intracellular niche and the potential contribution to persistence and recurrence.

IBCs formed by UPEC and other species confer tolerance to antibiotics, as well as protection from innate and adaptive immune responses (358, 361). Further, the

Figure 2.8. Phosphodiesterase activity of paired NTHI clinical isolates. icc was sequenced from ten pairs of clinical isolates of NTHI. (A) One set of paired isolates, NTHI1521MEE and NTHI1521NP, contained genetic variation in icc that resulted in changes to five amino acids (highlighted in red). (B) Phosphodiesterase activity of paired clinical isolates NTHI1521MEE and NTHI1521NP was determined using the PDELight HTS cAMP phosphodiesterase kit and activity reported as relative light units (RLU). Three biological replicates were assayed in technical triplicate and statistical significance was determined using a one–tailed Student’s t test. Error bars indicate standard error of the mean (SEM).

74 intracellular environment provides access to nutrients that are limited in the urine (362,

363, 408). IBCs are critical for persistence and recurrence for urinary tract infection

(361). These studies have shifted the clinical perspective from a focus on recurrence due to reinoculation to include recurrences from an intracellular bladder reservoir. The implications of these findings have already impacted approaches to treat and prevent recurrent infection (409, 410). Based upon the UPEC paradigm that IBCs lead to an intrabladder reservoir for reinfection, it is of interest to speculate that IBCs in the middle ear provide a viable population of bacteria to initiate recurrent OM in the absence of reinoculation.

NTHI was recovered from middle ear aspirates from 57% of children with persistent OM at all visits during a longitudinal study (411). Furthermore, the same strain of NTHI was isolated from the same COPD patient in a longitudinal study, despite periods of negative culture (54). As such, the observed persistence of NTHI at multiple niches requires new approaches to prevent and treat infection. Persistent bacterial populations can arise stochastically, but also as a result of stressors that include, among others, nutrient limitation (412-416). The use of heme-iron restriction as a nutritional stressor dramatically increases the longevity of NTHI in stationary phase cultures

(Figures 2.1, 2.2). In a preclinical model of OM, the strain adapted though transient heme-iron restriction (RM33) exhibited increased IBC formation as compared with the parent, suggesting that adaptation to fluctuations in nutrient availability promotes an intracellular lifestyle.

Our discovery of IBCs in vivo combined with the ability of NTHI to form IBCs in cultured epithelial cells allows investigation of differential trafficking of bacteria either to the lysosome or for productive IBC formation. In our study we focused on nutrient

75 limitation, but the complexity of the host environment suggests that other immune stressors may contribute to NTHI microevolution during disease. Additional investigations delineated mechanisms of invasion that promote IBC formation, and these data are described in Chapter 3.

In two independent cultures, we observed microevolution of icc which encodes the only cAMP phosphodiesterase in Haemophilus. Icc degrades cAMP to modulate the intracellular concentration of this important signaling molecule. DNA import is regulated by cytoplasmic cAMP levels (150) and we and others have shown that defects in cAMP phosphodiesterase increases competence [Figure 2.4D; (150)]. Bacteria can utilize exogenous DNA for horizontal gene transfer, DNA repair, or as a nutritional source (417,

418). Use as a nutritional source is a compelling concept and is supported by the observations that DNA is readily available in host niches and only 10-15% of the DNA taken up by NTHI is incorporated intact into the genome, the remainder is degraded

(209, 419-421). The link between nutrition, cAMP levels and competence is supported by the requirement for nutrient limitation as an inducer of competence in Haemophilus

(422). Moreover, nutrient availability signals catabolite repression to control levels of cAMP through the enzymatic activities of adenylate cyclase and cAMP phosphodiesterase combined with the cAMP-binding protein CRP (144, 423). In addition to the canonical CRP binding site in E. coli (CRP-N), Haemophilus has an additional regulatory site (CRP-S) that uses Sxy for the regulation of competence (143, 146, 424,

425). Thus, there is overlap between CRP-N and CRP-S directed regulation and a close link between sensing nutritional status and regulation of competence in H. influenzae

(145). Although we used competence as a tool to determine changes in the levels of cAMP in the adapted strain (RM33), the contribution of competence or any other

76 member of the CRP regulon in the phenotypes observed during infection remain to be elucidated. Icc is likely not the only target of adaptation in vivo. During disease, fluctuation in nutrients as well as immune stressors may influence adaptation through direct modulation of cAMP levels, indirectly through mutation or regulation in other members of the cAMP regulon, or other as yet unidentified processes.

These studies provide a new model to delineate the role of cAMP in the development of IBCs. In our study, we observe a decrease in cAMP phosphodiesterase activity following nutrient restriction with an increase in persistence, consistent with the role of cAMP and catabolite repression in nutrient limited conditions. In contrast, other studies provide evidence that a decrease in cAMP levels is associated with increased antibiotic-mediated persister cell formation (426, 427). Thus, the levels of cAMP appear to modulate persistence through multiple mechanisms based upon the type of stressor encountered.

NTHI was not recovered from the mucosal tissues of some ears (Figure 2.5D), yet we observed IBCs in all sections evaluated (Figure 2.6C), suggesting that NTHI within IBCs may not be liberated using standard homogenization techniques or that this population is viable but not culturable following intracellular growth. It should be noted that the IBCs counted represent the number present in a 4µm thin section, which suggests that our quantitation will underestimate the magnitude of intracellular NTHI. In fact, recent studies have demonstrated that enumeration of viable bacteria does not accurately represent the number of bacteria within IBCs (428). Recent studies have also determined that UPEC IBCs are metabolically active in the bladder (360). Future studies will evaluate the metabolic activity of NTHI in IBCs in the middle ear.

Collectively, our observations reveal microevolution in response to nutrient limitation leading to persistence in long-term culture, and that this adapted strain exhibited a predilection for productive growth within the epithelial environment which resulted in increased IBC formation during experimental OM. Perhaps even more interesting, we observed an increase in IBC formation of the parent at 28 days during experimental OM, suggesting that genotypic or phenotypic adaptation occurs during infection. As the availability of nutrients change due to the dynamic immune response, we speculate that the intracellular environment provides nutrients leading to bacterial proliferation and IBC formation. Moreover, the presence of IBCs during OM provides the first opportunity to evaluate the role of intracellular populations in chronicity and quiescence as a new paradigm for recurrent OM. This model provides a new platform for the identification of novel diagnostic and therapeutic approaches for OM. Our prior proteomic studies demonstrate an increase in the actin filament branching protein,

Arp2/3, during acute experimental OM and that inhibition of Arp2/3 decreases NTHI invasion as a potential adjunct therapy (298). Finally, it is of interest to determine whether intracellular populations contribute to the other NTHI-mediated diseases such as conjunctivitis, sinusitis, pneumonia, and exacerbations in COPD and cystic fibrosis

(429, 430).

Chapter 3. Transient nutrient deprivation promotes macropinocytosis-dependent

intracellular bacterial community development by Haemophilus†.

Abstract

Nutrient limitation restricts bacterial growth in privileged sites such as the middle ear.

Transient heme-iron restriction of nontypeable Haemophilus influenzae (NTHI), the major causative agent of chronic and recurrent otitis media (OM), promotes new and diverse phenotypes that can influence planktonic, biofilm, and intracellular lifestyles of

NTHI. However, the bacterial responses to nutrient restriction that impact intracellular fate and survival of NTHI are unknown. In this work, we provide evidence for the role of transient heme-iron restriction in promoting the formation of intracellular bacterial communities (IBCs) of NTHI both in vitro and in vivo in a preclinical model of OM. We show that transient heme-iron restriction of NTHI results in significantly increased invasion and intracellular populations that escape or evade the endolysosomal pathway for increased intracellular survival. In contrast, NTHI continuously exposed to heme-iron traffics through the endolysosomal pathway for degradation. The use of pharmacological inhibitors revealed that prior heme-iron status does not appear to influence NTHI internalization through endocytic pathways. However, inhibition of macropinocytosis altered the intracellular fate of transiently restricted NTHI for degradation in the

† A version of this chapter has been published as follows: Hardison RL, Heimlich DR, Harrison A, Beatty WL, Rains S, Moseley MA, Thompson JW, Justice SS, and Mason KM. 2018. Transient nutrient deprivation promotes macropinocytosis-dependent intracellular bacterial community development. mSphere 3:e00286-18. 79 endolysosomal pathway. Furthermore, prevention of macropinocytosis significantly reduced the number of IBCs in cultured middle ear epithelial cells providing evidence for the feasibility of this approach to reduce OM persistence. These results reveal that microenvironmental cues can influence the intracellular fate of NTHI leading to new mechanisms for survival during disease progression.

Importance

Otitis media is the most common bacterial infection in childhood. Current therapies are limited in the prevention of chronic or recurrent otitis media which leads to increased antibiotic exposure and represents a significant socioeconomic burden. In this study, we delineate the effect of nutritional limitation on the intracellular trafficking pathways used by nontypeable Haemophilus influenzae (NTHI). Moreover, transient limitation of heme-iron led to the development of intracellular bacterial communities that are known to contribute to persistence and recurrence in other diseases. New approaches for therapeutic interventions were revealed through the use of pharmacological inhibition of macropinocytosis that reduce the production of intracellular bacterial communities and promote trafficking through the endolysosomal pathway. This work demonstrates the importance of an intracellular niche for NTHI and provides new approaches for intervention for acute, chronic and recurring episodes of otitis media.

3.1 Introduction

Bacterial pathogens exploit diverse host niches to survive and cause disease.

Opportunistic pathogens, by definition, are especially adaptable as the host microenvironment can shift rapidly during the transition from a commensal to a pathogen. Physiological stressors occur as a result of innate and adaptive host immune 80 defense mechanisms and can drive phenotypic changes that promote diverse bacterial lifestyles. Understanding the mechanisms underlying responses to microenvironmental cues that influence bacterial survival in both extracellular (planktonic and biofilm) and intracellular (vacuolar and cytosolic) environments is critical to ultimately prevent chronic and/or recurrent diseases.

As an opportunistic pathogen, nontypeable Haemophilus influenzae (NTHI) is a major cause of otitis media (OM), exacerbations of chronic obstructive pulmonary disease, and sinusitis, among others. Many NTHI infections can return despite prior antibiotic treatment, and NTHI is the most commonly isolated microbe from recurrent episodes of OM (47, 54, 411, 431-434).

The survival and pathogenicity of NTHI, a heme-iron auxotroph, depends on the ability to acquire iron from the external environment. NTHI resides as a commensal in the nasopharynx but can migrate into the sterile middle ear under permissive conditions such as an upper respiratory viral infection or Eustachian tube dysfunction. At the onset of OM, the host tightly sequesters heme-iron and other essential nutrients through nutritional immunity. Over the course of infection, inflammation and immune responses cause fluctuations in iron availability. NTHI has multiple mechanisms for iron acquisition and responds to these changes by upregulating core iron and heme responsive genes during experimental OM (384). Additionally, heme-acquisition genes were more prevalent in NTHI strains isolated from the middle ears of children with OM compared to strains isolated from the throats of healthy children (383). These studies underscore the importance for NTHI heme-iron acquisition in development of disease.

Iron limitation can induce phenotypic changes that promote virulence and survival in bacteria, including stimulation of biofilm formation in S. aureus and increased adherence of P. aeruginosa to epithelial cells (435, 436). We have recently 81 demonstrated that transient heme-iron restriction of NTHI alters biofilm architecture and morphology, increases survival in a preclinical model of OM, and promotes invasion and intracellular bacterial community (IBC) formation within chinchilla middle ear epithelial cells (59, 60, 116).

While historically considered an extracellular pathogen, intracellular NTHI are observed within respiratory and middle ear epithelia (47, 55, 299, 303, 437). NTHI is internalized by host cells through cell type specific mechanisms that include clathrin- mediated endocytosis, actin remodeling, microtubules, and lipid rafts (55, 171, 295, 300,

304, 305, 312, 350). Further, multiple studies have demonstrated that NTHI may enter cells through macropinocytosis (171, 313). Once internalized, NTHI typically traffic via the endolysosomal pathway resulting in degradation within lysosomes (55, 303).

However, the contribution of intracellular NTHI to the disease progression of recurrent and chronic OM, and importantly, how nutritional immunity and resulting heme-iron restriction may influence the intracellular fate of NTHI remains underexplored.

Our prior studies have revealed a potential viable intracellular population of NTHI in response to heme-iron limitation (59, 60). Based upon these observations, we hypothesized that prior heme-iron restriction of NTHI alters intracellular bacterial trafficking and results in increased bacterial survival within host cells. Herein, we report that heme-iron restriction of NTHI leads to productive invasion into epithelial cells in vitro and in vivo that promotes IBC formation in chinchilla middle ear epithelium during experimental OM. In contrast to NTHI that were continuously exposed to heme-iron, we observed populations of transiently restricted NTHI that did not co-localize with markers of the early and late endolysosomal pathway, suggesting evasion or escape from early endocytic vacuoles. Further, our data demonstrate that entry via macropinocytosis

82 contributes to intracellular survival and evasion of endolysosomes by transiently restricted NTHI. Together, these data reveal a novel role for the response of NTHI to changes in heme-iron availability in promoting an intracellular lifestyle, and advance our understanding of how NTHI may survive within the host middle ear epithelium to cause acute and recurrent OM.

3.2 Materials and Methods

Bacterial strains, cell lines, and media

NTHI strain 86-028NP is a minimally passaged clinical isolate which has been sequenced and characterized in the chinchilla model of OM (60, 103). A green fluorescent protein (GFP) reporter strain of 86-028NP was generated by electroporation of plasmid pGM1.1 as previously described (60, 116). For routine culture, NTHI was grown on chocolate agar plates (Fisher Scientific, Pittsburgh PA). For routine liquid culture, NTHI was grown in an iron-depleted defined iron source (DIS) medium supplemented with β-NAD (112). Where indicated, DIS was supplemented with 2 µg/mL heme (Millipore Sigma, St. Louis, MO). NTHI was transiently restricted or continuously exposed to heme-iron in 24-hour liquid culture as previously described (60). Briefly,

◦ strain 86-028NP or 86-028NP(pGM1.1) were grown overnight at 37 C in a 5% CO2 atmosphere on chocolate agar plates. Cells were resuspended in DIS, adjusted to an

OD490 of 0.65 and diluted 1:10 into pre-warmed DIS medium containing either 0 or 2

◦ µg/mL heme. Cultures were grown for 24 hours statically at 37 C, 5% CO2 and normalized to an OD490 of 0.37 in DIS containing 2 µg/mL heme to generate parallel transiently restricted and continuously exposed cultures, respectively.

Normal human bronchial epithelial (NHBE; Lonza, Allendale, NJ) cells were cultured with bronchial epithelial basal media (Lonza) supplemented according to manufacturer’s specifications in cell culture-treated flasks or plates (Corning, Corning,

◦ NY) at 37 C with 90% humidity and 5% CO2. Primary chinchilla middle ear epithelial

(CMEE) cells were isolated from adult chinchilla middle ear mucosa and cultured in

◦ CMEE growth medium at 37 C with 90% humidity and 5% CO2 as previously described

(60, 438).

Ethics statement

All animal experiments were carried out in strict accordance with the accredited conditions in the Guide for the Care and Use of Laboratory Animals of the National

Institutes of Health. The protocol was approved by the Institutional Animal Care and Use

Committee at The Research Institute at Nationwide Children’s Hospital. All experimental procedures were performed under xylazine and ketamine anesthesia, and all efforts were made to minimize suffering.

Chinchilla model of otitis media

Healthy adult chinchillas (Chinchilla lanigera) were obtained from Rauscher’s chinchilla ranch (LaRue, OH). To assess formation of intracellular bacterial communities

(IBCs) within middle ear epithelial cells, three cohorts of five chinchillas were transbullarly challenged with 1077 CFU of transiently restricted or 1542 CFU of continuously exposed NTHI in a total volume of 300 µL saline per ear. Two days following middle ear challenge, the animals were sacrificed and the middle ear inferior bullae removed and fixed in buffered 4% paraformaldehyde (PFA) in DPBS for 24 hours.

Fixed bullae were decalcified in 0.35M ethylenediaminetetraacetic acid and 0.1 M Tris

(pH 6.95) and embedded in paraffin. For immunohistochemistry, 4 µm sections were deparaffinized in xylene followed by antigen retrieval as previously described (298).

Immunohistochemistry and fluorescent labeling of IBCs in vivo

To assess the formation of IBCs in the chinchilla middle ear, immunohistochemistry was performed on middle ear sections. Non-specific antibody binding was reduced by incubation for 10 minutes with 0.01% sodium borohydride and the inherent fluorescence of the sample was quenched by incubation for 10 minutes with

CAS-Block Histochemical Reagent (Thermo Fisher Scientific, Waltham, MA). The specimens were then incubated for 30 minutes in Image-iT FX Signal Enhancer (Thermo

Fisher Scientific) to further prevent nonspecific association of dyes with the tissue. NTHI in middle ear sections were detected with a chinchilla antiserum to NTHI outer membrane proteins (OMP) diluted 1:25 in CAS block buffer and incubated overnight at

4˚C, as previously described (59). The NTHI-antibody complexes were visualized using

Protein A conjugated with Alexa Fluor 488 (Thermo Fisher Scientific) diluted 1:100 in

DPBS for 1 hour at room temperature. Host cell membranes were visualized using wheat germ agglutinin (WGA) conjugated with Alexa Fluor 594 (final concentration 5

µg/mL, Thermo Fisher Scientific) and DNA was visualized with Hoescht 33342 (Thermo

Fisher Scientific). Coverslips were mounted on slides with ProLong Gold Antifade

(Thermo Fisher Scientific) and imaged with an Axiovert 200M inverted epifluorescence microscope equipped with the Apotome attachment for improved fluorescence resolution and an Axiocam MRM CCD camera and Axiovision software (Carl Zeiss, Inc.,

Thornwood, NY). Images were processed in Photoshop (Adobe, San Jose, CA) using the levels function applied to every pixel within the image to enhance visualization of fluorescent structures. For rendered images, 0.6 µm optical sections were acquired at 85

0.6 µm intervals, all images were stacked and 3 dimensional interpolation was performed in Image J (NIH). Orthogonal images to visualize the depth of the intracellular communities was performed in Image J (NIH) and a single optical section is depicted in the panel.

Co-culture of nutritionally conditioned NTHI with epithelial cells in vitro

NHBEs were seeded on glass coverslips in 24 well cell culture dishes (Corning) at 1.9 x 104 cells per well or in 8-well Nunc chamber slides (Thermo Fisher Scientific) at

8000 cells per well and grown to 90% confluency. CMEE cells were cultured as previously described (60). Transiently restricted or continuously exposed 86-

028NP(pGM1.1) were co-cultured with epithelial monolayers at a multiplicity of infection

(MOI) of 25 bacteria per cell. At 1-hour post-infection of NHBEs, cells were washed once with DPBS prior to the addition of fresh bronchial epithelial growth medium (BEGM) and subsequently incubated for 4 or 24 hours depending upon the experiment. For CMEEs, at 5 hours post infection, cells were washed once with DPBS prior to the addition of fresh CMEE growth medium (60) containing 100 µg/ml gentamicin for 30 minutes.

CMEEs were washed twice with DPBS prior to the addition of fresh medium for a total incubation period of 24 hours. Following incubation, monolayers were washed twice with

DPBS and fixed on ice for 1 hour with 4% PFA. After fixation monolayers were washed twice with DPBS and host cell membranes were visualized with WGA conjugated to either Alexa Fluor 594 or 350 (final concentration 5 µg/mL) dependent upon other fluorophores used in each experiment. Coverslips were mounted on slides using

ProLong Gold Antifade reagent. Images were acquired and processed as described above.

Gentamicin protection assays

Quantitative gentamicin protection assays were adapted for use with NTHI from an established protocol (439). NHBE cells were seeded at 1.9 x 104 cells per well and grown to 90% confluency in 24 well cell-culture plates (Corning). Cells were cultured with transiently restricted or continuously exposed NTHI at a multiplicity of infection (MOI) of

25 per cell for 1 hour. To determine the total number of bacteria associated with the cells

(intracellular and extracellular), monolayers were lysed with 0.05% saponin for 5 minutes at room temperature, scraped for 10 seconds with mini cell scrapers, and collected in microcentrifuge tubes. The concentration of saponin and duration of cell scraping needed to sufficiently lyse the cells was determined empirically and confirmed by microscopy. The concentration of saponin did not affect the viability of NTHI. To determine the total number of intracellular bacteria, NHBE monolayers were treated with

100 µg/mL gentamicin in BEGM for 30 minutes to eradicate extracellular bacteria. The concentration and duration of gentamicin treatment sufficient to kill extracellular bacteria was determined empirically. Following gentamicin treatment, monolayers were lysed with 0.05% saponin, scraped with mini cell scrapers for 10 seconds per well, and samples were collected in microcentrifuge tubes. All samples were serially diluted and plated to enumerate viable CFU. The percent of viable bacteria remaining from the original inoculum following gentamicin treatment was calculated by dividing the number of viable bacteria recovered from each well by the number of viable bacteria originally inoculated into each well and multiplying by 100. An invasion index was determined by dividing the percent of viable intracellular bacteria (gentamicin treated) by the percent of total intracellular and extracellular bacteria (no gentamicin present). Duplicate wells were averaged for each condition and time point, and the experiment was performed in

87 biological triplicate. Statistical significance was determined by Student’s t-test, p<0.05

(Graphpad Prism, La Jolla, CA).

Intracellular survival assays

To determine the ability of the transiently restricted or continuously exposed

NTHI populations to survive within epithelial cells in vitro, an intracellular survival assay was performed as follows. Cells were co-cultured with transiently restricted or continuously exposed NTHI at an MOI of 25 per cell for 90 minutes, washed twice with

DPBS, and treated with 100 µg/mL gentamicin for 30 minutes at 37˚C, 5% CO2 to eradicate extracellular bacteria. Monolayers were washed twice with DPBS and subjected to a second incubation period in antibiotic-free medium for the time indicated in each experiment. Monolayers were treated a second time with 100 µg/mL gentamicin for 30 minutes at 37˚C, 5% CO2, washed twice with DPBS, lysed with 0.05% saponin for

5 minutes at room temperature and scraped for 10 seconds with mini cell scrapers.

Samples were collected in microcentrifuge tubes and serially diluted and plated on chocolate agar to enumerate viable intracellular bacteria. The percent of viable intracellular bacteria was calculated relative to the inoculum. Duplicate wells were averaged for each condition and time point. Eight biological replicates were performed for the 2-hour time point and five biological replicates were performed for the 24-hour time point. Significance at each time point was determined by Student’s t-test, p<0.05

(Graphpad Prism).

Pharmacological inhibition of endocytic pathways

To determine uptake mechanisms of continuously exposed and transiently restricted NTHI into host epithelial cells, pharmacological compounds were used to

88 systematically inhibit endocytic pathways. These compounds are listed in Table 3.1.

Noncytotoxic concentrations of each inhibitor and vehicle were established through 3-

(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assays of uninfected cells using a cytotoxicity detection kit (Promega, Madison, WI). The following pharmacological compounds were used to inhibit endocytosis: 5-(N-ethyl-N-isopropyl)- amiloride (EIPA; Cayman Chemical, Ann Arbor, MI) was reconstituted in dimethyl sulfoxide (DMSO) to generate a 10 mg/mL working stock and used at a working concentration of 60 µM in cell culture medium to inhibit macropinocytosis; chlorpromazine hydrochloride (Sigma-Aldrich Corp., St. Louis, MO) was resuspended in cell culture medium at a working concentration of 30 µM to inhibit clathrin-mediated endocytosis; cytochalasin D (CytoD) from Zygosproium mansonii (Sigma-Aldrich) was reconstituted in DMSO to generate a 1 mg/mL stock solution and used at a working concentration of 10 µM in cell medium to inhibit F actin polymerization; methyl-β- cyclodextrin (MβCD; Sigma-Aldrich) was used at working concentrations of 5 mM and 1 mM in cell medium to extract cholesterol and reduce lipid rafts, made fresh in cell culture medium for same-day use. For gentamicin protection assays and fluorescence microscopy, inhibitor concentrations remained constant for pretreatment and infection period, except for MβCD which was used to pretreat cells at a concentration of 5 mM and then reduced to 1 mM during the infection period to prevent cytotoxicity. Cells were

◦ pretreated with EIPA or MβCD for 2 hours at 37 C, 5% CO2, while pretreatment with

⁰ cytochalasin D or chlorpromazine was for 1 hour at 37 C, 5% CO2. These inhibitor concentrations and treatment times were determined empirically using fluorescent control cargo conjugates as described below.

Positive Compound Abbreviation Target Mechanism of Action Control Inhibits actin Cytochalasin D CytoD F-actin Phalloidin polymerization Clathrin/Receptor- Dynamin and/or AP-2 Chlorpromazine CPZ Mediated recruitment to clathrin Transferrin Endocytosis coated vesicles Methyl-β- Extraction of cholesterol Cholera Toxin MβCD Lipid rafts cyclodextrin from cell membrane subunit B 5-(N-ethyl-N- Inhibits Na+/H+ 70,000 MW isopropyl) EIPA Macropinocytosis exchangers Dextran amiloride

Table 3.1 Pharmacological compounds used to inhibit endocytosis in this study.

Fluorescence microscopy of control cargo conjugates upon pharmacological inhibition

Experimental conditions for pharmacological inhibition were determined through optimal inhibition of the following fluorescent control cargo conjugates: cholera toxin subunit B Alexa Fluor 488 conjugate (Thermo Fisher Scientific; 2 µg/mL final concentration), transferrin Alexa Fluor 488 conjugate (Thermo Fisher Scientific; 50

µg/mL final concentration), and 70000 MW Dextran Alexa Fluor 488 conjugate (Thermo

Fisher Scientific; final concentration 50 µg/mL). NHBE cells were grown to 90% confluency on coverslips and were pretreated with cell medium with or without inhibitor followed by the addition of cargo conjugates and incubation for 1 hour (transferrin, dextran) at 37⁰C or 20 minutes on ice followed by 1 hour at 37⁰C (cholera toxin subunit

B). Cells were washed twice with DPBS and fixed with 4% PFA overnight at 4⁰C. PFA was removed and cells were washed twice with DPBS before coverslips were mounted on slides using ProLong Gold Antifade. Cells were imaged and processed as described above. Cytochalasin D reduces cellular F-actin polymerization (440), and pretreatment time and activity of cytochalasin D was monitored by selectively staining for F-actin

⁰ polymers. Cells were pretreated with the inhibitor for 1 hour at 37 C, 5% CO2, washed twice with DPBS, and fixed in 4% PFA overnight at 4⁰C. Following fixation, cells were washed twice with DPBS and permeabilized with 0.1% Triton X-100 in DPBS for 10 minutes. F-actin was then stained using Alexa Fluor 350 phalloidin according to manufacturer protocol (Thermo Fisher Scientific). Cells were pre-incubated with Image- iT FX Signal Enhancer (Thermo Fisher Scientific) for 30 minutes. A 6.6 µM stock solution of Alexa Fluor 350 was prepared by reconstituting in methanol. A working solution was prepared by diluting 5 µL of the methanolic stock solution into 200 µL of DPBS per

91 coverslip to be stained. Cells were incubated with Alexa Fluor 350 phalloidin for 20 minutes at room temperature and then washed twice with DPBS and visualized on an

Axiovert 200M inverted epifluorescent microscope.

Fluorescence microscopy following infection of cells treated with pharmacological inhibitors

To determine the effect of pharmacological inhibition on internalization of nutritionally conditioned NTHI, NHBEs were seeded at 1.9 x 104 cells per well and grown to 90% confluency on glass coverslips in 24 well plates. Monolayers were pretreated with each inhibitor before and during incubation with transiently restricted or continuously exposed NTHI strain 86-028NP(pGM1.1) at an MOI of 25 per cell for the time points described in each experiment. Following infection, cells were washed twice with DPBS and fixed with 4% PFA overnight at 4˚C. Cells were then washed twice with DPBS and incubated with WGA conjugated to Alexa Fluor 594 (final concentration of 5 µg/mL in

DPBS) for 10 minutes at room temperature. Coverslips were washed twice with DPBS and mounted on glass microscope slides using ProLong Gold Antifade reagent. Images were acquired and processed as described above. Experiments were performed in biological triplicate and representative images are shown.

Gentamicin protection and intracellular survival assays with pharmacological inhibitors

The gentamicin protection and intracellular survival assay protocols were modified to include a period of pretreatment with cell culture medium containing the inhibitor, as indicated, followed by co-culture with transiently restricted or continuously exposed NTHI as described above. 92

Immunolabeling of endolysosomal markers

To examine co-localization of NTHI with endolysosomal markers, NHBE monolayers were seeded at 8000 cells per well and grown to 90% confluency in 8 well chamber slides. Cells were infected for a total of 4 or 24 hours with transiently restricted or continuously exposed NTHI strain 86-028NP(pGM1.1) at an MOI of 25 per cell.

Monolayers were washed twice with DPBS and fixed overnight at 4◦C in 4% PFA. Cells were then washed twice with DPBS, permeabilized in 1% Triton X-100 in DPBS for 10 minutes, and blocked with CAS-Block for 30 minutes. Host cell membranes were visualized using WGA conjugated to either Alexa Fluor 350 or 594 (final concentration of

5 µg/mL), and early endosomes and lysosomes were visualized by incubation with rabbit anti-early endosomal antigen 1 (EEA1; 1:1000; Abcam, Eugene, OR) or rat anti-LAMP1

(1:500; Developmental Hybridoma Studies Bank, Iowa City, IA) for 1 hour at room temperature, respectively. Cells were washed twice with DPBS and incubated with donkey anti-rabbit IgG conjugated to Alexa Fluor 594 (1:500; Thermo Fisher Scientific) for EEA1 or donkey anti-rat Alexa Fluor 594 (1:500, Thermo Fisher Scientific) for LAMP1 for 1 hour at room temperature. Following immunolabeling, cells were washed twice with

DPBS and coverslips were mounted on slides using Slowfade Diamond Antifade

Mounting reagent. All images were acquired and processed as described above.

Experiments were performed in triplicate and representative images are shown.

Quantification of the number of colocalization events per cell was performed by visually counting 200 total individual cells from three independent experiments for each condition. Statistical significance was determined by a two tailed Mann-Whitney U-test

(Graphpad prism).

Quantification of LAMP1 staining patterns

NHBE cells were co-cultured with either transiently restricted or continuously exposed NTHI strain 86-028NP(pGM1.1) for 5 hours and immunolabeled with rabbit anti-

LAMP1 (Abcam, 1:500) and donkey anti-rabbit IgG conjugated to Alexa Fluor 594

(1:500, Thermo Fisher Scientific) to visualize LAMP1. Immunofluorescence staining patterns of LAMP1 were visually categorized as diffuse, punctate, or a combination.

Quantification of LAMP1 staining patterns was completed by visual counting of ~100 individual cells by two independent investigators.

Transmission electron microscopy

NHBEs were seeded and grown to 80% confluency in 8-well chamber slides and inoculated with nutritionally conditioned NTHI at MOI of 25 per cell. At 90 minutes post- infection, monolayers were washed with DPBS, cell medium was replaced, and infection allowed to progress for a total of 4 or 24 hours as indicated. Monolayers were washed

◦ with DPBS, incubated with trypsin (Corning) for 5 minutes at 37 C, 5% C02, and trypsin was inactivated with Trypsin Neutralizing Solution (TNS; Lonza). Samples were collected in 15 mL conical tubes and pelleted by centrifugation at 500 x g for 10 minutes. Cell pellets were washed by resuspending in DPBS, followed by centrifugation at 1000 x g for

10 minutes. Pellets were then fixed in 4% PFA /0.05% glutaraldehyde in DPBS for one hour on ice. Pellets were stored in cold PBS at 4˚C. For immunolocalization of EEA1 and

LAMP1, samples were embedded in 10% gelatin and infiltrated overnight with 2.3M sucrose/20% polyvinyl pyrrolidone at 4˚C. Samples were trimmed, frozen in liquid nitrogen, and sectioned with a Leica Ultracut UCT7 cryo-ultramicrotome (Leica

Microsystems Inc., Bannockburn, IL). Fifty nanometer sections were blocked with 5%

94 fetal bovine serum/5% normal goat serum for 30 minutes and subsequently incubated with rabbit anti-EEA1 antibody (1:500, Abcam) or rabbit anti-LAMP1 (1:200, Abcam) for

1 hour followed by secondary anti-rabbit IgG antibody conjugated to 18 nm colloidal gold

(1:30, Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) for 1 hour.

Sections were stained with 0.3% uranyl acetate/2% methyl cellulose and viewed on a

JEOL 1200 EX transmission electron microscope (JEOL USA Inc., Peabody, MA) equipped with an AMT 8-megapixel digital camera and AMT Image Capture Engine

V602 software (Advanced Microscopy Techniques, Woburn, MA). All labeling experiments were conducted in parallel with controls omitting the primary antibody.

These controls were consistently negative at the concentration of colloidal gold conjugated secondary antibodies used in these studies.

Visual localization of intra- vs extracellular NTHI

To confirm the subcellular localization of NTHI, NHBE monolayers were co-cultured with continuously exposed or transiently restricted NTHI strain 86-028NP in an 8-well chamber slide (as described in the Materials and Methods) for 4 hours and fixed with 4%

PFA. Cells were then either permeabilized with 0.1% Triton X-100 for 10 minutes at room temperature or non-permeabilized. NTHI were labeled with anti-OMP antiserum diluted 1:25 in CAS-Block for 3 hours at room temperature. Cells were washed twice with DPBS and NTHI was detected with protein A-Alexa Fluor 488 diluted 1:500 in DPBS for 1 hour at room temperature. Cells were then washed twice with DPBS and coverslips were mounted on slides using ProLong Gold with DAPI (Thermo Fisher Scientific) to visualize host cell and bacterial nuclei, pseudo colored white. In permeabilized cells, both intracellular and extracellular NTHI are observed as green/white. In non- permeabilized cells, intracellular NTHI are observed as white and extracellular NTHI are 95 observed as green. All images were taken with an Axiovert 200M inverted epifluorescence microscope. Experiments were performed in triplicate and representative images are shown.

Cytotoxicity Assay

To determine if nutritionally conditioned NTHI was cytotoxic to mammalian cells at the time points used in this study, NHBE cells were seeded into 96 well cell culture plates at 1.9e4 cells per well and allowed to grow to 80-90% confluency. Transiently restricted or continuously exposed NTHI strain 86-028NP were co-cultured with the confluent NHBEs at an MOI of 25 per cell for 4 or 24 hours. Cytotoxic effect of nutritionally conditioned NTHI was determined by measuring lactate dehydrogenase

(LDH) release using the CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega) according to the manufacturer’s protocol.

Proteomic analysis of transiently restricted and continuously exposed NTHI

Sample preparation and protein isolation: NTHI strain 86-028NP(pGM1.1) was continuously exposed to, or restricted of heme-iron for 24 hours as described in Bacterial

Strains, Cell Lines, and Media. Cultures were normalized to an OD490 of 0.37 and 5 mL of each culture was centrifuged at 4000 x g for 15 minutes at 4˚C. Pellets were resuspended in 1 mL 50mM ammonium bicarbonate buffer (pH 8) and lysed by high pressure cell (20,000 psi; One Shot Model, Constant Systems LTD, Kennesaw, GA).

Samples were centrifuged at 20,000 x g at 4˚C for 15 minutes and the supernatants were stored at -80˚C prior to analyses. After thawing, the samples were vortexed and centrifuged at 15,000 x g at 4˚C. The soluble fractions were lyophilized overnight to dryness and resuspended in 50 µL of 0.2% acid labile surfactant (ALS-1). Protein

96 concentrations of the supernatants were determined by mini-Bradford assay (Bio-Rad

Laboratories, Inc., Hercules, CA). Samples with concentrations ≥ 0.3 µg/µl (n=3 per condition) were chosen for further sample processing and proteomic analysis. 10 µg of protein per sample was normalized to 0.2 µg/µL in 0.2% ALS-1 then denatured at 40˚C for 10 minutes. Samples were then reduced with dithiothreitol at a final concentration of

10 mM for 15 minutes at 80˚C, followed by alkylation with 20 mM final concentration iodoacetamide at room temperature for 30 minutes. Samples were cooled to room temperature, then digested with 1 µg sequencing grade trypsin (Promega) overnight at

37˚C. Digestion was quenched and acid-labile detergent cleaved by addition of 7 µL of

10:20:70 v/v/v trifluoroacetic acid/acetonitrile/water and incubation at 60˚C for two hours.

After cooling to 4˚C, samples were centrifuged at 15,000 x g for 2 minutes and the supernatant was transferred into total recovery vials (Waters, Milford, MA). A study pool quality control (SPQC) sample was made by mixing 5 µL from each sample into a separate vial.

Quantitative sample analysis: Each sample was analyzed on a “nano-flow” LC-

MS/MS using a Waters nanoAcquity LC interfaced to a Thermo Q Exactive HF with a nanoelectrospray ionization source with tune parameters: spray voltage 1.8 kV and capillary temperature 250˚C. For quantitative analysis, 3.5 µL (450 ng protein equivalent) was analyzed per injection. A single-pump trapping configuration was used for LC separation, including trapping on a 180 µm x 20 mm Symmetry C18 5 um column

(Waters) and analytical separation on a 75 µm x 250 mm 1.8 µm particle HSS T3 column (Waters). Trapping utilized 5 µL/min at 99.9/0.1 v/v water/acetonitrile for five minutes, while analytical separation used a gradient from 5-30% acetonitrile (0.1% formic acid) over 90 minutes at 0.4 µL/min and a column temperature of 55˚C.

Label-free quantitative LC-MS/MS on the Q Exactive HF utilized an MS1 scan from m/z 375-1600 at 120,000 resolution, an AGC target of 3e6 ions, and a maximum injection time (IT) of 50 msec. The MS1 scan was followed by MS/MS (MS2) of the top

12 most abundant ions at 1.2 m/z isolation width, 30,000 resolution, AGC target of 1e4 ions, normalized collision energy (NCE) of 27 V, and dynamic exclusion of 20 sec.

The SPQC sample was used to condition the column prior to beginning the analysis of the sample set, and also before, during, and after the samples for quality control purposes. The run order was an interwoven block design. Data analysis was performed in Rosetta Elucidator v4.0 (Rosetta Biosoftware, Inc., Seattle, WA) feature detection and accurate mass and retention time alignment. Relative peptide abundance was calculated based on area-under-the-curve (AUC) of aligned features across all the runs. MS/MS (*.mgf) data from each run was searched against a custom-built protein sequence database containing a deduplicated aggregate of H. influenzae species, taxonomy downloaded from pubmed.gov (NCBI RefSeq entries), Uniprot

(www.uniprot.org), as well as contaminants from the cRAPome (http://crapome.org/).

The database also contained a reversed- sequence “decoy” database for false discovery rate (FDR) determination. Mascot v2.4 (Matrix Science, Inc., Boston, MA) was used to search DDA data. Amino acid modifications allowed in database searching included fixed carbamidomethyl cys (+57), and variable deamination of Asn and Gln (+1) and oxidation of Met (+16). Tryptic enzyme cleavage rules were followed with up to 2 missed cleavages, a peptide tolerance of +/- 5 ppm and a product tolerance of +/- 0.02 Da. Data was processed to the PeptideTeller data curation algorithm to determine FDR and was annotated at 0.5% peptide FDR. A total of 6973 peptides and 1183 proteins were quantified. Peptide intensities were robust mean normalized across all samples, and

98 relative protein abundance was calculated as the simple sum of the peptide intensities to all samples. Peptides were annotated to the single most likely protein under the principles of Occam’s Razor using the ProteinProphet algorithm (441). Summary for the proteomics data are contained in Table 3.2 (sample identifier information). The %CV for the technical replicates analyzed throughout the study showed excellent analytical reproducibility, with a median 4.9% including all proteins, or 4.0% including proteins with

2 or more peptides. Differential expression determination was based on calculation of the fold-change (+ reflecting upwards in the direction of S, negative reflecting upwards in the direction of NS), as well as a t-test after log2 transformation of protein intensity values. Student’s t-Test value reported in Table 3.2 are not corrected for multiple hypothesis. Raw proteomics data for this experiment have been made available on the

MassIVE data repository at ftp://massive.ucsd.edu/MSV000082399.

3.3 Results

Transient heme-iron restriction of NTHI promotes the formation of intracellular bacterial communities in an experimental model of OM

Heme-iron restriction of NTHI alters bacterial interactions with host epithelial cells. In our prior studies using strains transiently depleted of heme-iron, we observed an increase in the populations of intracellular NTHI within cultured epithelial cells (59, 60).

Although NTHI can internalize into epithelial cells, the fate is typically nonproductive due to trafficking into degradative pathways (55, 303). In contrast, when NTHI are transiently restricted for heme-iron, we observe large intracellular clusters of bacteria reminiscent of

IBCs growing within cultured middle ear epithelial cells (60). Within clinical biopsies,

99 intracellular bacteria have been observed in middle ear epithelium obtained from children with OM (407). Therefore, we sought to determine whether prior heme-iron restriction of NTHI would promote the formation of IBCs by NTHI within the chinchilla middle ear mucosal epithelium in a preclinical model of OM. The prototypical NTHI strain

86-028NP was cultured in DIS medium in the presence or absence of 2 µg/mL heme- iron for 24 hours resulting in two cultures, one continuously exposed to heme-iron (CE) and one transiently restricted of heme-iron (TR; Figure 3.1A). The cultures were normalized for viable counts and inoculated directly into the chinchilla middle ear (Figure

3.1A). Two days post inoculation, middle ears were processed to evaluate the formation of IBCs within the middle ear mucosae (Figure 3.1A). The anatomy of the middle ear includes epithelial mucosae typically consisting of 2-3 layers of cells overlaying a bony septum (Figure 3.1B). NTHI was visualized within thin sections using an antiserum directed against NTHI OMPs (green, Figure 3.1C). In addition, counterstains were included to visualize the host cell membranes (red) and DNA (blue) (Figure 3.1C). We first observed that infection with either transiently restricted or continuously exposed

NTHI (Figure 3.1D, F) resulted in the expansion of the middle ear epithelium as compared with the sham treated ear (Figure 3.1B). Infection with transiently-restricted

NTHI resulted in the formation of IBCs that filled the volume of the cells and multiple

IBCs were observed within each thin section (Figure 3.1D). The staining of IBCs using the antiserum was specific for NTHI as evidenced by the absence of fluorescence in the green channel when the primary antibody was excluded (Figure 3.1E). In contrast to the transiently restricted NTHI, intracellular clusters of NTHI were not readily observed within the mucosal layer of chinchilla middle ears infected with continuously exposed NTHI

100

Figure 3.1. Transient heme-iron restriction of NTHI promotes intracellular bacterial community formation in a preclinical model of otitis media. (A) Schematic representation of environmental heme-iron conditioning of NTHI strain 86-028NP and inoculation of chinchilla middle ears. Two days post inoculation, middle ears were excised, fixed, and paraffin embedded and thin sections from sham-treated, transiently restricted, or continuously exposed infected ears were processed for microscopy. TR, transiently restricted NTHI; CE, continuously exposed NTHI. (B) Four-nanometer thin sections of sham-treated middle ear bullae were stained with hematoxylin and eosin (left panel) or processed for immunofluorescence microscopy and stained with wheat germ agglutinin and 4′,6-diamidino-2-phenylindole (right panel). The lumen, mucosa, and bone are denoted for anatomical orientation. (C) Immunofluorescence micrograph gallery depicting the surface staining of transiently restricted NTHI by anti-OMP labeling (green), staining of the host cell membrane by wheat germ agglutinin (red), and staining of host and bacterial DNA by Hoechst stain (blue). (D) Representative images of transiently restricted NTHI IBCs within middle ear mucosal epithelium, including a three-dimensional rendering of a series of optical sections (left panel) to depict the orthogonal views of IBCs filling the entirety of the epithelial cell and visualized using the fluorophores depicted in panel C. Multiple IBCs were observed in each thin section. (E) A no-primary-antibody control of a thin section of a middle ear infected with transiently restricted NTHI. The thin section is sequential to those in panel D and depicts the specificity of NTHI labeling. (F) Representative image of middle ears infected with continuously exposed NTHI depicts the absence of IBCs in these thin sections. Bar, 10 µm for all images. (G) Quantification of IBCs in thin sections of ears infected with transiently restricted or continuously exposed NTHI. Three sequential thin sections were counted for three ears from each cohort. Significance was determined using a two-tailed Student t test.

101

(Figure 3.1F). Although we did observe rare instances of intracellular populations of continuously exposed NTHI, the number of IBCs per thin section was significantly lower in ears infected with continuously exposed NTHI compared to transiently restricted NTHI

(Figure 3.1G). Continuously exposed NTHI resided mainly in a biofilm with immune cell infiltrate localized on top of the epithelial mucosae. Hence, as observed in other highly recurrent infections (388), we observe a productive intracellular lifestyle for NTHI during experimental OM as a consequence of nutritional status.

Heme-iron restriction significantly increases intracellular survival of NTHI

The differences in the formation of IBCs between the transiently restricted and continuously exposed NTHI suggest that nutritional conditioning increases the association with and survival within middle ear epithelial cells. Technical limitations prohibit quantification of intracellular bacteria at early time points in vivo. We therefore capitalized upon our in vitro culture model to evaluate the ability of NTHI to survive within primary normal human bronchial epithelial cells (NHBE; (59, 60)). Gentamicin protection assays were used to distinguish between NTHI association (adherent and invaded

NTHI) and invasion (protected intracellular NTHI) following transient heme-iron restriction. There was no significant difference in either association or the initial invasion of transiently restricted or continuously exposed NTHI with NHBEs (Figure 3.2A).

However, when invasion was normalized for the total number of bacteria associated with the cells, we observed a statistically significant increase in the intracellular populations of the transiently restricted NTHI as compared with the continuously exposed (Figure

3.2B). Consistent with our in vivo and in vitro observations (Figure 3.1C, 3.1D, 3.2B), there was a statistically significant increase in the intracellular survival of transiently restricted NTHI 102

Figure 3.2. Transiently restricted NTHI invades and survives within human epithelial cells in intracellular bacterial communities. (A) Normal human bronchial epithelial (NHBE) cells were co- cultured with either transiently restricted (TR) or continuously exposed (CE) NTHI for 1 h, and total bacterial association (no gentamicin) or intracellular bacteria only (+ gentamicin) were quantified. Results are depicted as the percentage of inoculum that remained viable with or without gentamicin treatment. Statistical significance was determined using a two-tailed Student t test of the means from duplicate wells from each of three biological replicates. Error bars indicate standard errors of the means. ns, not significant. (B) The invasion index of transiently restricted (TR) and continuously exposed (CE) NTHI defined as the number of viable intracellular bacteria when normalized for total association. Statistical significance was determined using a two-tailed Student t test on the mean from technical duplicates from each of three biological replicates. Error bars represent standard errors of the means. (C) Intracellular survival of transiently restricted (TR) and continuously exposed (CE) NTHI following infection of NHBE cells for 2 or 24 h. Results are depicted as the percentage of inoculum that remained viable following gentamicin treatment. Statistical significance was determined using a two-tailed Student t test on the mean from duplicates from each of eight biological replicates for the 2-h time point and five biological replicates for the 24-h time point. Error bars represent standard errors of the means. (D to G) NTHI strain 86- 028NP(pGM1.1) expressing green fluorescent protein was transiently restricted (TR) or continuously exposed (CE) to heme-iron and then co-cultured with NHBE cells for 4 (D and E) or 24 (F and G) hours. Epithelial cell membranes were labeled with wheat germ agglutinin conjugated to Alexa Fluor 594 (red), and bacteria were visualized by GFP fluorescence (green). Bar, 10 µm. Images depict the formation of early IBCs by TR NTHI as early as 4 h post-inoculation (D) that progress to mature IBCs at 24 h (F). CE NTHI localizes in circular compartments (E, inset) and does not form IBCs 24 h post-inoculation (G).

103 following gentamicin treatment for at least 24 hours (Figure 3.2C). Neither the continuously exposed nor transiently restricted NTHI were cytotoxic to NHBE cells at any time point (Figure 3.3). Although there was a trend for an increase in the association of the continuously exposed NTHI as compared with the transiently restricted NTHI (Figure

3.2A), there was no significant difference in the percentage of intracellular bacteria upon titration of the multiplicity of infection (Figure 3.4). These observations suggest that there are differences in the subcellular localization of the transiently restricted and continuously exposed NTHI. We would predict that the transiently restricted NTHI population occupies niches that allow productive growth whereas continuously exposed

NTHI proceed through a degradative pathway. Using fluorescent reporter strains, we observed that both transiently restricted and continuously exposed NTHI were internalized into NHBE monolayers at four hours (Figure 3.2D, E). For the transiently restricted NTHI, we observed intracellular bacteria that resemble the early stages of IBC formation (Figure 3.2D). In contrast, the continuously exposed NTHI appeared to be compressed into circular structures suggesting that the bacteria are within membrane enclosed compartments (Figure 3.2E). The intracellular location of NTHI is supported by the use of specific antisera in the presence and absence of permeabilization of host cell membranes to identify intracellular and extracellular bacteria, respectively (Figure 3.5).

Consistent with our previously published reports using primary respiratory epithelial cells

(59, 60), we observed increased intracellular populations of transiently restricted NTHI within the cytoplasm of NHBEs at 24 hours post inoculation (Figure 3.2F), while intracellular populations of continuously exposed NTHI decreased and did not progress to IBC formation (Figure 3.2G). Therefore, we provide evidence that transiently restricted and continuously exposed NTHI reside in different subcellular locations using gentamicin protection, specific antibody labeling and optical sectioning approaches. Taken together, 104

Figure 3.3. Nutritionally conditioned NTHI is not cytotoxic to NHBE cells. The potential cytotoxicity of transiently restricted (TR) or continuously exposed (CE) NTHI to NHBE cells at 4 or 24 h was assessed by the release of lactate dehydrogenase (LDH). Experiments were performed in biological duplicate, and data are reported as the fold change in cytotoxicity, compared to uninfected cells. Error bars represent standard errors of the means. The maximum release of LDH was determined by the lysis of all cells with 9% Triton X-100. Neither TR nor CE NTHI displayed a cytotoxic effect on NHBE cells at any time point used in this study.

105

Figure 3.4. Effect of multiplicity of infection on intracellular survival of NTHI. Intracellular survival assays at 2 h post inoculation were performed with NHBE cells inoculated with transiently restricted (TR) or continuously exposed (CE) bacteria at a multiplicity of infection (MOI) of 1, 12.5, 25, 50, or 100 per cell. The percentage of viable inoculum remaining following gentamicin protection is reported. Data represent the means from duplicate wells for each of three biological replicates with standard errors of the means.

106

Figure 3.5. Intracellular localization of nutritionally conditioned NTHI in the presence and absence of permeabilization. To confirm the intracellular localization of transiently restricted (TR) or continuously exposed (CE) NTHI in NHBE cells, confluent monolayers were co-cultured with NTHI for 4 hours. Cells were permeabilized or non-permeabilized in parallel and analyzed by immunofluorescence. Epithelial cell membranes were visualized by wheat germ agglutinin conjugated to Alexa Fluor 594 (red), and DNA (bacterial and host) was visualized by 4′,6-diamidino-2-phenylindole (DAPI) and pseudo colored white. NTHI was labeled with anti-OMP and detected with protein A-Alexa Fluor 488 (green). In non- permeabilized cells (left), intracellular bacteria are white (indicated by white arrows) and extracellular bacteria are green/white (indicated by green arrows). TR NTHI is frequently observed within cells (top left panel) while CE NTHI is mostly external with few bacteria residing intracellularly. In permeabilized cells (right), all bacteria regardless of nutritional condition or localization are visualized as green/white. Bar, 10 µm.

107 our data demonstrate that transient heme-iron restriction of NTHI promotes formation of

IBCs within epithelial cells.

Heme-iron restriction of NTHI alters bacterial trafficking to early endosomes

The ability of NTHI to grow within the cell following transient restriction of heme- iron, combined with the differential subcellular localization of the continuously exposed

NTHI suggests that nutritional conditioning promotes NTHI escape or evasion of vacuolar trafficking to gain access to the cytoplasm. Consistent with published observations, once internalized into epithelial cells, continuously exposed NTHI appears to remain contained within circular compressed compartments (Figure 3.2E). Thus, we would predict that the continuously exposed NTHI is trafficked through the endolysosomal pathway (55, 303). The observation of significant increases in intracellular survival of NTHI following prior heme-iron restriction of the bacteria (Figure

3.2C), led us to hypothesize that continuously exposed NTHI would be more often associated with markers of the endolysosomal pathway than the transiently restricted

NTHI. To evaluate the differential trafficking of the transiently restricted and continuously exposed NTHI, NHBE cells were co-cultured with NTHI and subsequently examined for bacterial colocalization with the specific marker of early endosomes, early endosomal antigen 1 (EEA1; (442)). We observed continuously exposed NTHI within EEA1- containing membrane compartments at 4 hours post infection (Figure 3.6A, B, E). The

EEA1 signal was distributed around the periphery of the entire compartment (Figure

3.6E, red arrows). Co-localization of continuously exposed NTHI within EEA1-containing membranes was observed throughout the NHBE cells. In contrast, transiently restricted

NTHI did not appear to be localized within EEA1-containing compartments and

108

Figure 3.6. Transient heme-iron conditioning of NTHI alters trafficking to early endosomes. (A) NHBE cells were co-cultured with either transiently restricted (TR) or continuously exposed (CE) NTHI strain 86-028NP(pGM1.1) for 4 h. Epithelial cell membranes were labeled with wheat germ agglutinin conjugated to Alexa Fluor 350 (blue), and bacteria were visualized by GFP fluorescence (green). Early endosomes were labeled with rabbit antibody to early endosomal antigen 1 (EEA1) protein and visualized with donkey anti-rabbit IgG conjugated to Alexa Fluor 594 (red). Representative images are shown for each condition (TR or CE) with individual and merged fluorescence images shown for depiction of colocalization. Colocalization of bacteria with EEA1 is observed as either yellow (merged) or red EEA1 label closely surrounding clusters of green NTHI bacteria. Bar, 10 µm. (B) Additional images representative of those depicted in panel A. (C) The number of colocalization events per cell was quantified by visual assessment of 200 individually infected cells. Statistical significance was determined by Mann-Whitney U test, and error bars represent standard errors of the means. (D and E) Transmission electron microscopy of NHBE cells co-cultured with TR or CE conditioned NTHI for 4 h and subsequently immunolabeled to detect bacterial association with EEA1. Early endosomes were labeled with rabbit antibody to EEA1 and visualized with anti-rabbit IgG antibody conjugated to an 18-nm colloidal gold particle. EEA1-containing vesicles devoid of bacteria (labeled B) are indicated by yellow arrows, while red arrows indicate EEA1- containing vesicles associated with bacteria. Bar, 500 nm.

109 appeared free in the cytoplasm (Figure 3.6A, B). In addition, EEA1 appeared to be associated with smaller vesicles that do not contain NTHI (Figure 3.6D, yellow arrows).

In some cases, we did observe EEA1 in close proximity to the transiently restricted

NTHI, but the protein was not distributed peripherally around the entire membrane compartment (Figure 3.6D, red arrow). The number of NTHI and EEA1 co-localization events was quantitatively assessed from 200 independent cells infected with either transiently restricted or continuously exposed NTHI and visualized by fluorescent microscopy (Figure 3.6C). The observation of a significant association of continuously exposed NTHI within EEA-1 membrane bound compartments is consistent with prior observations (55, 303). However, the transiently restricted NTHI demonstrates a different intracellular localization pattern with significantly fewer co-localization events per cell (Figure 3.6C). Based upon these observations, we predict that transiently restricted NTHI would not enter into the later stages of the lysosomal pathway, and promote intracellular survival.

Infection with transiently restricted NTHI changes LAMP1 distribution

Endosomes function to sort and direct internal cargo for eventual fusion with lysosomes for targeted degradation (443). Given the strong association of continuously exposed NTHI with EEA1, we predict that the fate of the continuously exposed NTHI would involve continuation through the endolysomal pathway. To test this, we investigated the co-localization of Lysosomal Associated Membrane Protein-1 (LAMP1) with transiently restricted or continuously exposed NTHI. At 24 hours transiently restricted NTHI did not co-localize with LAMP1, whereas there was evident colocalization of continuously exposed NTHI with LAMP1 (Figure 3.7A, B). These data suggest that continuously exposed NTHI traffick to the lysosomes whereas transiently 110

 Figure 3.7. Transient heme-iron restriction of NTHI alters lysosomal trafficking and biogenesis. (A and B) NHBE cells were co-cultured with transiently restricted (TR) or continuously exposed (CE) NTHI strain 86- 028NP(pGM1.1) and assessed for colocalization of NTHI with LAMP1 by immunofluorescence at 24 h post inoculation. Cell membranes were visualized with wheat germ agglutinin conjugated to Alexa Fluor 350 (blue), and NTHI was visualized by GFP fluorescence (green). LAMP1 bar, 10 µm. (C and D) Colocalization of TR and CE NTHI with LAMP1 was assessed by TEM. LAMP1 was labeled with rabbit anti-LAMP1 and detected with anti-rabbit IgG conjugated to 18-nm colloidal gold. LAMP1-positive vesicles devoid of bacteria (labeled B) are indicated by yellow arrows, and LAMP1-positive vesicles associated with bacteria are indicated with red arrows. Bar, 500 nm. (E to G) LAMP1 staining patterns depict differences in lysosomal biogenesis in NHBE cells infected with TR NTHI (F) or CE NTHI (G) at 5 h post inoculation. Uninfected cells are included for comparison (E). LAMP1 was visualized using rat anti-LAMP1 and detected with anti-rat IgG conjugated to Alexa Fluor 594 (red). TR NTHI infection promotes a diffuse LAMP1 staining pattern while CE NTHI infection promotes a punctate and often perinuclear staining pattern in NHBE cells. Uninfected NHBE cells display a combination of both diffuse and punctate staining. Bar, 10 µm. (H) LAMP1 staining patterns were quantified by visual counting of 100 individual NHBE cells from each condition (infected with either TR or CE NTHI). Statistical significance was determined using a two-tailed Student t test.

111 restricted NTHI avoid or escape this pathway. Vacuole-associated LAMP1 were also observed by immunogold transmission electron microscopy (TEM, Figure 3.7C, D). As we observed with EEA1, LAMP1 was distributed around vacuolar compartments that contained continuously exposed NTHI (Figure 3.7D, red arrows). In contrast, while we did observe transiently restricted NTHI associated with LAMP1 (Figure 3.7C, red arrows), we more often observed LAMP1 containing vesicles devoid of NTHI (Figure

3.7C, yellow arrows). Interestingly, we observed striking differences in the subcellular distribution of LAMP1 signal by immunofluorescence microscopy, dependent on whether

NTHI was transiently restricted or continuously exposed (Figure 3.7F, G). Uninfected

NHBE cells displayed LAMP1 signal that consisted of a mixture of diffuse and punctate staining patterns (Figure 3.7E). Infection of NHBEs with continuously exposed NTHI

(Figure 3.7G) results in a punctate LAMP1 staining pattern that appears to be centrally located within the cell. In marked contrast, infection of NHBEs with transiently restricted

NTHI (Figure 3.7F) results in a LAMP1 staining pattern that is almost entirely diffuse, suggesting that localization of LAMP1 is altered. Visual quantification of staining patterns per ~100 cells confirmed that infection with transiently restricted NTHI promotes a diffuse

LAMP1 distribution in close to 90% of cells while infection with continuously exposed

NTHI promotes this phenotype in only 20% of cells, with the majority of cells displaying a staining pattern of punctate LAMP1 (Figure 3.7H). Taken together, these data demonstrate that transiently restricted NTHI evades the lysosomes and may alter lysosomal biogenesis.

112

Internalization of NTHI into epithelial cells is prevented with pharmacological inhibition of endocytosis

The differential trafficking of transiently restricted and continuously exposed NTHI could be a consequence of the mechanism of internalization. Proteomic analysis indicates that the outer membrane profiles are similar between the transiently restricted and continuously exposed NTHI, suggesting that these two populations could use similar adhesins for initial interactions with the epithelial cells (Table 3.2). NTHI is internalized into the cell through multiple endocytic pathways (58). Therefore, we hypothesized that transiently restricted and continuously exposed NTHI may be internalized into epithelial cells through different pathways. To investigate this, NHBE monolayers were pretreated with a panel of pharmacological inhibitors to target endocytic pathways (Table 3.1) prior to co-culture with transiently restricted or continuously exposed NTHI. The effective dosage for each inhibitor was empirically determined with the use of appropriate indicators of endocytosis (i.e. cargo or cellular ultrastructure; Table 3.1 and Figure 3.8A).

There was no significant difference in the association of NTHI with epithelial cells in the presence or absence of the inhibitors (Figure 3.9). In the absence of inhibition (Figure

3.8B, untreated), intracellular populations were observed for both the transiently restricted and continuously exposed NTHI, consistent with our prior observations

(Figures 3.2D, E). We observed inhibition of internalization for both transiently restricted and continuously exposed NTHI in the presence of cytochalasin D (CytoD; an inhibitor of actin polymerization), chlorpromazine (CPZ; an inhibitor of clathrin or receptor-mediated endocytosis) and methyl-β-cyclodextrin (MβCD; an inhibitor of caveolae/lipid raft domains) as evidenced by the presence of extracellular bacteria on the cell periphery

(Figure 3.8B). The extent of NTHI internalization in the presence and absence of the

113

Table 3.2 Assessment of differentially expressed proteins of NTHI transiently restricted and continuously exposed to heme-iron.

114

Figure 3.8. Pharmacological inhibition of endocytosis pathways reveals that nutritionally conditioned NTHI is internalized into cells through multiple mechanisms. (A) Representative images depicting inhibition of uptake of known fluorescent cargo conjugates (Alexa Fluor 488, green) into NHBE cells in the presence of each pharmacological inhibitor [CPZ, chlorpromazine; MβCD, methyl-β- cyclodextrin; EIPA, 5-(N-ethyl-N-isopropyl)-amiloride]. The optimal concentration of cytochalasin D (CytoD) was determined by staining F-actin with phalloidin conjugated to Alexa Fluor 350. Epithelial cell membranes were labeled with wheat germ agglutinin conjugated to Alexa Fluor 594 (red). Bar, 10 µm. (B) Fluorescence microscopy was used to determine uptake of transiently restricted (TR) or continuously exposed (CE) NTHI strain 86-028NP(pGM1.1) following pharmacological inhibition of endocytosis pathways. NHBE cells were pretreated with pharmacological inhibitors prior to a 4-h incubation with TR or CE NTHI strain 86-028NP(pGM1.1). NTHI was visualized by GFP fluorescence (green), and epithelial cell membranes were labeled with wheat germ agglutinin conjugated to Alexa Fluor 594 (red). The far- right panel depicts infected cells with no inhibitor (Untreated) for comparison. Each experiment was performed in three biological replicates, and representative images are shown. Bar, 10 µm. (C) Viable intracellular bacteria were enumerated following gentamicin treatment of NHBE cells that were incubated with transiently restricted (TR) or continuously exposed (CE) NTHI for 1 h in the presence of pharmacological inhibitors compared to untreated controls. Statistical significance was determined by analysis of variance with means from duplicate wells from three independent biological replicates, and error bars represent standard errors of the means (**, P < 0.01; ****, P < 0.0001). The invasion index was calculated as the number of viable intracellular bacteria divided by total associated bacteria.

115

Figure 3.9. Association of nutritionally conditioned NTHI with NHBE cells is unchanged in the presence of pharmacological inhibitors. Total association (intracellular and extracellular) of transiently restricted (TR) or continuously exposed (CE) NTHI with NHBE cells pretreated with the indicated pharmacological inhibitors was determined at 1 h post-inoculation. Data represent the means from duplicate wells from each of three biological replicates, and the standard error of the mean is shown. Addition of pharmacological inhibitors did not affect the total association of either TR or CE NTHI with NHBE cells. CytoD, cytochalasin D; CPZ, chlorpromazine; MβCD, methyl-β-cyclodextrin; EIPA, 5-(N- ethyl-N-isopropyl)-amiloride.

116 inhibitors was further quantified using gentamicin protection (Figure 3.8C). We observed statistically significant decreases in internalization of transiently restricted NTHI upon addition of these three inhibitors, suggesting that internalization of transiently restricted

NTHI can occur through endocytic pathways. Although significantly reduced, treatment with these three inhibitors did not completely abolish internalization of transiently restricted NTHI, suggesting an alternate route of entry. Although the initial invasive populations of continuously exposed NTHI are significantly reduced from the transiently restricted NTHI (Figures 3.2B, 3.8C), treatment with CytoD, CPZ, and MβCD almost completely abolished internalization of continuously exposed NTHI (Figure 3.8B, C).

Inhibition of macropinocytosis promotes trafficking of transiently restricted NTHI to early endosomes and decreases intracellular survival

NTHI has been shown to use macropinocytosis for internalization into epithelial cells (171, 313). Since we observed uptake of both transiently restricted and continuously exposed NTHI through endocytic mechanisms (Figure 3.8B, C), we next asked whether there was a difference in the uptake of these two NTHI populations by macropinocytosis. The effective dosage for 5-(n-ethyl-n-isopropyl) amiloride (EIPA; an inhibitor of macropinocytosis) was empirically determined with the use of 70,000 MW

Dextran (Figure 3.8A). Pretreatment of NHBE monolayers with EIPA had no effect on the internalization of the continuously exposed NTHI. In contrast, EIPA reduced but did not significantly inhibit internalization of transiently restricted NTHI compared to untreated cells (Figure 3.8B, C). Thus, our data reveal a subpopulation of transiently restricted NTHI that are susceptible to inhibition of macropinocytosis that was not observed with the continuously exposed NTHI.

117

In addition to the subtle differences in the internalization of continuously exposed or transiently restricted NTHI in the presence of EIPA, we observed a striking difference in the intracellular localization of transiently restricted NTHI following uptake in EIPA- treated cells. Microscopic analysis revealed that in the presence of EIPA, transiently restricted NTHI now appeared to traffic similar to the continuously exposed NTHI and were localized within circular structures, resembling vesicles or membrane compartments (Figure 3.8B, inset). This was a remarkable difference compared to untreated cells, where transiently restricted NTHI appear free in the cytoplasm in early

IBCs (Figure 3.8B, untreated and Figure 3.2D). In contrast, EIPA did not alter the subcellular localization of the continuously exposed NTHI (Figure 3.8B). Therefore, we hypothesized that EIPA promotes trafficking of transiently restricted NTHI through the endolysosomal pathway.

To determine the fate of transiently restricted NTHI in the presence of EIPA, pretreated NHBEs were co-cultured with transiently restricted or continuously exposed

NTHI to visualize co-localization with EEA1. As expected, treatment with EIPA did not alter the trafficking of continuously exposed NTHI to early endosomes (Figure 3.10C, D).

Consistent with our hypothesis, we observed significantly increased co-localization of transiently restricted NTHI with EEA1 in the presence of EIPA as shown by quantitative assessment of fluorescent images (Figure 3.10A, B). These data suggest that blocking macropinocytosis does not prevent bacterial entry, but rather promotes trafficking of transiently restricted NTHI to the endocytosis pathway.

Based upon the observation that transiently restricted NTHI is diverted into the endolysosomal pathway upon treatment with EIPA, we predicted that inhibition of macropinocytosis would decrease the intracellular survival of transiently restricted NTHI. 118

Figure 3.10. Inhibition of macropinocytosis redirects transiently restricted NTHI to the endolysosomal pathway and decreases intracellular survival. (A to D) NHBE cells were pretreated with 60 µM EIPA prior to co-culture with transiently restricted (TR) or continuously exposed (CE) NTHI strain 86-028NP(pGM1.1) and visualized for colocalization of NTHI with EEA1 by fluorescence microscopy. NTHI was visualized by GFP fluorescence (green), EEA1 was labeled with rabbit antibody to EEA1 and visualized with donkey anti-rabbit IgG conjugated to Alexa Fluor 594 (red), and epithelial cell membranes were visualized with wheat germ agglutinin conjugated to Alexa Fluor 350 (blue). Representative images depict colocalization of TR (A) or CE (C) NTHI with EEA1 as observed by yellow fluorescence or red EEA1 labeling closely surrounding clusters of green NTHI bacteria. Bar, 10 µm. (B) EIPA pretreatment of NHBE cells significantly increases the number of colocalization events of TR NTHI with EEA1 in infected cells compared to infected cells that were not treated with EIPA. (D) Colocalization of CE NTHI with EEA1 does not significantly change in the presence or absence of EIPA. Statistical significance was determined by Mann-Whitney U test of colocalization events from a total of 200 independent cells from three biological assays. (E) Viable intracellular bacteria were enumerated following gentamicin treatment of NHBE cells infected with TR or CE NTHI in the presence and absence of EIPA. Statistical significance was determined by two-tailed Student’s t test of the means for duplicate wells from three independent biological replicates, and error bars represent standard errors of the means. (F to H) Intracellular bacterial communities were enumerated in chinchilla middle ear epithelial cells incubated with TR NTHI in the presence or absence of EIPA. (F and G) Bacteria were visualized by GFP fluorescence (green), epithelial cell members were stained with wheat germ agglutinin conjugated to Alex Fluor 594 (red), and host and bacterial DNA were labeled with Hoechst stain (blue). Bar, 10 µm. (H) Statistical significance was determined by two-tailed Student’s t test of the mean from duplicate wells from each of three independent biological replicates, and error bars represent standard errors of the means.

119

As previously observed (Figure 3.2C), the intracellular survival of continuously exposed

NTHI was decreased as compared with the transiently restricted NTHI (Figure 3.10E).

EIPA treatment significantly reduced the intracellular survival of transiently restricted

NTHI (Figure 3.10E). In addition, EIPA also reduced the intracellular survival of the continuously exposed NTHI, although this was not significant (Figure 3.10E). Taken together, these data suggest that entry through macropinocytosis is the primary pathway that promotes intracellular survival of transiently restricted NTHI.

To determine whether the long term survival associated with productive growth leading to IBC formation is inhibited by EIPA, cultured primary chinchilla middle ear epithelial (CMEE) cells were co-cultured with transiently restricted NTHI in the presence and absence of EIPA. Consistent with our observations using the preclinical model of

OM (Figure 3.1), IBCs were readily observed in cultured CMEE cells at 24 hours (Figure

3.10F). Pretreatment with EIPA significantly reduced the number of IBCs formed by transiently restricted NTHI (Figure 3.10G, H). Taken together, these data suggest that entry via the macropinocytosis pathway promotes intracellular survival of transiently restricted NTHI leading to productive IBC formation.

3.4 Discussion

Recent studies have revealed that NTHI can use multiple pathways for entry into host cells and typically traffics through the endolysosomal pathway for degradation within lysosomes (58). While we also observed that multiple mechanisms are used for

NTHI internalization, our studies demonstrated that transient limitation of an essential nutrient redirects the fate of NTHI from a detrimental outcome to intracellular growth

120 leading to IBCs (Figures 3.1, 3.2, and 3.11). Pharmacological inhibition of various endocytic pathways demonstrated that the uptake of both transiently restricted and continuously exposed NTHI involves actin remodeling and can occur through clathrin- and lipid-raft dependent mechanisms (Figure 3.8). Remarkably, inhibition of macropinocytosis ablated the formation of IBCs due to trafficking of NTHI through the endolysosomal pathway (Figures 3.8, 3.10, and 3.11). The formation of a viable intracellular population may provide a myriad of benefits including protection from the host immune system, access to intracellular nutrients, or a respite from competing bacteria in polymicrobial infections. Additionally, intracellular reservoirs may be protected from antibiotics that are the first line of treatment in many cases of OM (444). Taken together, nutrient availability is an important environmental cue for phenotypic responses culminating in an intracellular reservoir for NTHI.

Intracellular lifestyles are commonly associated with chronic and recurrent infections. IBCs were first described for UPEC and were later found to act as reservoirs for recurrent infection (351-357). UPEC IBCs have been identified in bladder biopsies and the urine of patients with urinary tract infections (445-448). More recently, the diversity of pathogens that form IBCs has expanded to include Klebsiella, Proteus, and

Helicobacter (363-366). We now provide evidence for IBC formation in the pathogenesis of NTHI-mediated experimental OM (Figure 3.1). Recent studies have demonstrated that

UPEC IBCs remain metabolically active in the bladder (360), and intracellular populations of NTHI within endosomes were found to remain metabolically active for 24 hours (303). Future studies will investigate the metabolic activity of NTHI IBCs and the contribution to persistence and reoccurrence of OM.

121

Figure 3.11. Proposed model for differential trafficking of transiently restricted NTHI through macropinocytosis resulting in IBC formation. In the absence of EIPA (top), both transiently restricted (TR, left) and continuously exposed (CE, right) NTHI enter the cells through endolysosomal pathways: clathrin-mediated endocytosis (blue circles), lipid raft/caveolae-mediated endocytosis (orange rectangles), and macropinocytosis (membrane ruffling by actin polymerization, red). The ability to enter the cell through these various pathways appears to be independent of prior heme-iron status. Once internalized, CE NTHI (yellow bacteria) traffics to the early endosomes (red circles) and finally to the lysosomes (purple circles), where the bacteria are degraded. TR NTHI (green bacteria) also enters the cells through endolysosomal pathways and traffics to the early endosomes and lysosomes. In contrast, the subpopulation of transiently restricted NTHI that enters through macropinocytosis either completely evades or escapes this pathway (indicated by “?”) to form intracellular bacterial communities in the cell cytoplasm. In the presence of the macropinocytosis inhibitor EIPA (bottom), trafficking of continuously exposed NTHI through the endolysosomal pathway remains unchanged. Transiently restricted NTHI, entering the cell through clathrin- or lipid raft/caveolae-mediated endocytosis, now localizes to the early endosomes. This shift in trafficking targets TR NTHI for degradation by the endolysosomal pathway and significantly decreases intracellular survival of this population.

122

An important lingering question is the mechanism underlying the ability of transiently restricted NTHI to evade the endocytic pathway. Pharmacological inhibition suggests that the majority of NTHI, regardless of heme-iron status, are taken up through actin-mediated pathways. Furthermore, when either clathrin- or lipid raft-mediated endocytosis pathways were inhibited, we observed similar decreases in viable intracellular populations. Recent studies demonstrate that in addition to the removal of cholesterol to prevent lipid raft-mediated endocytosis, MβCD also impedes the formation of receptor-mediated clathrin-coated endocytic vesicles (449). Prior studies demonstrate that platelet activating factor receptor-mediated endocytosis predominates over macropinocytosis for uptake of NTHI (313). Thus, it appears that the subpopulation of transiently restricted NTHI which are taken up by macropinocytosis may be primed for increased survival within or escape from macropinosomes compared to the continuously exposed NTHI.

To determine the potential bacterial factors that contribute to differential trafficking and increased intracellular survival of NTHI, we compared the proteome of transiently restricted and continuously exposed NTHI (Table 3.2). Differential uptake and trafficking could result from differences in the composition of the outer membrane.

Transiently restricted NTHI demonstrated an increase in transferrin-binding protein 2, but there were no significant differences in adhesins or other OMPs. Interestingly, we observed increased levels of lipoate-protein ligase B (LipB) in the transiently restricted population. LipB is a member of the lipoic acid biosynthetic pathway with the lipoyl synthase LipA (Table 3.2). Lipoate-dependent metabolism is required for optimal cytosolic replication and virulence of L. monocytogenes and B. pseudomallei (450-452).

Host-derived lipoic acid is also important for C. trachomatis intracellular growth and

123 development (453). Ongoing studies will determine the contribution of lipid modifications to invasion and intracellular survival of NTHI in epithelial cells.

The existence of a viable, protected intracellular population of NTHI promoted by transient heme-iron restriction reveals new opportunities for the design of novel therapeutics for chronic and reoccurring OM. Our data underscore the importance to consider extracellular and intracellular niches of NTHI for optimal therapeutic approaches. We demonstrated that blocking macropinocytosis significantly decreased intracellular survival of NTHI by redirecting internalized bacteria through the endolysosomal pathway for degradation (Figure 3.11). Similarly, we have previously shown that inhibition of the actin related protein complex, Arp2/3, prevents NTHI uptake into epithelial cells (298). Therefore, it is tempting to speculate that the use of macropinocytosis or actin inhibitors could prevent NTHI persistence and reduce recurrent episodes of OM. Indeed, our data demonstrate the feasibility of the use of macropinocytosis inhibition as independent or adjunct therapies to reduce the intracellular burden of NTHI. Although our data contribute to the growing body of knowledge for the diverse lifestyles and persistent nature of NTHI, a greater understanding of the bacterial factors contributing to IBC formation will be necessary in order to design specific and targeted therapeutics.

124

Chapter 4. Cytokine responses to transiently heme-iron restricted NTHI-mediated infection.

4.1 Introduction

As an opportunistic pathogen, the migration of NTHI from the commensal environment of the nasopharynx to sites of infection stimulate an innate immune response by the host. NTHI-mediated infection often results in a robust inflammatory response by the host which can become prolonged and lead to tissue damage (454).

Inflammation is the body’s initial response to infection, and acts to alert the host to the presence of foreign stimuli or bodily insult. Characterized by heat, pain, and swelling, inflammation at the site of infection stimulates the release of pro-inflammatory mediators, such as cytokines and chemokines, to attract leukocytes for eradication of the pathogen or foreign stimuli. Inflammation within the middle ear space is routinely observed during both acute and chronic OM and may contribute to complications of OM such as hearing loss. However, a subset of OM is subclinical and asymptomatic, often recurring despite repeated antibiotic therapies. The ability for NTHI to persist within the middle ear mucosae without stimulating an initial robust inflammatory response may contribute to recurrent OM.

Epithelial cells are a first line of defense to the external environment and mount an inflammatory response through the secretion of cytokines (455). NTHI infects multiple types of epithelium within the host, including the middle ear and lung epithelia. NTHI simulates the release of pro-inflammatory cytokines and antimicrobial peptides from human airway and middle ear epithelial cells through upregulation of the NF-kB pathway

125 and through synergistic action with endogenous factors (456, 457). Previous work in our laboratory suggests a differential immune response to infection with heme-iron restricted

NTHI, both in vitro and in vivo (59, 60). Loss of Sap transporter function, an ABC transporter involved in NTHI heme-iron acquisition, results in reduced inflammatory cytokine production by epithelial cells stimulated with a sapA mutant in vitro (59).

Further, the observation that heme-iron status of NTHI may dictate the severity of the host inflammatory response can be extended to studies of NTHI under environmental heme-iron restriction. Specifically, our prior studies demonstrated that infection with

NTHI that has been continuously exposed to heme-iron results in a pro-inflammatory response with immune cell infiltrate and significant tissue damage in a preclinical model of OM (60). In contrast, infection with NTHI that has been transiently restricted of heme- iron results in a marked decrease in tissue destruction in the middle ear in vivo (60).

Therefore, we performed a pilot study to investigate cytokine secretion by epithelial cells in response to infection with transiently restricted or continuously exposed

NTHI. We utilized both in vitro cell culture and an established in vivo infection model to examine cytokine production in response to nutritionally conditioned NTHI. We observed that transient heme-iron restriction of NTHI results in a trend towards decreased cytokine production by infected epithelial cells both in vitro and in vivo. Further, granulocyte macrophage colony-stimulating factor (GM-CSF) production is significantly decreased in vitro in response to infection of respiratory epithelium with transiently restricted NTHI compared to continuously exposed NTHI. The data presented herein reinforce the importance of investigating immune responses to transiently restricted NTHI and will direct future studies towards more targeted investigations into the role of GM-CSF during

NTHI-mediated OM.

126

4.2 Materials and methods

Bacterial strains, cell lines, and media

NTHI strain 86-028NP is a minimally passaged clinical isolate which has been sequenced and characterized in the chinchilla model of OM (60, 103). For routine culture, NTHI was grown on chocolate agar plates (Fisher Scientific, Pittsburgh PA). For routine liquid culture, NTHI was grown in an iron-depleted defined iron source (DIS) medium supplemented with β-NAD (112). Where indicated, DIS was supplemented with

2 µg/mL heme (Millipore Sigma, St. Louis, MO). NTHI was transiently restricted or continuously exposed to heme-iron in 24-hour liquid culture as previously described (60).

◦ Briefly, strain 86-028NP was grown overnight at 37 C in a 5% CO2 atmosphere on chocolate agar plates. Cells were resuspended in DIS, adjusted to an OD490 of 0.65 and diluted 1:10 into pre-warmed DIS medium containing either 0 or 2 µg/mL heme. Cultures

◦ were grown for 24 hours statically at 37 C, 5% CO2 and normalized to an OD490 of 0.37 in DIS containing 2 µg/mL heme to generate parallel transiently restricted and continuously exposed cultures, respectively.

◦ NY) at 37 C with 90% humidity and 5% CO2.

Ethics statement

All animal experiments were carried out in strict accordance with the accredited conditions in the Guide for the Care and Use of Laboratory Animals of the National

Institutes of Health. The protocol was approved by the Institutional Animal Care and Use

127

Chinchilla model of otitis media

Healthy adult chinchillas (Chinchilla lanigera) were obtained from Rauscher’s chinchilla ranch (LaRue, OH). For initial studies of the cytokine response to in vivo infection with NTHI, two cohorts of three chinchillas were transbullarly challenged with

863 CFU of transiently restricted or 866 CFU of continuously exposed NTHI in a total volume of 300 µL saline per ear. Chinchillas were monitored over the 14-day study period by otoscopy and tympanometry. Middle ear effusions were collected by epitympanic taps on days 4, 7, 10, and 14. Bacterial burden in middle ear effusions was determined by serial dilution and plating on chocolate agar plates for viable CFU/mL at time of collection. Remaining samples were frozen at -80°C until processing for detection of cytokines.

Detection of NTHI-infected epithelial cell cytokine production in vitro

NHBE cells were seeded onto transwell membranes at 2.3e6 cells per transwell in a 6-well plate and grown to confluency. Cells were inoculated with transiently restricted or continuously exposed NTHI strain 86-028NP at an MOI of 25 in triplicate wells. One hour following inoculation, cell media was aspirated to remove non-adherent bacteria, and the epithelial cell surface was washed once with 500 µL 1X DPBS. At 24 hours post-inoculation, spent medium was collected from the basolateral surface of the transwells. Transwells of uninfected NHBEs were included as a control and used to calculate fold change in cytokine production. Cytokines present in the spent medium

128 were detected by the Proteome Profiler Human Cytokine Array kit (R&D Systems, catalog #ARY005, Minneapolis, MN) as previously described (59). Samples were incubated with biotinylated detection antibodies and the resulting cytokine-antibody complex was bound to a cognate antibody immobilized on a nitrocellulose membrane.

Relative amounts of each cytokine were detected by measuring streptavidin-horseradish peroxidase (HRP) chemiluminescence. The fold-change in cytokine production from uninfected control NHBE transwells was determined by measuring the pixel density at each antibody spot using ImageJ.

Detection of inflammatory cytokines in middle ear effusions

Analysis of inflammatory cytokines present in chinchilla middle ear effusions was performed using the Human Inflammatory Cytokine Multi-Analyte ELISArray Kit

(QIAGEN). This kit provides simultaneous detection of interleukin (IL)-1α, IL-1β, IL-2, IL-

4, IL-6, IL-8, IL-10, IL-12, IL-17A, IFN-ɣ, TNF-α and GM-CSF from a single sample.

Effusions collected from the left ear at day 7 post-inoculation were chosen for this initial investigation because we were able to recover effusions with sufficient volume for testing from all left ears in each cohort (Table 4.1). Samples were diluted 1:5 in sample dilution buffer and the array was performed according to manufacturer’s instructions. The absorbance levels of the cytokines were measured on a plate reader at 450 nm.

Standard errors of the mean were calculated from samples collected from three independent animals.

129

Table 4.1. Volumes of recovered chinchilla middle ear fluid during experimental OM. (L) = left ear; (R) = right ear. Gray shading indicates no sample. Animal 100 and animal 153 reached endpoint criteria and were sacrificed prior to D+10 of the study.

130

4.3 Results

Epithelial cell cytokine responses to nutritionally conditioned NTHI in vitro

Previous studies in our lab revealed that heme-iron restriction of NTHI alters epithelial cell homeostasis and results in decreased severity of OM in the chinchilla model (59, 60). Chinchillas infected with transiently restricted NTHI have attenuated inflammation during the course of disease and decreased tissue damage in the middle ear compared to animals infected with continuously exposed NTHI (60). As cytokines play a significant role in modulation of the inflammatory response during disease progression, we hypothesized that cytokine production by epithelial cells stimulated with

NTHI would be dampened compared to cells stimulated by continuously exposed NTHI.

Using a transwell model of respiratory epithelial cell growth, we analyzed inflammatory cytokine secretion in spent NHBE supernatants at 24 hours post-infection as a starting point. We observed an overall trend towards decreased pro-inflammatory cytokine production by NHBEs stimulated with transiently restricted NTHI compared to continuously exposed NTHI (Figure 4.1). Further, within the panel of cytokines assayed, the level of Granulocyte Macrophage- Colony Stimulating Factor (GM-CSF) in spent supernatants of NHBEs stimulated with transiently restricted NTHI was significantly decreased compared to NHBEs stimulated with continuously exposed NTHI (Figure 4.1).

GM-CSF is a pro-inflammatory cytokine with primary functions that include regulating the survival, proliferation, and differentiation of granulocyte-macrophage populations (458).

Although not significant, we also observed a notable decrease in the production of

Granulocyte-Colony Stimulating Factor (G-CSF), IL-1A, and IL-6 by NHBEs in response to stimulation with transiently restricted NTHI compared to continuously exposed NTHI

(Figure 4.1). Taken together, our in vitro data suggest that cytokine production by

131

Figure 4.1. In vitro cytokine production by respiratory epithelial cells stimulated with NTHI transiently restricted or continuously exposed to heme-iron. NHBE cells were grown to confluency on transwell membranes and stimulated with transiently restricted (TR) or continuously exposed (CE) NTHI at an MOI of 25:1. Cytokines present in the spent cell culture supernatants at 24 hours were assayed and the fold change was calculated relative to uninfected cells. Data represent two biological replicates that each included triplicate wells per condition. Statistical significance for each cytokine was determined by Student’s t-test. *, P < 0.05.

132 epithelial cells is dampened following infection with transiently restricted NTHI compared to continuously exposed NTHI.

Pilot investigation of cytokine responses to nutritionally conditioned NTHI in vivo during experimental OM

Several inflammatory cytokines have been identified in middle ear effusions collected from children with OM, illustrating the utility of analyzing effusions for the presence of inflammatory mediators (459, 460). Our observation of a trend toward lower cytokine production by epithelial cells in vitro in response to transiently restricted NTHI led us to investigate the cytokine responses in vivo in the middle ear during experimental

OM. Cohorts of chinchillas were infected transbullarly with either transiently restricted or continuously exposed NTHI and monitored for a total infection period of 14 days. At days

4, 7, 10, and 14, middle ear fluid was collected by epitympanic taps. There was no significant difference in the bacterial burden of transiently restricted or continuously exposed NTHI over the course of infection, although by 10 days post-infection one chinchilla in each cohort reached end-point criteria and was euthanized (Figure 4.2A and

Table 4.1). Therefore, we chose to analyze cytokine levels in the middle ear fluids collected at seven days post-inoculation as sufficient sample volume was recovered from each animal at this time point.

As with our in vitro study, we observed a slight trend towards decreased levels of cytokines in effusions collected from animals infected with transiently restricted NTHI compared to continuously exposed NTHI; however, none of the differences in the cytokines assayed were statistically significant (Figure 4.2B). Further, in contrast to our in vitro data, levels of GM-CSF in middle ear fluid appeared similar between ears infected with transiently restricted and continuously exposed NTHI (Figure 4.2B). 133

Figure 4.2. Transiently restricted or continuously exposed NTHI cytokine stimulation in the chinchilla middle ear. (A) Bacterial burden of chinchilla middle ear fluids (effusions) collected over the time course of infection with either transiently restricted (TR) or continuously exposed (CE) NTHI. Fluids were collected by epitympanic tap, serially diluted, and plated for viable CFU. (B) Data represents the relative amounts of the listed inflammatory cytokines present in the chinchilla middle ear fluid as determined by ELISArray. Mean + SE of fluids collected from three independent animals at seven days post-inoculation with transiently restricted or continuously exposed NTHI.

134

Therefore, GM-CSF may be more important for early responses to NTHI stimulation.

Further investigation of levels of pro-inflammatory cytokines produced by epithelial cells in response to nutritionally conditioned NTHI across a broad range of time points both in vitro and in vivo will be warranted. Additionally, larger scale studies to investigate in vivo cytokine responses to transiently restricted or continuously exposed NTHI will require larger cohorts of animals to account for inability to recover effusion across the time course.

4.4 Discussion

Increasing our understanding of how microenvironmental cues impact host-pathogen interactions will provide a platform for development of innovative therapies to address chronic or recurrent OM. The host inflammatory response to invading pathogens in the middle ear, when prolonged under chronic or recurrent disease conditions, can lead to tissue damage and adverse patient outcomes, such as hearing loss. To survive and cause OM, NTHI must be able to sense and adapt to complex host environments and fluctuations in nutrient (i.e. heme-iron) availability. Transient heme-iron restriction of

NTHI triggers numerous pathogenic changes which include intracellular persistence and altered disease severity in NTHI-mediated experimental OM in the chinchilla model

(Chapter 2,3; 60). Disease progression is mediated by host innate immune factors such as cytokines, chemokines, and responding myeloid cells. Therefore, defining the expression profile of cytokines and other inflammatory mediators in response to infection with transiently restricted NTHI will be an important first step in revealing mechanisms of

NTHI persistence during chronic or recurrent OM. Data presented in this chapter suggest a dampened cytokine response by host epithelial cells in the presence of transiently restricted NTHI compared to continuously exposed NTHI, both in vitro and in 135 vivo. Further, we observed a significant decrease in GM-CSF levels in spent epithelial cell culture supernatant following infection with transiently restricted NTHI compared to continuously exposed NTHI (Figure 4.1). These studies, taken together with the findings in Chapters 2 and 3, narrow the technological approach toward determining changes in immune responses that may mediate the observed attenuation of disease severity in response to transiently restricted NTHI (60).

GM-CSF plays an important role in the host response to infectious agents by regulating leukocyte function (461). Specifically, GM-CSF promotes survival, migration, and differentiation of leukocytes (458). In studies of P. aeruginosa and S. aureus- mediated exacerbations of CF, epithelial cell production of GM-CSF in response to these microbes significantly enhanced the survival of polymorphonuclear leukocytes (PMNs) at the infection site (462). GM-CSF also modulates the inflammatory response in the middle ear; neutralization of GM-CSF inhibits endotoxin-mediated inflammation during experimental OM (463). We observed significantly decreased levels of GM-CSF produced by epithelial cells in vitro at 24 hours post infection with transiently restricted

NTHI compared to continuously exposed NTHI (Figure 4.1). However, we did not observe differences in GM-CSF levels in middle ear fluid at day seven of experimental

OM in the chinchilla (Figure 4.2). It is possible that GM-CSF plays a role in early responses to NTHI-mediated OM, and additional investigation into the levels of GM-CSF in the middle ear fluid and tissue homogenates at early time points of infection will be an important next step. Similarly, a broader range of in vitro time points will need to be investigated to generate a more detailed snapshot of epithelial cell inflammatory cytokine responses and confirm the observed differences in GM-CSF production. As many inflammatory cytokines, GM-CSF included, are necessary for stimulating the migration of immune cells to the infection site, it will also be important to quantify infiltrating myeloid 136 cells in the middle ear fluids collected from transiently restricted or continuously exposed

NTHI-infected animals. We hypothesize that middle ear fluids collected from transiently restricted NTHI-infected animals will have significantly decreased levels of immune cell infiltrate compared to continuously exposed NTHI-infected animals.

Overall, the data presented in this chapter further our understanding of the inflammatory response to nutritionally conditioned NTHI. These pilot studies will provide an important starting point for future studies to define both pro- and anti-inflammatory responses to nutritionally conditioned NTHI by host epithelial cells. Future investigations will also focus on understanding possible mechanisms underlying transiently restricted

NTHI dampening of the immune response, such as changes in epithelial cell signaling cascades in response to NTHI.

137

Chapter 5. General Discussion

5.1 Research Findings

Advances in the fields of microbiology and immunology have significantly progressed our understanding of bacterial pathogenesis, yet there remains a critical need for development of effective therapies and vaccine candidates for many infectious diseases. Chronic and recurring episodes of NTHI-mediated OM continue to be a significant socioeconomic burden to both patients and caregivers. Further, healthcare professionals must continually make decisions on antibiotic use or surgical interventions.

As a result, otitis media remains a major driver of the over-prescription and/or misuse of antibiotics and directly contributes to the current global crisis of antibiotic-resistant bacteria.

Achieving the goal of preventing chronic and recurring episodes of otitis media requires that we shift our approach to understanding microbial pathogenesis. Recent studies illuminating the adaptive nature of bacterial pathogens during chronic disease make it clear that we cannot focus solely on bacterial mechanisms of virulence, just as we similarly cannot focus solely on the host response to disease. The development of novel therapeutics that are effective at targeting bacteria in chronic infections requires that we appreciate the marriage of bacterial and host responses – that ultimately, the nature of infectious disease is a call-and-response between a host defense strategy and bacterial countermeasures for survival in a hostile environment. Microevolution of a

138 pathogen for adaptation to changing environments requires that we as researchers also evolve our thinking; a pathogen which has “historically” been considered to act or respond one way, in one disease setting, should not be assumed to act the same under even slightly different conditions.

The goal of the project presented in this dissertation was to use this approach to investigate the role of nutritional immunity in NTHI pathogenesis and persistence. As

NTHI is auxotrophic for heme-iron, a crucial element for NTHI survival is the ability to regulate heme-iron acquisition and utilization during the shift from a nasopharyngeal commensal to an opportunistic pathogen in the middle ear. NTHI has multiple heme-iron binding proteins and transporters that aid in obtaining heme-iron from the environment, and Fur regulation of gene expression in response to changing iron levels allows NTHI to respond accordingly to iron limitation (77). In support of this, heme-acquisition genes are significantly upregulated in NTHI strains isolated from the middle ear of children with

OM when compared to NTHI strains isolated from the nasopharynx of healthy children

(393, 394). Transient heme-iron restriction induces phenotypic changes in NTHI, including altered biofilm formation, increased survival in a chinchilla model of otitis media, and the formation of intracellular bacterial communities (60). However, the mechanisms behind how NTHI uses these phenotypic changes to persist in a disease setting, and how we might eventually exploit bacterial adaptation for therapeutic development, remains largely unknown.

Chapter 2 reveals that transient heme-iron restriction of NTHI results in long-term survival in broth culture for at least five weeks. We demonstrate that heme-restriction- 139 induced long-term survival selects for a mutation in icc, a gene that encodes a cAMP phosphodiesterase, and that this mutation increases transformation. This mutation also promotes persistence of NTHI in IBCs in a preclinical model of OM, despite clinical signs of a resolved infection. This work further investigates the nature of IBC formation and reveals, for the first time, that transiently heme-restricted NTHI which are taken up by epithelial cells via macropinocytosis escape or evade the endolysosomal pathway and can survive in the cell cytoplasm (Chapter 3). Additionally, pharmacological inhibition of macropinocytosis significantly decreases IBC formation in chinchilla middle ear epithelial cells. Adaptation of NTHI towards an intracellular lifestyle may alter the response of epithelial cells to the invading pathogen. One such outcome may be altered cytokine secretion and/or signaling. Chapter 4 details preliminary findings that suggest the cytokine response to transiently restricted NTHI is lessened compared to the cytokine response induced by continuously exposed NTHI. As cytokines are mediators of the immune response, dampened cytokine production by epithelial cells in response to transiently restricted NTHI may shift the disease severity of OM and subsequently lessen the inflammatory response to NTHI.

With this work we have increased our understanding of the mechanisms that

NTHI use to persist in chronic or recurring infection, and how NTHI adaptation to changing microenvironments may be hindering our current efforts to design effective therapeutics. The following discussion sections will focus on remaining questions, additional studies, and hypotheses that have arisen due to the knowledge gained from this project.

140

5.2 cAMP and competence: Implications for persistence

This work identified a naturally occurring mutation in icc following transient heme- iron restriction in two independent long-term stationary phase cultures (Chapter 2). The mutations were not identical. The mutation in RM33 produces an amino acid change in the active site of the encoded cAMP phosphodiesterase, while the mutation in AR20 truncates Icc, with only the C-terminal portion of the protein encoded. We would, therefore, predict that, in both isolates, Icc is non-functional. The RM33 mutation indeed renders Icc non-functional, as demonstrated by significantly decreased cAMP phosphodiesterase activity, and this would be predicted to be similar for the AR20 isolate

(Figure 2.4). An important distinction which warrants further investigation is whether long-term survival of transiently restricted NTHI is due to the initial transient heme-iron restriction or the acquisition of a mutation in icc and subsequent decrease in phosphodiesterase activity. As a first step in addressing this question, we explored the ability of nutritionally conditioned RM33 to outcompete or extend the survival of the parental strain under long-term stationary phase culture conditions. If a mutation in icc is sufficient to induce long-term survival, then one would expect that the RM33 strain would exhibit long-term survival in static culture without an initial transient heme-iron restriction.

In support of this hypothesis, we observed that this was the case, as RM33 that had been continuously exposed to heme-iron was viable for at least two weeks, similar to the transiently restricted parental strain 86-028NP (Figure 5.1A). When cultured together, continuously exposed RM33 outcompeted the parental strain that had also been continuously exposed to heme-iron (Figure 5.1C). Similarly, transiently restricted RM33 outcompeted the continuously exposed parental strain (Figure 5.1B). There was no survival advantage conferred to transiently heme-iron restricted RM33 compared to continuously exposed RM33 when cultured together (Figure 5.1D), 141

Figure 5.1 Prior transient heme-iron restriction is not required for induction of long-term survival of RM33 in stationary phase culture. (A) Individual long-term stationary phase cultures of NTHI strain 86-028NP or RM33 that were transiently restricted or continuously exposed to heme-iron. Cultures were serially diluted and plated for viable CFU/mL every 24 hours for 14 days. (B) NTHI strain 86- 028NP/pSPEC1 that was continuously exposed to heme-iron and RM33/pGZRS1.1 transiently restricted from heme-iron were mixed at a 1:1 ratio and cultured in long-term stationary phase. (C) NTHI strain 86- 028NP/pSPEC1 that was continuously exposed to heme-iron and RM33/pGZRS1.1 that was continuously exposed to heme-iron were mixed at a 1:1 ratio and cultured in long-term stationary phase. (D) NTHI strain RM33/pSPEC1 that was continuously exposed to heme-iron and RM33/pGZRS1.1 that was transiently restricted for heme-iron were mixed at a 1:1 ratio and cultured in long-term stationary phase. For (B-D), cultures were serially diluted and plated on chocolate agar plates containing appropriate antibiotics for selection of marked strains every 24 hours to determine viable CFU/mL.

142 illustrating that RM33 does not appear to require an initial period of heme-iron restriction to enter long-term stationary phase. Further experiments utilizing a strain with an engineered deletion of icc in a long-term survival assay without prior heme-iron restriction would provide additional insight into the specific contributions of Icc and/or heme-iron status of NTHI in inducing long-term survival. It is our working hypothesis that the increase in longevity is induced initially by the limitation of heme-iron which promotes the acquisition of, and then selection for, mutations in the icc gene.

These mutations further drive a persistent phenotype. Although the only genetic change that was observed in both parallel, independent experiments was mutation in icc, we cannot rule out the possibility of other potential changes (i.e. epigenetic) occurring in response to transient heme-iron restriction.

As a result of mutation in icc, we observed significantly decreased cAMP phosphodiesterase activity in RM33 compared to the parent strain, suggesting that

RM33 has increased levels of intracellular cAMP. cAMP regulates transformation efficiency in NTHI, and increased cAMP correlates with increased competence (143). A link between competence and bacterial persistence has been observed for several pathogens. Natural competence promotes the persistence of H. pylori in chronic infection, and mutating the H. pylori cytosolic competence factor DprA decreases persistence (464). Similarly, the competence stimulating peptide (CSP)-comDE regulatory circuit of Streptococcus mutans is directly involved in persister development

(465). Our observation that the persistent NTHI isolate RM33 displays significantly increased transformation efficiency in the absence of exogenous cAMP (Figure 2.4) suggests a role for DNA uptake and transformation in NTHI persistence during periods of nutrient stress. Thus, we speculate that an increased transformation efficiency would provide an advantage for NTHI in the context of nutritional immunity during OM. We 143 previously showed that the amino acid valine is significantly decreased in the middle ear environment during acute otitis media (298). Valine inhibits the induction of competence in NTHI (466), so a diminution of valine in the middle ear during the early stages of infection suggest an environment that may be favorable for transformation and, thus, may provide a survival advantage for NTHI.

5.3 Proteomic insights into RM33 persistence

Our observation that transient heme-iron restriction of NTHI induces long-term survival in stationary phase revealed an in vitro model for studying bacterial factors that influence NTHI persistence. Using this model, we are now able to generate additional libraries of persistent isolates from NTHI grown in long-term culture induced by transient heme-iron restriction. These libraries of persistent NTHI isolates collected in vitro will be useful for phenotypic studies into how NTHI isolates adapted to a nutrient-starved environment interact with other pathogens or host cells.

One important phenotype that will be important to characterize is biofilm lifestyle of the persistent strain compared to the parental strain. Previous studies have demonstrated that biofilms play a significant role in bacterial persistence in chronic disease, and NTHI forms biofilms in the middle ears of chinchillas during OM (209, 221,

226). Additionally, we have already shown striking changes in the architecture and morphology of NTHI biofilms due to transient heme-iron restriction; specifically, we observed that transiently heme-restricted NTHI biofilms formed tall, filamentous towers while biofilms continuously exposed to heme-iron displayed a mat-like architecture with bacteria retaining their coccoid morphology (60). An important question is whether the phenotypic changes induced by transient heme-restriction are stable and contribute to the survival of the microbe in the long-term. Interestingly, when the RM33 isolate was 144 continuously exposed to heme-iron it still maintained the tower formation of the initial transiently restricted NTHI biofilms compared to the mat-like architecture of the parental strain 86-028NP (Figure 5.2A, B, D). We further demonstrated that this phenotype was maintained in mixed culture, and that there was clear segregation of the parental and

RM33 biofilms evident when using fluorescent reporter strains (Figure 5.2C).

Morphological changes (for example, altered biofilm architecture) can promote bacterial survival and persistence in response to antimicrobial agents or innate immune cells

(359, 467). Therefore, to further our understanding of pathogenic mechanisms that promote NTHI persistence, it is important to determine NTHI proteins that contribute to altered biofilm formation by the persistent RM33 strain.

To identify bacterial factors that may be contributing to the persistence and distinct phenotypes of transiently restricted NTHI, we generated proteomic profiles of biofilms formed by both the RM33 and parental strain (86-028NP) using LC/MS-MS. We identified 751 proteins with a peptide count greater than one. Proteins with significant changes in amount between the parent and RM33 biofilms are listed in Table A.1.

Compared to biofilms formed by the parent strain, the amounts of 27 of these proteins were significantly increased in biofilms formed by RM33. Two proteins were significantly decreased in RM33 compared to the parental biofilms. As expected based on our observation of increased transformation efficiency in RM33, 68% of the significantly increased proteins in the RM33 biofilm compared to the parental biofilm have roles in competence. The Haemophilus competence regulon is co-regulated by cAMP-binding protein CRP and the Sxy protein (143). CRP activates gene transcription when levels of cAMP are high, so we speculate that an inactivating mutation in icc may result in increased amounts of cellular cAMP and, thus, constitutive expression of CRP-regulated genes. In Haemophilus and E. coli, CRP regulates genes with roles in nutrient uptake 145

Figure 5.2. RM33 forms biofilms with distinct architecture that remain segregated from the parent in mixed culture. NTHI strain 86-028NP/pGM1.1 (expressing GFP) (A) or RM33/pKM1.1 (expressing mCherry) (B) were inoculated in separate wells of a chamberslide and incubated for a total of 48 hours to allow biofilm growth. Biofilms were fixed with 4% paraformaldehyde imaged by fluorescence microscopy. The heights of each biofilm are indicated in the top right corner of each panel. Images are representative of five independent experiments. (C) 86-028NP/pGM1.1 and RM33/pKM1.1 were mixed at a 1:1 ratio and inoculated on chamberslides. Biofilms were fixed at 48 hours post-inoculation and imaged by fluorescence microscopy. Green, 86-028NP; red, RM33. Image is representative of three independent experiments. (D) Heights of single strain biofilms were measured from twenty random fields in five independent experiments. Statistical analysis was performed by Student’s t-test.

146 and utilization (468). Many of the significantly increased proteins identified in RM33 biofilms have roles in bacterial metabolism. Changes in bacterial metabolism are often associated with bacterial persistence, with the ability for a bacterium to shift primary metabolic pathways to use a different intermediate or terminal electron acceptor for energy production, providing a survival advantage in both spent culture (long-term stationary phase) and in the host. Similarly, upregulation of biosynthetic pathways can have additional effects, such as the production of metabolites, that can initiate signaling pathways or interact with other bacterial or host components. Along these lines, we observed significant increases in proteins involved in tryptophan biosynthesis in the

RM33 biofilms when compared to the parent biofilms (Figure 5.3A).

The proteins involved in tryptophan biosynthesis in NTHI strain 86-028NP are encoded by genes in the tryptophan biosynthesis operon (trpABCDEG). Regulation of the tryptophan biosynthesis operon occurs during biofilm formation by other pathogens but with differing outcomes depending on the stage of the biofilm cycle. For example,

Salmonella upregulates expression of genes in the trp operon during late stage biofilm formation (72 hours), whereas expression of the trp operon in E. coli is restricted to early biofilm stages (<7 hours) (469, 470). Our data demonstrate that protein products of the trp operon increase in later stage RM33 biofilm (48 hours) in vitro, but the production of tryptophan biosynthesis enzymes in early stages of NTHI biofilm formation remains to be determined.

Once synthesized, or imported into the bacterial cell, tryptophan can be catabolized by tryptophanase, encoded by tnaA, to produce ammonia, pyruvate, and indole. Indole is an aromatic metabolite with diverse functions and roles in signaling, biofilm formation, and persistence (154). Interestingly, at 24 hours a quantitative indole assay revealed that supernatants of biofilms formed by RM33 contained less indole than 147

Figure 5.3. Enzymes in the tryptophan biosynthesis pathway are significantly increased in biofilms formed by RM33 compared to biofilms formed by the parent strain. Quantitative liquid chromatography – tandem mass spectrometry (LC-MS/MS) was performed on triplicate samples of 48 hour biofilms formed by RM33 or the parent. (A) Enzymes of the tryptophan biosynthesis pathway are significantly increased in biofilms formed by RM33 compared to biofilms formed by the parent. Enzymes are listed in gray, and * denotes statistical significance (RM33 compared to the parent). Tryptophan conversion to indole is catalyzed by TrpAB. (B) Indole concentrations in the supernatant of biofilms formed by 86-028NP or RM33 after 24 or 48 hours of growth. Indole concentration was determined by modified Kovac’s assay. Biofilms formed by RM33 produce lower concentrations of indole compared to the parent biofilms, despite increased production of tryptophan biosynthesis enzymes in the RM33 biofilms. (C) Concentration of indole that remains biofilm-associated after 48 hours of biofilm growth. RM33 or 86- 028NP biofilms (after removal of supernatant) were lysed by high-pressure cell and the lysates were assayed for indole by modified Kovac’s assay.

148 that of the parent (Figure 5.3B). This was intriguing as we expected to observe significantly increased indole in the supernatant because of the increased production of components of the tryptophan biosynthesis pathway in RM33 biofilms. One possible reason for this disparity is that the RM33 strain does not export measurable amounts of indole to the supernatant. To investigate this, we performed a quantitative assay to measure the amount of indole that remained associated with the biofilm as opposed to being released into the supernatant. This assay revealed that RM33 is producing some indole that remains in the biofilm itself, as opposed to the parental biofilms, which did not have any measurable indole within the biofilm (Figure 5.3C). Overall, though, most of the indole produced by both biofilm populations was in the supernatant, suggesting that indole is indeed either being actively exported or diffusing out of the bacterial cell in both

RM33 and the parent biofilms (Figure 5.3B). A second possible hypothesis for the disparity in indole production between the parent and RM33 biofilms is that RM33 is not catabolizing tryptophan at the same rate as the parental biofilms, despite producing significantly increased amounts of tryptophan biosynthesis enzymes. Related to this hypothesis, we did not observe a significant increase in the amount of the tryptophanase

TnaA in the RM33 biofilms. Therefore, it is likely that the bacteria are increasing the products of the trp operon not primarily for breakdown of tryptophan into indole, but rather to use tryptophan for an alternative purpose; for example, as a nutrient source or incorporation into a different bioactive pathway as an amino acid.

Enzymes with roles in tryptophan biosynthesis are increased in RM33 biofilms compared to the parent, yet significantly less indole is produced when compared to the parental biofilms during early biofilm formation. Indole signaling has been linked to bacterial persistence via interplay with bacterial cAMP levels. RM33 contains a non- functional cAMP phosphodiesterase as a result a mutation in icc, so the observation that 149 levels of both indole and cAMP are altered in RM33 compared to the parent strain is intriguing. The role of indole and cAMP in inducing bacterial persistence occurs through several distinct and contrasting mechanisms. For example, indole signaling induced persistence of E. coli, and deletion of tnaA, the gene encoding tryptophanase, decreased persistence (412). In contrast to these findings, a second study observed that a toxin/antitoxin system in E. coli contributed to persistence by reducing tryptophanase activity with a concomitant decrease in the production of indole (416). Similarly, a cAMP phosphodiesterase, DosP, increased E. coli persistence by decreasing tryptophanase activity and subsequently decreasing the concentration of indole (427). Our results allow us to propose a unique model for the role of indole in NTHI persistence in which an inactivating mutation in the gene encoding cAMP phosphodiesterase, Icc, is associated with increased production of tryptophan biosynthesis enzymes, yet decreased indole production compared to the parental biofilms.

Our finding that, during biofilm formation, the persistent isolate RM33 increases production of the tryptophan biosynthesis pathway, but not tryptophanase, is intriguing.

Expression of tnaA in E. coli is known to be dually regulated by cAMP-CRP and tryptophan-induced transcription antitermination (471). Therefore, it is possible that tryptophan biosynthesis and catabolism may be similarly regulated by cAMP-CRP in

NTHI. Future studies will focus on investigating the role of tryptophan synthesis and catabolism in NTHI persistence, particularly in the context of indole signaling and biofilm formation. We have generated tnaA and trpAB single and double deletion mutants in

NTHI 86-028NP for this purpose. It will be of interest to use these mutants to determine if the tryptophan biosynthesis and tryptophanase enzymes contribute not only to biofilm formation by RM33, but to initial induction of persistence in the long-term stationary phase culture model, or survival of NTHI in vivo in the chinchilla middle ear. 150

Overall, our data provide proteomic insight into mechanisms of biofilm formation by RM33. We identified significant increases in proteins involved in both competence and tryptophan biosynthesis in the RM33 biofilms when compared to the parental strains. Use of our proteomics data, in conjunction with metabolomic profiles of RM33 and 86-028NP biofilms, will be useful to further our understanding of biofilm formation and persistence of NTHI following adaptation to transient heme-iron restriction.

5.4 Adaptation to an intracellular lifestyle

A central focus of this dissertation was to investigate the mechanisms underlying

IBC formation by transiently heme-restricted NTHI. Our prior studies described a role for transient heme-iron restriction in promoting the formation of IBCs that appear to fill the volume of the cytoplasm in chinchilla middle ear epithelial cells in vitro (60). We have now expanded on our original observation and demonstrated a role for NTHI IBCs both in an additional in vitro cell model, human respiratory epithelium, and in vivo within the chinchilla middle ear mucosae during acute and chronic otitis media (Chapters 2 and 3).

The observation that NTHI can form IBCs in epithelial cells immediately following transient heme-iron restriction, and do not require adaptation to a long-term stationary phase culture or acquisition of a mutation in icc, suggests that IBC formation is not driven solely by genetic mutation or decreased phosphodiesterase activity. However, we observed that the icc mutation did confer an advantage in vivo as RM33 formed significantly increased numbers of IBCs as compared to the parental strain as early as seven days post-inoculation. The early establishment of an intracellular niche for RM33 during acute OM may provide a survival advantage that extends into chronic OM, as we observed that the RM33 IBCs persisted to at least 28 days post-infection (Figure 2.6).

151

Successful adaptation of NTHI to an intracellular lifestyle means that the pathogen can actively subvert (or simply have mechanisms for surviving) host endocytic processes both at initial entry and during intracellular trafficking. The remainder of this section will discuss areas of future investigation into the process of NTHI invasion and escape into the cytosol. Chapter 3 detailed our observations that IBC formation is driven by escape or evasion of the endolysosomal pathway. Transiently restricted NTHI avoided colocalization with both endosomes and lysosomes and established a niche in the host cytosol. In direct contrast, continuously exposed NTHI was observed in both endosomes and lysosomes and displayed significantly decreased intracellular survival compared to transiently restricted NTHI (Figure 3.2, 3.6, 3.7). Further, we demonstrated that, while transiently restricted NTHI use multiple points of entry into epithelial cells, escape from host trafficking pathways and resulting IBC formation is likely spawned from the subpopulation of NTHI which enter through the non-specific macropinocytosis pathway (Figure 3.11).

Macropinocytosis is subverted by multiple intracellular pathogens, including Shigella, Salmonella, Legionella, and Mycobacteria (472-475). These pathogens can enter host cells by macropinocytosis and then either survive within a modified macropinosome or escape into the cytosol (314). Macropinocytosis has also been recently identified as a major route of entry for a number of both DNA and RNA viruses, including viruses of the upper respiratory tract such as HRV 8 and 14, IAV, and

RSV (476, 477). In 90-95% of OM cases, bacterial OM is preceded by a viral URI

(28). This presents an interesting question; whether a prior viral URI would prime subsequent bacterial entry through macropinocytosis, especially as some viruses can specifically induce continuous macropinocytosis in host cells (478, 479).

152

In a similar vein, it will be important to determine whether transiently restricted

NTHI itself is inducing macropinocytosis in respiratory epithelial cells for uptake via this endocytic pathway. Of the bacterial pathogens that have the ability to hijack macropinocytosis, many can induce macropinosome formation through virulence factors and toxins secreted by Type III/IV secretion systems (314). However, there is evidence that some bacteria can induce macropinocytosis without the use of a

Type III/IV secretion system. For example, spent Mycobacterium tuberculosis culture medium is sufficient to induce macropinosome formation in epithelial cells, indicating that bacterially secreted proteins can trigger macropinocytosis (475). Similarly, NTHI must be viable to initiate cytoskeletal rearrangements in airway epithelium, suggesting that NTHI also produces secreted factor(s) that stimulate membrane ruffling (171).

Ongoing and future studies will be focused on determining whether transiently restricted NTHI induces macropinocytosis in respiratory epithelium. However, the nutritional status of NTHI (transiently restricted for heme-iron versus continuously exposed) did not produce a significant difference in the amount of viable bacteria taken up by macropinocytosis when this pathway was inhibited by the addition of EIPA (Figure

3.10). One possible reason for this observation is that NTHI’s nutritional status produces no difference in induction of macropinocytosis in epithelial cells, but rather that the subpopulation of transiently restricted NTHI which are taken up by macropinocytosis may be primed for increased survival within or escape from macropinosomes as compared to the continuously exposed NTHI. To determine if transiently restricted or continuously exposed NTHI differ in their ability to survive within or escape from macropinosomes, it will be important to assess production and localization of host cell components involved in membrane ruffling and formation of the

153 macropinosome during infection. This will also provide additional insight into how macropinocytosis may be promoting IBC formation.

Macropinosome formation is driven by membrane ruffling in response to actin polymerization near the plasma membrane, initiated by Rho GTPases, Cdc42, and Rac

(314). These proteins facilitate coordination of the actin branch nucleating Arp2/3 complex. The plasma membrane then extends into ruffles which eventually fuse back into the membrane itself. Additional regulation of macropinosome formation occurs through p21 activating kinases (PAKs), and sorting nexins (SNXs) (314). Specifically,

SNX1, 5, 18, and 33 directly correlate with macropinosome frequency, as overexpression of these SNXs significantly increases the number of macropinosomes formed per cell (480). Our previous proteomic studies investigating the host response to

NTHI-mediated acute OM demonstrated that infection with NTHI strain 86-028NP induces a significant increase in Arp2/3 in chinchilla middle ear mucosal tissue (298).

Further, a recent study that defined the transcriptome of an episode of NTHI-mediated acute OM in a murine model revealed that NTHI infection stimulates significant increases in the transcripts of genes encoding Arp2/3, a number of Rho-GTPases and

Racs, Cdc42, and SNX18 (230). Taken together, these studies suggest a host microenvironment that may be favorable for NTHI entry into host cells by actin reorganization and macropinocytosis.

This work utilized pharmacological inhibition of macropinocytosis to examine IBC formation, but the use of macropinocytosis stimulators such as protein kinase C, phorbol esters, small G proteins, or growth factors in parallel studies would provide additional evidence for the role for macropinocytosis in IBC formation (481). The fact that growth factors can induce macropinocytosis raises the question of whether infection affects epithelial cell growth factor production, and how this may change the rate of 154 macropinocytosis in neighboring cells. Our data suggest that transiently restricted NTHI do in fact alter epithelial cell production of some cytokines and growth factors

(specifically colony stimulating factors) (Chapter 4). Colony simulating factors play important roles in regulating innate immune cell migration, proliferation, and survival

(458). Further, colony stimulating factors have also been shown to induce macropinocytosis through stimulation of the PI3K signaling cascade (482). UPEC suppresses cytokine production by epithelial cells, a process which aids the initial invasion and early IBC formation by UPEC (483). Therefore, we speculate that modulation of epithelial cytokine and growth factor production may aid initial NTHI entry and early IBC formation in neighboring cells, either through protection from immune cell infiltrate to allow time for invasion, or possibly by inducing macropinocytosis for initial entry.

The establishment of a niche in the cytosol of host cells requires vacuolar escape or evasion of the endolysosomal pathway. Successful intracellular pathogens such as Shigella, Listeria, or Salmonella employ a variety of mechanisms to either modify or escape endocytic vesicles. Both Shigella flexneri and L. monocytogenes secrete lytic proteins to rapidly escape from early vacuoles into the cytosol and thus do not co- localize with early or late endocytic proteins (484, 485).

Salmonella enterica serovar Typhimurium secretes effector proteins to modify the acidity of host endosomal compartments and promote survival in a Salmonella-containing- vacuole (SCV) enriched with endosomal features such as EEA1 (486). NTHI-containing vacuoles (NTHI-CV) decorated with either EEA1 or LAMP1 were identified within epithelial cells (303). This observation is similar to our characterization of continuously exposed NTHI co-localizing with EEA1 and LAMP1 markers. In contrast, we identified a large proportion of transiently restricted NTHI that do not co-localize with EEA1 at 4 155 hours, which provides evidence that this population is likely escaping from an early vacuole and avoids being trafficked to the degradative environment of the endolysosome (Figure 3.6). The NTHI genome does not encode type III or type IV secretion systems (103), which suggests that transiently restricted NTHI must escape from the endolysosomal pathway using an unknown mechanism that is independent of known secreted effector proteins. Also, we observed populations of transiently restricted

NTHI which appear to fill the volume of cells (Figures 3.1, 3.2), so we hypothesize that these populations are not contained within a membrane.

We observed significant changes in the distribution of LAMP1 throughout the cell following infection with continuously exposed NTHI as compared to cells infected with transiently restricted NTHI (Figure 3.7). This suggests that lysosomal biogenesis, and possibly activity, is altered between the two populations. Continuously exposed NTHI induced a perinuclear localization of lysosomes, while the transiently restricted NTHI induced a more dispersed localization of lysosomes in infected cells. Perinuclear localization of lysosomes has been associated with Rab protein activity and an increased probability of autophagosome formation when compared to cells with dispersed lysosomes (487-489). Therefore, there could be differences in the autophagic response of cells towards transiently restricted NTHI compared to continuously exposed

NTHI. The autophagy pathway can be subverted by intracellular pathogens to prevent trafficking to the lysosome (490, 491). However, the role of autophagy in NTHI invasion and trafficking is not well understood. NTHI induces autophagy in HEp-2 cells (492) but studies of autophagy in other epithelial cell lines did not reveal NTHI co-localization with the autophagy marker LC3 (303). Therefore, it will be of interest in future studies to examine interactions of transiently restricted NTHI with autophagosomes as a potential mechanism for evasion of the endolysosomal pathway prior to IBC formation. 156

5.5 NTHI as a cytosolic pathogen

It will be an important evolution of this project to determine both the bacterial factors and nutritional requirements for NTHI survival in the cytosol. It is evident that prior heme-iron restriction primes NTHI to escape or evade the endolysosomal pathway in epithelial cells, but it is unknown how NTHI is able to acquire iron or other essential nutrients in the epithelial cell cytosol. Surprising little is known about the microenvironment of the mammalian cell. The cell cytosol has low levels of magnesium, sodium, and calcium ions, and intracellular iron is bound to heme, ferritin, transferrin, or iron-sulfur cluster-containing enzymes. Theoretically, the cell cytosol is a nutrient-rich environment with carbon and nitrogen sources. However, a key study provided evidence that the mammalian cytoplasm is likely nutrient restrictive, and that metabolic adaptation to an intracellular environment, either through prior nutrient stress or the act of entering and escaping the endolyosomal pathway, is a requirement for survival in the cytosol.

Microinjection of several different bacterial species directly into the mammalian cell cytosol revealed that only those species that are adapted to life in the intracellular environment (such as Listeria) can survive (493). Current knowledge of the nutritional requirements for cytosolic growth of intracellular pathogens have come largely from studies utilizing auxotrophic mutants or comparative gene expression profiles of intracellular bacteria with broth-grown bacteria (341, 494). There have been several recent observations that contributed to our knowledge of the nutritional requirements for intracellular survival and replication of other cytosolic pathogens. Rickettsia utilizes host- derived pyruvate as a carbon source and requires serine, proline, and glycine for intracellular replication (495-497). S. flexneri requires guanine, thymine, and aromatic amino acids (498-500). Listeria requires host-derived lipoyl peptides for cellular respiration and host-derived hexose phosphates as a carbon source (450, 451, 501, 157

502). Similarly, B. pseudomallei requires host-derived lipoate for cellular respiration, and additional requirements include purine, histidine, and p-aminobenzoaic acid synthesis for intracellular survival (452). Overall themes from these investigations reveal the importance of amino acid biosynthetic pathways, as well as the ability to efficiently acquire and utilize host-derived nutrients. Successful intracellular pathogens may use both de novo synthesis pathways in addition to scavenging from the environment to conserve energy. For example, Listeria significantly upregulates genes involved in carbon and nitrogen metabolism during intracellular replication in addition to upregulating genes involved in the transport of nutrients across the bacterial membrane

(503).

A clear opportunity for advancing our knowledge of the mechanisms of NTHI survival in the cytosol lies in future studies to define the gene expression, proteome, and metabolome profiles of NTHI existing as IBCs in epithelial cells. A recent study demonstrated the utility of dual RNA-Seq to simultaneously monitor transcriptional changes of both NTHI and the host epithelial cell environment throughout 72 hours of in vitro infection, and revealed that NTHI downregulates central metabolic pathways in addition to increasing production of membrane-bound transporter proteins (201).

Although this study provided great insight into NTHI-epithelial cell interactions, it comes with the caveat that the methodology did not allow for isolation of intracellular bacterial populations to determine gene expression profiles specific to NTHI within host cells.

Further, our data indicate that growth of NTHI in nutrient-replete medium prior to co- culture with epithelial cells (the growth conditions used in the dual RNA-Seq study) does not induce intracellular bacterial community formation, and instead results in the trafficking of intracellular NTHI through the endolysosomal pathway for degradation.

Additional RNA-Seq and proteomics studies that focus specifically on the intracellular 158 populations of NTHI, using our model system of prior transient heme-iron restriction to induce IBC formation by NTHI, will allow further insight into the bacterial factors and metabolic adaptations required for NTHI growth and survival as a cytosolic pathogen.

As we began to design and implement these studies, we took a different, but required, first step towards understanding how transient heme-iron restriction promotes a cytosolic lifestyle for NTHI by comparing the proteomes of the transiently restricted and continuously exposed NTHI inocula that were used in our studies. We hypothesized that transient heme-iron restriction may drive a unique protein profile for NTHI that would provide insight into possible survival advantages for NTHI even before entry and adaptation to an intracellular niche. As shown in Chapter 3, we identified 15 proteins whose amounts were significantly altered in the transiently restricted population compared to the continuously exposed population (Table 3.2). Several of the proteins identified are encoded by genes that are Fur regulated, an expected finding in a population of NTHI that has been heme-iron restricted (77). With regards to proteins that were significantly decreased in the transiently restricted NTHI population compared to the continuously exposed NTHI, most of the proteins identified either contain or are a subunit of an enzyme that requires iron-sulfur clusters for enzymatic activity (for example, FrdAB, DmsA, NrfA) (504-507). Further, NrfA, a subunit of an ammonia- forming nitrite reductase, is a tetraheme cytochrome (507). Thus, reduced production of these non-essential, heme-containing proteins during heme-restricted conditions is a logical outcome in order to preserve heme-iron stores. Focusing on the proteins that significantly increased in amount in the transiently restricted population compared to the continuously exposed population, interesting roles may exist for the tellurite resistance protein TehB, lipoate-protein ligase B LipB (discussed in Chapter 3), and the imidazole glycerol phosphate synthase subunit HisF in priming NTHI for intracellular survival and 159 adaptation. Of note, we also observed a significant increase in transferrin binding protein

2, Tbp2, in the transiently restricted NTHI population. As iron in mammalian cells is sequestered bound to ferritin or transferrin, this may indicate an advantage for transiently restricted NTHI to already have increased levels of Tbp2 prior to entering the cell.

Our observation that the amount of TehB increases following heme-iron restriction corroborates earlier studies that showed a significant increase in the expression of tehB in NTHI 86-028NP, via a Fur-independent mechanism (508).

Characterization of TehB in Haemophilus demonstrated that it provides resistance to oxidative stress, facilitates heme utilization, and is required for virulence in a rat model of bacteremia (508). Interestingly, genes involved in tellurite resistance are associated with intracellular survival in multiple species. There is a direct correlation between the expression of a tellurite resistance gene and the ability for Corynebacterium diptheriae, the causative agent of diphtheria, to survive within intracytoplasmic compartments of epithelial cells (509). Further, Yersinia pestis expresses tellurite resistance proteins, including TehB, during its intracellular lifecycle in macrophages (510). Tellurite is a metalloid that is toxic to gram-negative bacteria and is found in the environment; it is, therefore, unlikely that NTHI, which grows only in a human host, would naturally be exposed to tellurite. It is thus likely that TehB in Haemophilus has a primary function separate from tellurite resistance. It is proposed that TehB in Haemophilus encodes a

SAM-dependent methyltransferase involved in small molecule modification (508). This is interesting as our proteomic screen also identified two other putative SAM- methyltransferases with significantly increased amounts in transiently restricted NTHI

(QueA and an uncharacterized protein with domains frequently found in methyltransferases) compared to continuously exposed NTHI, suggesting that increased 160 methyltransferase activity may at least initially contribute to differences seen in transiently restricted NTHI intracellular trafficking and survival.

We also observed a significant increase in amount of an enzyme found in the histidine biosynthesis pathway, the imidazole glycerol phosphate synthase HisF.

Histidine is an essential amino acid for bacterial growth, and the histidine biosynthetic pathway is involved in de novo synthesis of nucleotides as well as nitrogen metabolism

(511). The his operon is highly conserved in Haemophilus, and NTHI strain 86-028NP possesses all eight genes in the his operon (512). NTHI’s ability to synthesize histidine has been proposed to confer a survival advantage in the middle ear, and the his operon is significantly more prevalent in NTHI OM strains when compared to commensal strains

(512). However, expression of members of the his operon was decreased when NTHI was co-cultured with bronchial epithelial cells (201). Histidine biosynthesis is one of the most energy-depleting metabolic pathways (513), so downregulation of histidine biosynthesis during NTHI infection may represent the bacterium’s effort to conserve energy. Alternatively, it could indicate access to exogenous histidine through the import of host-derived molecules. It is intriguing that transiently restricted NTHI significantly increases production of only the HisF protein, and not the other members of the biosynthetic pathway, especially as one study found that HisF in particular is required for intracellular replication of B. pseudomallei (452). Further investigation into the role of

HisF in NTHI intracellular lifestyles, and specifically if production of HisF remains increased once NTHI is established in an IBC, will be required.

Finally, in addition to determining the bacterial changes occurring that promote intracellular survival of NTHI in the cytosol, it will be important to investigate host changes occurring because of NTHI residence within epithelial cells. Dual RNA-Seq or proteomics studies that identify changes in host gene expression or protein production in 161 parallel to NTHI changes after IBCs are well established, and in a time course that explores both early (endolysosomal escape) and mid-stages of IBC formation, will be useful in this pursuit. Many intracellular pathogens have evolved mechanisms that alter host signaling networks to ensure an environment that promotes survival. For example, both Rickettsia and Listeria induce a long-lasting activation of NF-κB as a result of bacterial factors produced during cytosolic replication (514, 515). Specifically, this is of great benefit to Rickettsia, as NF-κB activation inhibits apoptosis and thus allows continued replication of Rickettsia in the host cell cytosol (516).

5.6 Microevolution in chronic infection

A major finding of this work is that periods of nutrient limitation seem to prime NTHI for adaptation to diverse lifestyles and persistence in chronic infection. NTHI responds to transient heme-iron restriction by becoming more adept at encountering environmental stressors. In our study, we observed microevolution of NTHI in vitro in the form of genetic mutation of icc, a gene which encodes a cAMP phosphodiesterase. This mutation resulted in decreased cAMP phosphodiesterase activity and increased transformation efficiency. However, selective pressures can drive bacterial adaptation through other mechanisms, including metabolic changes, mucoid switching, induction of a dormant state, or increased antibiotic resistance; all of which can influence the persistence of the bacteria within a specific niche.

Our observation that transient heme-iron restriction of NTHI induced a shift to a long- term stationary phase in two independent experiments, and that in both cases this persistent phenotype was accompanied by a mutation in icc, prompted us to ask if microevolution of NTHI would occur in a host environment during the progression of disease. We hypothesize that the nutrient fluctuations occurring during early colonization 162 and induction of inflammation, in addition to the presence of host immune factors that were not present in our in vitro studies, would result in the acquisition of and selection for genetic mutations in NTHI parental strain over time during an episode of OM. To begin to address this we infected cohorts of chinchillas with the NTHI strain 86-

028NP transbullarly and allowed infection to progress for seven days, at which point

NTHI was recovered from the chinchilla middle ear effusions or tissue homogenates and immediately inoculated into a second cohort of chinchillas. This process was then repeated twice more so that bacteria were sequentially passaged at seven-day intervals through a total of four cohorts of chinchillas.

Whole genome sequencing of the NTHI recovered at the end of the study (day 31, four passages through independent cohorts of chinchillas), revealed selection for 16 mutant strains (Table B.1). The types of mutations ranged from phase variation (both transcriptional and translational) to single nucleotide polymorphisms (intragenic and promoter regions) and a deletion. Phase variable genes identified in our list of mutations included hmw genes and lic genes. Other genes identified encode proteins involved in

LOS biosynthesis and hemoglobin-haptoglobin import. In regard to the latter, we observed mutations in three different hemoglobin-haptoglobin proteins, two of which have previously been observed to acquire mutations in chronic NTHI isolates from

COPD patients (517). Whole genome sequencing of isolated NTHI from intermediate passages will be necessary to reveal 1) when the mutations identified at day 31 first occurred, and 2) if other mutations occurred in earlier passages that did not provide a survival advantage and thus would not be observed at day 31. Interestingly, we observed that increased IBC formation was correlated with passaging; we observed a significantly increased number of IBCs forming in the middle ear after the final passage compared to the first passage (Figure 5.4). Our current findings from this study provide 163

Figure 5.4. NTHI IBC formation in middle ear mucosal tissue significantly increases after four sequential periods of infection in a chinchilla model of OM. NTHI was passaged through four rounds of sequential 7-day infection in the chinchilla middle ear as described in Additional Methods in Appendix C. After each passage (7 day infection), chinchilla middle ears were excised, fixed, embedded in paraffin. Thin sections processed for fluorescence microscopy by immunohistochemistry. NTHI was labeled with anti-OMP and detected with Protein A-Alexa Fluor 488. Quantification of IBCs was performed by visually counting the number of IBCs per thin section from six independent ears per time point.

164 evidence that microevolution of NTHI occurs in response to host selective pressure in the middle ear environment and is associated with an intracellular niche for NTHI as well as survival over multiple rounds of infection.

Our findings are further supported by a recent study which found that environmental pressures within the host often select for mutants of respiratory pathogens during a disease state; specifically, “hypermutable” strains of Haemophilus were more frequently isolated from the CF lung than the healthy lung, and the authors speculated that these mutations conferred a survival advantage to the bacteria (518). A larger study of microevolution of pathogens in the human respiratory tract was recently published by the

Murphy research group (519). This study revealed that a major mechanism for NTHI survival in the host was phase variation through slip strand repeat (SSR) mispairing leading to alterations in gene expression. The number of SSRs was positively correlated with the length of time the NTHI strain persisted in the human host (519). Additionally, the study identified a propensity for single nucleotide polymorphisms in the genes that encode the outer membrane proteins P2, P5, and Hap (519). In addition to NTHI,

Pseudomonas also displays microevolution as a response to the environment encountered in the human respiratory tract, specifically the CF lung (520). Changes in surface antigens, modulation of metabolic pathways, and loss of virulence associated traits of Pseudomonas led to the generation of strains that could persist for longer in the respiratory tract (520, 521). Interestingly, 11 of the 16 patho-adaptive P. aeruginosa mutants evolved to become more invasive in a model of murine chronic airway infection

(522). These studies, taken in consideration alongside the studies presented in this work, illustrate the role of selective pressure from the host environment (through nutritional or innate immunity, microenvironmental changes, or interaction with

165 neighboring microbes) in driving bacterial adaptation which ultimately leads to increased persistence within the host.

5.7 Concluding remarks

Our understanding of NTHI pathogenesis continues to evolve as we make technological and scientific advances in the fields of microbiology and immunology.

However, NTHI remains a major mediator of chronic and recurrent diseases, including

OM and exacerbations in COPD and CF. Further, targeted therapeutics for NTHI- mediated disease and efficacious vaccine candidates have remained elusive. Several mechanisms contribute to NTHI persistence as an otopathogen. Biofilm formation, evasion of host immune factors, and the ability to sense and adapt to fluctuations in nutrient availability enhance NTHI survival during disease progression. The work presented herein provide evidence for novel mechanisms for NTHI survival in the host, driven by an elegant response and adaptation to nutrient fluctuations. Our work demonstrates that transient heme-iron restriction induces genetic and phenotypic changes that promote NTHI competence, a transition to persistence in a long-term stationary phase state, and a novel lifestyle as a cytosolic pathogen in IBCs.

An immediate outstanding question is whether IBCs formed by NTHI represent a protected lifestyle for survival during antibiotic treatment in a clinical setting. We speculate that IBCs formed by NTHI represent a possible reservoir for reinfection that drives recurrent OM. In ongoing studies related to this work, we have observed IBCs persisting in the chinchilla middle ear mucosal tissue up to 70 days post challenge. This is at least 40 days beyond the resolution of disease as measured by clinical observations. Our findings provide the groundwork for additional studies into the role of

IBCs in driving recurrence of OM. 166

The development of targeted therapeutics to treat NTHI-mediated disease will now require that we consider several novel ideas about NTHI pathogenesis. First, we must reconsider the idea that NTHI is classically an extracellular pathogen – multiple lines of evidence now indicate that NTHI can invade and survive in the cytoplasm of host cells as an adaptive mechanism during disease progression. Second, we must consider the possibility that certain nutrient limiting conditions, as would be experienced in the host environment, can induce a persistent state in NTHI. Thirdly, we must recognize that microevolution of NTHI occurs in the host environment. Beginning to consider therapeutic design with these ideas will aid in our understanding of how to develop targeted approaches. Our work identified a role for the cAMP phosphodiesterase Icc in persistence of NTHI associated with increased competence. Continued investigation into the NTHI competence regulon, and how DNA uptake may contribute to NTHI persistence in culture and intracellularly, will provide insight into possible targets.

Further, identifying the required genes for NTHI escape from the endolysosomal pathway, and subsequent survival in the cytoplasm will be a necessary first step.

Avenues for targeting the intracellular population of NTHI should also be explored, including antibacterial small molecules that could be delivered into host cells.

Additionally, the synergistic use of inhibitors of NTHI entry into host cells, such as macropinocytosis, with current antibiotic regimens, could also be investigated.

Although many questions about the molecular mechanisms of NTHI persistence remain, our work provides an important step in advancing our knowledge of how NTHI uniquely adapts to changing environments to promote survival. Building on our findings to continue studies of NTHI pathogenesis will move us closer to the eradication of chronic and recurring OM.

167

REFERENCES

1. Musher DM. Haemophilus Species. In: th, Baron S, editors. Medical Microbiology. Galveston (TX)1996.

2. Wilfert CM. Epidemiology of Haemophilus influenzae type b infections. Pediatrics. 1990;85(4 Pt 2):631-5.

3. MacNeil JR, Cohn AC, Farley M, Mair R, Baumbach J, Bennett N, et al. Current epidemiology and trends in invasive Haemophilus influenzae disease--United States, 1989-2008. Clin Infect Dis. 2011;53(12):1230-6.

4. Agrawal A, Murphy TF. Haemophilus influenzae infections in the H. influenzae type b conjugate vaccine era. J Clin Microbiol. 2011;49(11):3728-32.

5. Boisvert AA, Moore D. Invasive disease due to Haemophilus influenzae type A in children in Canada's north: A priority for prevention. Can J Infect Dis Med Microbiol. 2015;26(6):291-2.

6. McConnell A, Tan B, Scheifele D, Halperin S, Vaudry W, Law B, et al. Invasive infections caused by haemophilus influenzae serotypes in twelve Canadian IMPACT centers, 1996-2001. Pediatr Infect Dis J. 2007;26(11):1025-31.

7. Shuel M, Hoang L, Law DK, Tsang R. Invasive Haemophilus influenzae in British Columbia: non-Hib and non-typeable strains causing disease in children and adults. Int J Infect Dis. 2011;15(3):e167-73.

8. Tsang RS, Sill ML, Skinner SJ, Law DK, Zhou J, Wylie J. Characterization of invasive Haemophilus influenzae disease in Manitoba, Canada, 2000-2006: invasive disease due to non-type b strains. Clin Infect Dis. 2007;44(12):1611-4.

9. Bender JM, Cox CM, Mottice S, She RC, Korgenski K, Daly JA, et al. Invasive Haemophilus influenzae disease in Utah children: an 11-year population-based study in the era of conjugate vaccine. Clin Infect Dis. 2010;50(7):e41-6.

10. Dworkin MS, Park L, Borchardt SM. The changing epidemiology of invasive Haemophilus influenzae disease, especially in persons > or = 65 years old. Clin Infect Dis. 2007;44(6):810-6.

11. Resman F, Svensjo T, Unal C, Cronqvist J, Brorson H, Odenholt I, et al. Necrotizing myositis and septic shock caused by Haemophilus influenzae type f in a previously healthy man diagnosed with an IgG3 and a mannose-binding lectin deficiency. Scand J Infect Dis. 2011;43(11-12):972-6.

12. Rubach MP, Bender JM, Mottice S, Hanson K, Weng HY, Korgenski K, et al. Increasing incidence of invasive Haemophilus influenzae disease in adults, Utah, USA. Emerg Infect Dis. 2011;17(9):1645-50.

168

13. Slack MP, Azzopardi HJ, Hargreaves RM, Ramsay ME. Enhanced surveillance of invasive Haemophilus influenzae disease in England, 1990 to 1996: impact of conjugate vaccines. Pediatr Infect Dis J. 1998;17(9 Suppl):S204-7.

14. Bou R, Dominguez A, Fontanals D, Sanfeliu I, Pons I, Renau J, et al. Prevalence of Haemophilus influenzae pharyngeal carriers in the school population of Catalonia. Working Group on invasive disease caused by Haemophilus influenzae. Eur J Epidemiol. 2000;16(6):521-6.

15. Faden H, Duffy L, Williams A, Krystofik DA, Wolf J. Epidemiology of nasopharyngeal colonization with nontypeable Haemophilus influenzae in the first 2 years of life. J Infect Dis. 1995;172(1):132-5.

16. Harabuchi Y, Faden H, Yamanaka N, Duffy L, Wolf J, Krystofik D. Nasopharyngeal colonization with nontypeable Haemophilus influenzae and recurrent otitis media. Tonawanda/Williamsville Pediatrics. J Infect Dis. 1994;170(4):862-6.

17. Puig C, Marti S, Fleites A, Trabazo R, Calatayud L, Linares J, et al. Oropharyngeal colonization by nontypeable Haemophilus influenzae among healthy children attending day care centers. Microb Drug Resist. 2014;20(5):450- 5.

18. Ortiz-Romero MDM, Espejo-Garcia MP, Alfayate-Miguelez S, Ruiz-Lopez FJ, Zapata-Hernandez D, Gonzalez-Pacanowska AJ, et al. Epidemiology of Nasopharyngeal Carriage by Haemophilus influenzae in Healthy Children: A Study in the Mediterranean Coast Region. Pediatr Infect Dis J. 2017;36(10):919- 23.

19. Leach AJ, Boswell JB, Asche V, Nienhuys TG, Mathews JD. Bacterial colonization of the nasopharynx predicts very early onset and persistence of otitis media in Australian aboriginal infants. Pediatr Infect Dis J. 1994;13(11):983-9.

20. Bernstein JM. Immunologic aspects of otitis media. Curr Allergy Asthma Rep. 2002;2(4):309-15.

21. Schilder AG, Chonmaitree T, Cripps AW, Rosenfeld RM, Casselbrant ML, Haggard MP, et al. Otitis media. Nat Rev Dis Primers. 2016;2:16063.

22. Rosa-Olivares J, Porro A, Rodriguez-Varela M, Riefkohl G, Niroomand-Rad I. Otitis Media: To Treat, To Refer, To Do Nothing: A Review for the Practitioner. Pediatr Rev. 2015;36(11):480-6; quiz 7-8.

23. Monasta L, Ronfani L, Marchetti F, Montico M, Vecchi Brumatti L, Bavcar A, et al. Burden of disease caused by otitis media: systematic review and global estimates. PLoS One. 2012;7(4):e36226.

24. Klein JO. Otitis media. Clin Infect Dis. 1994;19(5):823-33.

169

25. Ahmed S, Shapiro NL, Bhattacharyya N. Incremental health care utilization and costs for acute otitis media in children. Laryngoscope. 2014;124(1):301-5.

26. Kaplan B, Wandstrat TL, Cunningham JR. Overall cost in the treatment of otitis media. Pediatr Infect Dis J. 1997;16(2 Suppl):S9-11.

27. Klein JO. The burden of otitis media. Vaccine. 2000;19 Suppl 1:S2-8.

28. Pichichero ME. Otitis media. Pediatr Clin North Am. 2013;60(2):391-407.

29. Pelton SI. Otoscopy for the diagnosis of otitis media. Pediatr Infect Dis J. 1998;17(6):540-3; discussion 80.

30. Sharma SK, Casey JR, Pichichero ME. Reduced memory CD4+ T-cell generation in the circulation of young children may contribute to the otitis-prone condition. J Infect Dis. 2011;204(4):645-53.

31. Sharma SK, Casey JR, Pichichero ME. Reduced serum IgG responses to pneumococcal antigens in otitis-prone children may be due to poor memory B- cell generation. J Infect Dis. 2012;205(8):1225-9.

32. Eskola J, Kilpi T, Palmu A, Jokinen J, Haapakoski J, Herva E, et al. Efficacy of a pneumococcal conjugate vaccine against acute otitis media. N Engl J Med. 2001;344(6):403-9.

33. Morris MC, Pichichero ME. Streptococcus pneumoniae burden and nasopharyngeal inflammation during acute otitis media. Innate Immun. 2017;23(8):667-77.

34. McCormick DP, Chandler SM, Chonmaitree T. Laterality of acute otitis media: different clinical and microbiologic characteristics. Pediatr Infect Dis J. 2007;26(7):583-8.

35. Pumarola F, Mares J, Losada I, Minguella I, Moraga F, Tarrago D, et al. Microbiology of bacteria causing recurrent acute otitis media (AOM) and AOM treatment failure in young children in Spain: shifting pathogens in the post- pneumococcal conjugate vaccination era. Int J Pediatr Otorhinolaryngol. 2013;77(8):1231-6.

36. Heikkinen T, Thint M, Chonmaitree T. Prevalence of various respiratory viruses in the middle ear during acute otitis media. N Engl J Med. 1999;340(4):260-4.

37. Marom T, Nokso-Koivisto J, Chonmaitree T. Viral-bacterial interactions in acute otitis media. Curr Allergy Asthma Rep. 2012;12(6):551-8.

38. Siegel SJ, Roche AM, Weiser JN. Influenza promotes pneumococcal growth during coinfection by providing host sialylated substrates as a nutrient source. Cell Host Microbe. 2014;16(1):55-67.

170

39. Zanin M, Baviskar P, Webster R, Webby R. The Interaction between Respiratory Pathogens and Mucus. Cell Host Microbe. 2016;19(2):159-68.

40. Lieberthal AS, Carroll AE, Chonmaitree T, Ganiats TG, Hoberman A, Jackson MA, et al. The diagnosis and management of acute otitis media. Pediatrics. 2013;131(3):e964-99.

41. Coker TR, Chan LS, Newberry SJ, Limbos MA, Suttorp MJ, Shekelle PG, et al. Diagnosis, microbial epidemiology, and antibiotic treatment of acute otitis media in children: a systematic review. JAMA. 2010;304(19):2161-9.

42. Bodor FF, Marchant CD, Shurin PA, Barenkamp SJ. Bacterial etiology of conjunctivitis-otitis media syndrome. Pediatrics. 1985;76(1):26-8.

43. Van Dort M, Walden C, Walker ES, Reynolds SA, Levy F, Sarubbi FA. An outbreak of infections caused by non-typeable Haemophilus influenzae in an extended care facility. J Hosp Infect. 2007;66(1):59-64.

44. Autio TJ, Koskenkorva T, Narkio M, Leino TK, Koivunen P, Alho OP. Diagnostic accuracy of history and physical examination in bacterial acute rhinosinusitis. Laryngoscope. 2015;125(7):1541-6.

45. Brode SK, Ling SC, Chapman KR. Alpha-1 antitrypsin deficiency: a commonly overlooked cause of lung disease. CMAJ. 2012;184(12):1365-71.

46. Sethi S, Murphy TF. Infection in the pathogenesis and course of chronic obstructive pulmonary disease. N Engl J Med. 2008;359(22):2355-65.

47. Bandi V, Apicella MA, Mason E, Murphy TF, Siddiqi A, Atmar RL, et al. Nontypeable Haemophilus influenzae in the lower respiratory tract of patients with chronic bronchitis. Am J Respir Crit Care Med. 2001;164(11):2114-9.

48. Bandi V, Jakubowycz M, Kinyon C, Mason EO, Atmar RL, Greenberg SB, et al. Infectious exacerbations of chronic obstructive pulmonary disease associated with respiratory viruses and non-typeable Haemophilus influenzae. FEMS Immunol Med Microbiol. 2003;37(1):69-75.

49. Groenewegen KH, Wouters EF. Bacterial infections in patients requiring admission for an acute exacerbation of COPD; a 1-year prospective study. Respir Med. 2003;97(7):770-7.

50. Rosell A, Monso E, Soler N, Torres F, Angrill J, Riise G, et al. Microbiologic determinants of exacerbation in chronic obstructive pulmonary disease. Arch Intern Med. 2005;165(8):891-7.

51. Harrison F. Microbial ecology of the cystic fibrosis lung. Microbiology. 2007;153(Pt 4):917-23.

171

52. Pettigrew MM, Gent JF, Revai K, Patel JA, Chonmaitree T. Microbial interactions during upper respiratory tract infections. Emerg Infect Dis. 2008;14(10):1584-91.

53. Cardines R, Giufre M, Pompilio A, Fiscarelli E, Ricciotti G, Di Bonaventura G, et al. Haemophilus influenzae in children with cystic fibrosis: antimicrobial susceptibility, molecular epidemiology, distribution of adhesins and biofilm formation. Int J Med Microbiol. 2012;302(1):45-52.

54. Murphy TF, Brauer AL, Schiffmacher AT, Sethi S. Persistent colonization by Haemophilus influenzae in chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2004;170(3):266-72.

55. Clementi CF, Hakansson AP, Murphy TF. Internalization and trafficking of nontypeable Haemophilus influenzae in human respiratory epithelial cells and roles of IgA1 proteases for optimal invasion and persistence. Infect Immun. 2014;82(1):433-44.

56. Olszewska-Sosinska O, Zielnik-Jurkiewicz B, Stepinska M, Antos-Bielska M, Lau-Dworak M, Kozlowska K, et al. Persistence of non-typeable Haemophilus Influenzae in the pharynx of children with adenotonsillar hypertrophy after treatment with azithromycin. Pathog Dis. 2016;74(1):ftv106.

57. Gunn JS, Bakaletz LO, Wozniak DJ. What's on the Outside Matters: The Role of the Extracellular Polymeric Substance of Gram-negative Biofilms in Evading Host Immunity and as a Target for Therapeutic Intervention. J Biol Chem. 2016;291(24):12538-46.

58. Clementi CF, Murphy TF. Non-typeable Haemophilus influenzae invasion and persistence in the human respiratory tract. Front Cell Infect Microbiol. 2011;1:1.

59. Raffel FK, Szelestey BR, Beatty WL, Mason KM. The Haemophilus influenzae Sap transporter mediates bacterium-epithelial cell homeostasis. Infect Immun. 2013;81(1):43-54.

60. Szelestey BR, Heimlich DR, Raffel FK, Justice SS, Mason KM. Haemophilus responses to nutritional immunity: epigenetic and morphological contribution to biofilm architecture, invasion, persistence and disease severity. PLoS Pathog. 2013;9(10):e1003709.

61. Cline SD, Saleem S, Daines DA. Regulation of the vapBC-1 toxin-antitoxin locus in nontypeable Haemophilus influenzae. PLoS One. 2012;7(3):e32199.

62. Ren D, Walker AN, Daines DA. Toxin-antitoxin loci vapBC-1 and vapXD contribute to survival and virulence in nontypeable Haemophilus influenzae. BMC Microbiol. 2012;12:263.

63. Ren D, Kordis AA, Sonenshine DE, Daines DA. The ToxAvapA toxin-antitoxin locus contributes to the survival of nontypeable Haemophilus influenzae during infection. PLoS One. 2014;9(3):e91523. 172

64. Hamilton B, Manzella A, Schmidt K, DiMarco V, Butler JS. Analysis of nontypeable Haemophilous influenzae VapC1 mutations reveals structural features required for toxicity and flexibility in the active site. PLoS One. 2014;9(11):e112921.

65. Jin G, Pavelka MS, Jr., Butler JS. Structure-function analysis of VapB4 antitoxin identifies critical features of a minimal VapC4 toxin-binding module. J Bacteriol. 2015;197(7):1197-207.

66. Walling LR, Butler JS. Structural Determinants for Antitoxin Identity and Insulation of Cross Talk between Homologous Toxin-Antitoxin Systems. J Bacteriol. 2016;198(24):3287-95.

67. Mell JC, Shumilina S, Hall IM, Redfield RJ. Transformation of natural genetic variation into Haemophilus influenzae genomes. PLoS Pathog. 2011;7(7):e1002151.

68. Oliver A, Mena A. Bacterial hypermutation in cystic fibrosis, not only for antibiotic resistance. Clin Microbiol Infect. 2010;16(7):798-808.

69. Moxon R, Bayliss C, Hood D. Bacterial contingency loci: the role of simple sequence DNA repeats in bacterial adaptation. Annu Rev Genet. 2006;40:307- 33.

70. Srikhanta YN, Fox KL, Jennings MP. The phasevarion: phase variation of type III DNA methyltransferases controls coordinated switching in multiple genes. Nat Rev Microbiol. 2010;8(3):196-206.

71. Hiltke TJ, Sethi S, Murphy TF. Sequence stability of the gene encoding outer membrane protein P2 of nontypeable Haemophilus influenzae in the human respiratory tract. J Infect Dis. 2002;185(5):627-31.

72. Hiltke TJ, Schiffmacher AT, Dagonese AJ, Sethi S, Murphy TF. Horizontal transfer of the gene encoding outer membrane protein P2 of nontypeable Haemophilus influenzae, in a patient with chronic obstructive pulmonary disease. J Infect Dis. 2003;188(1):114-7.

73. Cholon DM, Cutter D, Richardson SK, Sethi S, Murphy TF, Look DC, et al. Serial isolates of persistent Haemophilus influenzae in patients with chronic obstructive pulmonary disease express diminishing quantities of the HMW1 and HMW2 adhesins. Infect Immun. 2008;76(10):4463-8.

74. Waldron KJ, Rutherford JC, Ford D, Robinson NJ. Metalloproteins and metal sensing. Nature. 2009;460(7257):823-30.

75. Waldron KJ, Robinson NJ. How do bacterial cells ensure that metalloproteins get the correct metal? Nat Rev Microbiol. 2009;7(1):25-35.

173

76. Harrison A, Baker BD, Munson RS, Jr. Overlapping and complementary oxidative stress defense mechanisms in nontypeable Haemophilus influenzae. J Bacteriol. 2015;197(2):277-85.

77. Harrison A, Santana EA, Szelestey BR, Newsom DE, White P, Mason KM. Ferric uptake regulator and its role in the pathogenesis of nontypeable Haemophilus influenzae. Infect Immun. 2013;81(4):1221-33.

78. Gangaidzo IT, Moyo VM, Mvundura E, Aggrey G, Murphree NL, Khumalo H, et al. Association of pulmonary tuberculosis with increased dietary iron. J Infect Dis. 2001;184(7):936-9.

79. Lounis N, Truffot-Pernot C, Bentoucha A, Robert J, Ji B, Grosset J. Efficacies of clarithromycin regimens against Mycobacterium xenopi in mice. Antimicrob Agents Chemother. 2001;45(11):3229-30.

80. Golz A, Netzer A, Goldenberg D, Westerman ST, Westerman LM, Joachims HZ. The association between iron-deficiency anemia and recurrent acute otitis media. Am J Otolaryngol. 2001;22(6):391-4.

81. Griffiths WJ, Kelly AL, Cox TM. Inherited disorders of iron storage and transport. Mol Med Today. 1999;5(10):431-8.

82. Lieu PT, Heiskala M, Peterson PA, Yang Y. The roles of iron in health and disease. Mol Aspects Med. 2001;22(1-2):1-87.

83. Andrews NC. Iron metabolism: iron deficiency and iron overload. Annu Rev Genomics Hum Genet. 2000;1:75-98.

84. Nairz M, Fritsche G, Brunner P, Talasz H, Hantke K, Weiss G. Interferon-gamma limits the availability of iron for intramacrophage Salmonella typhimurium. Eur J Immunol. 2008;38(7):1923-36.

85. Nairz M, Theurl I, Ludwiczek S, Theurl M, Mair SM, Fritsche G, et al. The coordinated regulation of iron homeostasis in murine macrophages limits the availability of iron for intracellular Salmonella typhimurium. Cell Microbiol. 2007;9(9):2126-40.

86. Moeck GS, Coulton JW. TonB-dependent iron acquisition: mechanisms of siderophore-mediated active transport. Mol Microbiol. 1998;28(4):675-81.

87. Chu BC, Garcia-Herrero A, Johanson TH, Krewulak KD, Lau CK, Peacock RS, et al. Siderophore uptake in bacteria and the battle for iron with the host; a bird's eye view. Biometals. 2010;23(4):601-11.

88. Aujla SJ, Chan YR, Zheng M, Fei M, Askew DJ, Pociask DA, et al. IL-22 mediates mucosal host defense against Gram-negative bacterial pneumonia. Nat Med. 2008;14(3):275-81.

174

89. Bachman MA, Miller VL, Weiser JN. Mucosal lipocalin 2 has pro-inflammatory and iron-sequestering effects in response to bacterial enterobactin. PLoS Pathog. 2009;5(10):e1000622.

90. Raffatellu M, George MD, Akiyama Y, Hornsby MJ, Nuccio SP, Paixao TA, et al. Lipocalin-2 resistance confers an advantage to Salmonella enterica serotype Typhimurium for growth and survival in the inflamed intestine. Cell Host Microbe. 2009;5(5):476-86.

91. Cassat JE, Skaar EP. Iron in infection and immunity. Cell Host Microbe. 2013;13(5):509-19.

92. Madejczyk MS, Ballatori N. The iron transporter ferroportin can also function as a manganese exporter. Biochim Biophys Acta. 2012;1818(3):651-7.

93. Yin Z, Jiang H, Lee ES, Ni M, Erikson KM, Milatovic D, et al. Ferroportin is a manganese-responsive protein that decreases manganese cytotoxicity and accumulation. J Neurochem. 2010;112(5):1190-8.

94. Jabado N, Jankowski A, Dougaparsad S, Picard V, Grinstein S, Gros P. Natural resistance to intracellular infections: natural resistance-associated macrophage protein 1 (Nramp1) functions as a pH-dependent manganese transporter at the phagosomal membrane. J Exp Med. 2000;192(9):1237-48.

95. Kehl-Fie TE, Chitayat S, Hood MI, Damo S, Restrepo N, Garcia C, et al. Nutrient metal sequestration by calprotectin inhibits bacterial superoxide defense, enhancing neutrophil killing of Staphylococcus aureus. Cell Host Microbe. 2011;10(2):158-64.

96. Damo SM, Kehl-Fie TE, Sugitani N, Holt ME, Rathi S, Murphy WJ, et al. Molecular basis for manganese sequestration by calprotectin and roles in the innate immune response to invading bacterial pathogens. Proc Natl Acad Sci U S A. 2013;110(10):3841-6.

97. Begum NA, Kobayashi M, Moriwaki Y, Matsumoto M, Toyoshima K, Seya T. Mycobacterium bovis BCG cell wall and lipopolysaccharide induce a novel gene, BIGM103, encoding a 7-TM protein: identification of a new protein family having Zn-transporter and Zn-metalloprotease signatures. Genomics. 2002;80(6):630- 45.

98. Hood MI, Skaar EP. Nutritional immunity: transition metals at the pathogen-host interface. Nat Rev Microbiol. 2012;10(8):525-37.

99. Wolschendorf F, Ackart D, Shrestha TB, Hascall-Dove L, Nolan S, Lamichhane G, et al. Copper resistance is essential for virulence of Mycobacterium tuberculosis. Proc Natl Acad Sci U S A. 2011;108(4):1621-6.

100. Panek H, O'Brian MR. A whole genome view of prokaryotic haem biosynthesis. Microbiology. 2002;148(Pt 8):2273-82. 175

101. White DC, Granick S. Hemin Biosynthesis in Haemophilus. J Bacteriol. 1963;85:842-50.

102. Morton DJ, Turman EJ, Hensley PD, VanWagoner TM, Seale TW, Whitby PW, et al. Identification of a siderophore utilization locus in nontypeable Haemophilus influenzae. BMC Microbiol. 2010;10:113.

103. Harrison A, Dyer DW, Gillaspy A, Ray WC, Mungur R, Carson MB, et al. Genomic sequence of an otitis media isolate of nontypeable Haemophilus influenzae: comparative study with H. influenzae serotype d, strain KW20. J Bacteriol. 2005;187(13):4627-36.

104. Cope LD, Thomas SE, Latimer JL, Slaughter CA, Muller-Eberhard U, Hansen EJ. The 100 kDa haem:haemopexin-binding protein of Haemophilus influenzae: structure and localization. Mol Microbiol. 1994;13(5):863-73.

105. Morton DJ, Smith A, Ren Z, Madore LL, VanWagoner TM, Seale TW, et al. Identification of a haem-utilization protein (Hup) in Haemophilus influenzae. Microbiology. 2004;150(Pt 12):3923-33.

106. Maciver I, Latimer JL, Liem HH, Muller-Eberhard U, Hrkal Z, Hansen EJ. Identification of an outer membrane protein involved in utilization of hemoglobin- haptoglobin complexes by nontypeable Haemophilus influenzae. Infect Immun. 1996;64(9):3703-12.

107. Morton DJ, Bakaletz LO, Jurcisek JA, VanWagoner TM, Seale TW, Whitby PW, et al. Reduced severity of middle ear infection caused by nontypeable Haemophilus influenzae lacking the hemoglobin/hemoglobin-haptoglobin binding proteins (Hgp) in a chinchilla model of otitis media. Microb Pathog. 2004;36(1):25-33.

108. Ren Z, Jin H, Morton DJ, Stull TL. hgpB, a gene encoding a second Haemophilus influenzae hemoglobin- and hemoglobin-haptoglobin-binding protein. Infect Immun. 1998;66(10):4733-41.

109. Gray-Owen SD, Loosmore S, Schryvers AB. Identification and characterization of genes encoding the human transferrin-binding proteins from Haemophilus influenzae. Infect Immun. 1995;63(4):1201-10.

110. Al Jubair T, Singh B, Fleury C, Blom AM, Morgelin M, Thunnissen MM, et al. Haemophilus influenzae stores and distributes hemin by using protein E. Int J Med Microbiol. 2014;304(5-6):662-8.

111. Morton DJ, Seale TW, Vanwagoner TM, Whitby PW, Stull TL. The dppBCDF gene cluster of Haemophilus influenzae: Role in heme utilization. BMC Res Notes. 2009;2:166.

176

112. Mason KM, Raffel FK, Ray WC, Bakaletz LO. Heme utilization by nontypeable Haemophilus influenzae is essential and dependent on Sap transporter function. J Bacteriol. 2011;193(10):2527-35.

113. Mason KM, Bruggeman ME, Munson RS, Bakaletz LO. The non-typeable Haemophilus influenzae Sap transporter provides a mechanism of antimicrobial peptide resistance and SapD-dependent potassium acquisition. Mol Microbiol. 2006;62(5):1357-72.

114. Mason KM, Munson RS, Jr., Bakaletz LO. A mutation in the sap operon attenuates survival of nontypeable Haemophilus influenzae in a chinchilla model of otitis media. Infect Immun. 2005;73(1):599-608.

115. Shelton CL, Raffel FK, Beatty WL, Johnson SM, Mason KM. Sap transporter mediated import and subsequent degradation of antimicrobial peptides in Haemophilus. PLoS Pathog. 2011;7(11):e1002360.

116. Vogel AR, Szelestey BR, Raffel FK, Sharpe SW, Gearinger RL, Justice SS, et al. SapF-mediated heme-iron utilization enhances persistence and coordinates biofilm architecture of Haemophilus. Front Cell Infect Microbiol. 2012;2:42.

117. Carpenter BM, Whitmire JM, Merrell DS. This is not your mother's repressor: the complex role of fur in pathogenesis. Infect Immun. 2009;77(7):2590-601.

118. Santana EA, Harrison A, Zhang X, Baker BD, Kelly BJ, White P, et al. HrrF is the Fur-regulated small RNA in nontypeable Haemophilus influenzae. PLoS One. 2014;9(8):e105644.

119. Petersen S, Young GM. Essential role for cyclic AMP and its receptor protein in Yersinia enterocolitica virulence. Infect Immun. 2002;70(7):3665-72.

120. Botsford JL. Cyclic nucleotides in procaryotes. Microbiol Rev. 1981;45(4):620-42.

121. Ahuja N, Kumar P, Bhatnagar R. The adenylate cyclase toxins. Crit Rev Microbiol. 2004;30(3):187-96.

122. Wolff J, Cook GH. Activation of thyroid membrane adenylate cyclase by purine nucleotides. J Biol Chem. 1973;248(1):350-5.

123. Levy H, Weiss S, Altboum Z, Schlomovitz J, Rothschild N, Glinert I, et al. The effect of deletion of the edema factor on Bacillus anthracis pathogenicity in guinea pigs and rabbits. Microb Pathog. 2012;52(1):55-60.

124. Cowell BA, Evans DJ, Fleiszig SM. Actin cytoskeleton disruption by ExoY and its effects on Pseudomonas aeruginosa invasion. FEMS Microbiol Lett. 2005;250(1):71-6.

177

125. Shevchenko LA, Mishankin BN. [Adenylate cyclase of the causative agent of plague: its purification and properties]. Zh Mikrobiol Epidemiol Immunobiol. 1987(12):15-20.

126. Inada T, Takahashi H, Mizuno T, Aiba H. Down regulation of cAMP production by cAMP receptor protein in Escherichia coli: an assessment of the contributions of transcriptional and posttranscriptional control of adenylate cyclase. Mol Gen Genet. 1996;253(1-2):198-204.

127. Richter W. 3',5' Cyclic nucleotide phosphodiesterases class III: members, structure, and catalytic mechanism. Proteins. 2002;46(3):278-86.

128. Matange N. Revisiting bacterial cyclic nucleotide phosphodiesterases: cyclic AMP hydrolysis and beyond. FEMS Microbiol Lett. 2015;362(22):1-9.

129. Fuchs EL, Brutinel ED, Klem ER, Fehr AR, Yahr TL, Wolfgang MC. In vitro and in vivo characterization of the Pseudomonas aeruginosa cyclic AMP (cAMP) phosphodiesterase CpdA, required for cAMP homeostasis and virulence factor regulation. J Bacteriol. 2010;192(11):2779-90.

130. Gancedo JM. Biological roles of cAMP: variations on a theme in the different kingdoms of life. Biol Rev Camb Philos Soc. 2013;88(3):645-68.

131. Gorke B, Stulke J. Carbon catabolite repression in bacteria: many ways to make the most out of nutrients. Nat Rev Microbiol. 2008;6(8):613-24.

132. Shimada T, Fujita N, Yamamoto K, Ishihama A. Novel roles of cAMP receptor protein (CRP) in regulation of transport and metabolism of carbon sources. PLoS One. 2011;6(6):e20081.

133. Barth E, Gora KV, Gebendorfer KM, Settele F, Jakob U, Winter J. Interplay of cellular cAMP levels, {sigma}S activity and oxidative stress resistance in Escherichia coli. Microbiology. 2009;155(Pt 5):1680-9.

134. Shenoy AR, Sreenath N, Podobnik M, Kovacevic M, Visweswariah SS. The Rv0805 gene from Mycobacterium tuberculosis encodes a 3',5'-cyclic nucleotide phosphodiesterase: biochemical and mutational analysis. Biochemistry. 2005;44(48):15695-704.

135. Be NA, Lamichhane G, Grosset J, Tyagi S, Cheng QJ, Kim KS, et al. Murine model to study the invasion and survival of Mycobacterium tuberculosis in the central nervous system. J Infect Dis. 2008;198(10):1520-8.

136. Podobnik M, Tyagi R, Matange N, Dermol U, Gupta AK, Mattoo R, et al. A mycobacterial cyclic AMP phosphodiesterase that moonlights as a modifier of cell wall permeability. J Biol Chem. 2009;284(47):32846-57.

178

137. Matange N, Podobnik M, Visweswariah SS. The non-catalytic "cap domain" of a mycobacterial metallophosphoesterase regulates its expression and localization in the cell. J Biol Chem. 2014;289(32):22470-81.

138. Kalivoda EJ, Brothers KM, Stella NA, Schmitt MJ, Shanks RM. Bacterial cyclic AMP-phosphodiesterase activity coordinates biofilm formation. PLoS One. 2013;8(7):e71267.

139. Domingues S, Nielsen KM, da Silva GJ. Various pathways leading to the acquisition of antibiotic resistance by natural transformation. Mob Genet Elements. 2012;2(6):257-60.

140. Pifer ML, Smith HO. Processing of donor DNA during Haemophilus influenzae transformation: analysis using a model plasmid system. Proc Natl Acad Sci U S A. 1985;82(11):3731-5.

141. Stewart GJ, Carlson CA. The biology of natural transformation. Annu Rev Microbiol. 1986;40:211-35.

142. Seitz P, Blokesch M. Cues and regulatory pathways involved in natural competence and transformation in pathogenic and environmental Gram-negative bacteria. FEMS Microbiol Rev. 2013;37(3):336-63.

143. Redfield RJ, Cameron AD, Qian Q, Hinds J, Ali TR, Kroll JS, et al. A novel CRP- dependent regulon controls expression of competence genes in Haemophilus influenzae. J Mol Biol. 2005;347(4):735-47.

144. Dorocicz IR, Williams PM, Redfield RJ. The Haemophilus influenzae adenylate cyclase gene: cloning, sequence, and essential role in competence. J Bacteriol. 1993;175(22):7142-9.

145. Macfadyen LP. Regulation of competence development in Haemophilus influenzae. J Theor Biol. 2000;207(3):349-59.

146. Cameron AD, Redfield RJ. CRP binding and transcription activation at CRP-S sites. J Mol Biol. 2008;383(2):313-23.

147. Sinha S, Mell JC, Redfield RJ. Seventeen Sxy-dependent cyclic AMP receptor protein site-regulated genes are needed for natural transformation in Haemophilus influenzae. J Bacteriol. 2012;194(19):5245-54.

148. Wise Jr. EM, Alexander SP, Powers M. Adenosine 3' : 5'-cyclic monophosphate as a regulator of bacterial transformation. Proc Natl Acad Sci U S A. 1973;70(2):471-4.

149. Fleischmann RD, Adams MD, White O, Clayton RA, Kirkness EF, Kerlavage AR, et al. Whole-genome random sequencing and assembly of Haemophilus influenzae Rd. Science. 1995;269(5223):496-512.

179

150. Macfadyen LP, Ma C, Redfield RJ. A 3',5' cyclic AMP (cAMP) phosphodiesterase modulates cAMP levels and optimizes competence in Haemophilus influenzae Rd. J Bacteriol. 1998;180(17):4401-5.

151. Imamura R, Yamanaka K, Ogura T, Hiraga S, Fujita N, Ishihama A, et al. Identification of the cpdA gene encoding cyclic 3',5'-adenosine monophosphate phosphodiesterase in Escherichia coli. J Biol Chem. 1996;271(41):25423-9.

152. Lee JH, Wood TK, Lee J. Roles of indole as an interspecies and interkingdom signaling molecule. Trends Microbiol. 2015;23(11):707-18.

153. Kim J, Park W. Indole: a signaling molecule or a mere metabolic byproduct that alters bacterial physiology at a high concentration? J Microbiol. 2015;53(7):421- 8.

154. Lee JH, Lee J. Indole as an intercellular signal in microbial communities. FEMS Microbiol Rev. 2010;34(4):426-44.

155. Yanofsky C, Horn V, Gollnick P. Physiological studies of tryptophan transport and tryptophanase operon induction in Escherichia coli. J Bacteriol. 1991;173(19):6009-17.

156. Li G, Young KD. Indole production by the tryptophanase TnaA in Escherichia coli is determined by the amount of exogenous tryptophan. Microbiology. 2013;159(Pt 2):402-10.

157. Kawamura-Sato K, Shibayama K, Horii T, Iimuma Y, Arakawa Y, Ohta M. Role of multiple efflux pumps in Escherichia coli in indole expulsion. FEMS Microbiol Lett. 1999;179(2):345-52.

158. Pinero-Fernandez S, Chimerel C, Keyser UF, Summers DK. Indole transport across Escherichia coli membranes. J Bacteriol. 2011;193(8):1793-8.

159. Hirakawa H, Inazumi Y, Masaki T, Hirata T, Yamaguchi A. Indole induces the expression of multidrug exporter genes in Escherichia coli. Mol Microbiol. 2005;55(4):1113-26.

160. Nikaido E, Giraud E, Baucheron S, Yamasaki S, Wiedemann A, Okamoto K, et al. Effects of indole on drug resistance and virulence of Salmonella enterica serovar Typhimurium revealed by genome-wide analyses. Gut Pathog. 2012;4(1):5.

161. Lee J, Jayaraman A, Wood TK. Indole is an inter-species biofilm signal mediated by SdiA. BMC Microbiol. 2007;7:42.

162. Saint-Ruf C, Garfa-Traore M, Collin V, Cordier C, Franceschi C, Matic I. Massive diversification in aging colonies of Escherichia coli. J Bacteriol. 2014;196(17):3059-73.

180

163. Chen CC, Walia R, Mukherjee KJ, Mahalik S, Summers DK. Indole generates quiescent and metabolically active Escherichia coli cultures. Biotechnol J. 2015;10(4):636-46.

164. Martin K, Morlin G, Smith A, Nordyke A, Eisenstark A, Golomb M. The tryptophanase gene cluster of Haemophilus influenzae type b: evidence for horizontal gene transfer. J Bacteriol. 1998;180(1):107-18.

165. Bansal T, Englert D, Lee J, Hegde M, Wood TK, Jayaraman A. Differential effects of epinephrine, norepinephrine, and indole on Escherichia coli O157:H7 chemotaxis, colonization, and gene expression. Infect Immun. 2007;75(9):4597- 607.

166. Oh S, Go GW, Mylonakis E, Kim Y. The bacterial signalling molecule indole attenuates the virulence of the fungal pathogen Candida albicans. J Appl Microbiol. 2012;113(3):622-8.

167. Kohli N, Crisp Z, Riordan R, Li M, Alaniz RC, Jayaraman A. The microbiota metabolite indole inhibits Salmonella virulence: Involvement of the PhoPQ two- component system. PLoS One. 2018;13(1):e0190613.

168. Chu P, Rodriguez AR, Arulanandam BP, Klose KE. Tryptophan prototrophy contributes to Francisella tularensis evasion of gamma interferon-mediated host defense. Infect Immun. 2011;79(6):2356-61.

169. Beatty WL, Belanger TA, Desai AA, Morrison RP, Byrne GI. Tryptophan depletion as a mechanism of gamma interferon-mediated chlamydial persistence. Infect Immun. 1994;62(9):3705-11.

170. Walker WT, Jackson CL, Allan RN, Collins SA, Kelso MJ, Rineh A, et al. Primary ciliary dyskinesia ciliated airway cells show increased susceptibility to Haemophilus influenzae biofilm formation. Eur Respir J. 2017;50(3).

171. Ketterer MR, Shao JQ, Hornick DB, Buscher B, Bandi VK, Apicella MA. Infection of primary human bronchial epithelial cells by Haemophilus influenzae: macropinocytosis as a mechanism of airway epithelial cell entry. Infect Immun. 1999;67(8):4161-70.

172. Reddy MS, Bernstein JM, Murphy TF, Faden HS. Binding between outer membrane proteins of nontypeable Haemophilus influenzae and human nasopharyngeal mucin. Infect Immun. 1996;64(4):1477-9.

173. Su YC, Mukherjee O, Singh B, Hallgren O, Westergren-Thorsson G, Hood D, et al. Haemophilus influenzae P4 Interacts With Extracellular Matrix Proteins Promoting Adhesion and Serum Resistance. J Infect Dis. 2016;213(2):314-23.

174. Hallstrom T, Singh B, Resman F, Blom AM, Morgelin M, Riesbeck K. Haemophilus influenzae protein E binds to the extracellular matrix by

181

concurrently interacting with laminin and vitronectin. J Infect Dis. 2011;204(7):1065-74.

175. Jalalvand F, Su YC, Morgelin M, Brant M, Hallgren O, Westergren-Thorsson G, et al. Haemophilus influenzae protein F mediates binding to laminin and human pulmonary epithelial cells. J Infect Dis. 2013;207(5):803-13.

176. Novotny LA, Bakaletz LO. Intercellular adhesion molecule 1 serves as a primary cognate receptor for the Type IV pilus of nontypeable Haemophilus influenzae. Cell Microbiol. 2016;18(8):1043-55.

177. Desvaux M, Parham NJ, Henderson IR. The autotransporter secretion system. Res Microbiol. 2004;155(2):53-60.

178. Voulhoux R, Bos MP, Geurtsen J, Mols M, Tommassen J. Role of a highly conserved bacterial protein in outer membrane protein assembly. Science. 2003;299(5604):262-5.

179. Jain S, Goldberg MB. Requirement for YaeT in the outer membrane assembly of autotransporter proteins. J Bacteriol. 2007;189(14):5393-8.

180. Lehr U, Schutz M, Oberhettinger P, Ruiz-Perez F, Donald JW, Palmer T, et al. C- terminal amino acid residues of the trimeric autotransporter adhesin YadA of Yersinia enterocolitica are decisive for its recognition and assembly by BamA. Mol Microbiol. 2010;78(4):932-46.

181. Barenkamp SJ, Leininger E. Cloning, expression, and DNA sequence analysis of genes encoding nontypeable Haemophilus influenzae high-molecular-weight surface-exposed proteins related to filamentous hemagglutinin of Bordetella pertussis. Infect Immun. 1992;60(4):1302-13.

182. Buscher AZ, Burmeister K, Barenkamp SJ, St Geme JW, 3rd. Evolutionary and functional relationships among the nontypeable Haemophilus influenzae HMW family of adhesins. J Bacteriol. 2004;186(13):4209-17.

183. Grass S, St Geme JW, 3rd. Maturation and secretion of the non-typable Haemophilus influenzae HMW1 adhesin: roles of the N-terminal and C-terminal domains. Mol Microbiol. 2000;36(1):55-67.

184. Buscher AZ, Grass S, Heuser J, Roth R, St Geme JW, 3rd. Surface anchoring of a bacterial adhesin secreted by the two-partner secretion pathway. Mol Microbiol. 2006;61(2):470-83.

185. Grass S, Buscher AZ, Swords WE, Apicella MA, Barenkamp SJ, Ozchlewski N, et al. The Haemophilus influenzae HMW1 adhesin is glycosylated in a process that requires HMW1C and phosphoglucomutase, an enzyme involved in lipooligosaccharide biosynthesis. Mol Microbiol. 2003;48(3):737-51.

182

186. St Geme JW, 3rd. Molecular and cellular determinants of non-typeable Haemophilus influenzae adherence and invasion. Cell Microbiol. 2002;4(4):191- 200.

187. Yamada S, Suzuki Y, Suzuki T, Le MQ, Nidom CA, Sakai-Tagawa Y, et al. Haemagglutinin mutations responsible for the binding of H5N1 influenza A viruses to human-type receptors. Nature. 2006;444(7117):378-82.

188. Atack JM, Day CJ, Poole J, Brockman KL, Bakaletz LO, Barenkamp SJ, et al. The HMW2 adhesin of non-typeable Haemophilus influenzae is a human- adapted lectin that mediates high-affinity binding to 2-6 linked N-acetylneuraminic acid glycans. Biochem Biophys Res Commun. 2018.

189. Barenkamp SJ. Outer membrane proteins and lipopolysaccharides of nontypeable Haemophilus influenzae. J Infect Dis. 1992;165 Suppl 1:S181-4.

190. Ecevit IZ, McCrea KW, Pettigrew MM, Sen A, Marrs CF, Gilsdorf JR. Prevalence of the hifBC, hmw1A, hmw2A, hmwC, and hia Genes in Haemophilus influenzae Isolates. J Clin Microbiol. 2004;42(7):3065-72.

191. Erwin AL, Nelson KL, Mhlanga-Mutangadura T, Bonthuis PJ, Geelhood JL, Morlin G, et al. Characterization of genetic and phenotypic diversity of invasive nontypeable Haemophilus influenzae. Infect Immun. 2005;73(9):5853-63.

192. Erwin AL, Sandstedt SA, Bonthuis PJ, Geelhood JL, Nelson KL, Unrath WC, et al. Analysis of genetic relatedness of Haemophilus influenzae isolates by multilocus sequence typing. J Bacteriol. 2008;190(4):1473-83.

193. St Geme JW, 3rd, Kumar VV, Cutter D, Barenkamp SJ. Prevalence and distribution of the hmw and hia genes and the HMW and Hia adhesins among genetically diverse strains of nontypeable Haemophilus influenzae. Infect Immun. 1998;66(1):364-8.

194. van Schilfgaarde M, van Ulsen P, van Der Steeg W, Winter V, Eijk P, Everts V, et al. Cloning of genes of nontypeable Haemophilus influenzae involved in penetration between human lung epithelial cells. Infect Immun. 2000;68(8):4616- 23.

195. St Geme JW, 3rd, de la Morena ML, Falkow S. A Haemophilus influenzae IgA protease-like protein promotes intimate interaction with human epithelial cells. Mol Microbiol. 1994;14(2):217-33.

196. Meng G, Spahich N, Kenjale R, Waksman G, St Geme JW, 3rd. Crystal structure of the Haemophilus influenzae Hap adhesin reveals an intercellular oligomerization mechanism for bacterial aggregation. EMBO J. 2011;30(18):3864-74.

197. Spahich NA, St Geme JW, 3rd. Structure and function of the Haemophilus influenzae autotransporters. Front Cell Infect Microbiol. 2011;1:5. 183

198. Hendrixson DR, St Geme JW, 3rd. The Haemophilus influenzae Hap serine protease promotes adherence and microcolony formation, potentiated by a soluble host protein. Mol Cell. 1998;2(6):841-50.

199. Klemm P, Vejborg RM, Sherlock O. Self-associating autotransporters, SAATs: functional and structural similarities. Int J Med Microbiol. 2006;296(4-5):187-95.

200. Fink DL, Green BA, St Geme JW, 3rd. The Haemophilus influenzae Hap autotransporter binds to fibronectin, laminin, and collagen IV. Infect Immun. 2002;70(9):4902-7.

201. Baddal B, Muzzi A, Censini S, Calogero RA, Torricelli G, Guidotti S, et al. Dual RNA-seq of Nontypeable Haemophilus influenzae and Host Cell Transcriptomes Reveals Novel Insights into Host-Pathogen Cross Talk. MBio. 2015;6(6):e01765- 15.

202. Klausen M, Heydorn A, Ragas P, Lambertsen L, Aaes-Jorgensen A, Molin S, et al. Biofilm formation by Pseudomonas aeruginosa wild type, flagella and type IV pili mutants. Mol Microbiol. 2003;48(6):1511-24.

203. Limoli DH, Jones CJ, Wozniak DJ. Bacterial Extracellular Polysaccharides in Biofilm Formation and Function. Microbiol Spectr. 2015;3(3).

204. Morey P, Viadas C, Euba B, Hood DW, Barberan M, Gil C, et al. Relative contributions of lipooligosaccharide inner and outer core modifications to nontypeable Haemophilus influenzae pathogenesis. Infect Immun. 2013;81(11):4100-11.

205. Hong W, Pang B, West-Barnette S, Swords WE. Phosphorylcholine expression by nontypeable Haemophilus influenzae correlates with maturation of biofilm communities in vitro and in vivo. J Bacteriol. 2007;189(22):8300-7.

206. Swords WE, Moore ML, Godzicki L, Bukofzer G, Mitten MJ, VonCannon J. Sialylation of lipooligosaccharides promotes biofilm formation by nontypeable Haemophilus influenzae. Infect Immun. 2004;72(1):106-13.

207. Jurcisek JA, Brockman KL, Novotny LA, Goodman SD, Bakaletz LO. Nontypeable Haemophilus influenzae releases DNA and DNABII proteins via a T4SS-like complex and ComE of the type IV pilus machinery. Proc Natl Acad Sci U S A. 2017;114(32):E6632-E41.

208. Jones EA, McGillivary G, Bakaletz LO. Extracellular DNA within a nontypeable Haemophilus influenzae-induced biofilm binds human beta defensin-3 and reduces its antimicrobial activity. J Innate Immun. 2013;5(1):24-38.

209. Jurcisek JA, Bakaletz LO. Biofilms formed by nontypeable Haemophilus influenzae in vivo contain both double-stranded DNA and type IV pilin protein. J Bacteriol. 2007;189(10):3868-75.

184

210. Jurcisek JA, Bookwalter JE, Baker BD, Fernandez S, Novotny LA, Munson RS, Jr., et al. The PilA protein of non-typeable Haemophilus influenzae plays a role in biofilm formation, adherence to epithelial cells and colonization of the mammalian upper respiratory tract. Mol Microbiol. 2007;65(5):1288-99.

211. Goodman SD, Obergfell KP, Jurcisek JA, Novotny LA, Downey JS, Ayala EA, et al. Biofilms can be dispersed by focusing the immune system on a common family of bacterial nucleoid-associated proteins. Mucosal Immunol. 2011;4(6):625-37.

212. Novotny LA, Jurcisek JA, Goodman SD, Bakaletz LO. Monoclonal antibodies against DNA-binding tips of DNABII proteins disrupt biofilms in vitro and induce bacterial clearance in vivo. EBioMedicine. 2016;10:33-44.

213. Novotny LA, Jurcisek JA, Ward MO, Jr., Jordan ZB, Goodman SD, Bakaletz LO. Antibodies against the majority subunit of type IV Pili disperse nontypeable Haemophilus influenzae biofilms in a LuxS-dependent manner and confer therapeutic resolution of experimental otitis media. Mol Microbiol. 2015;96(2):276-92.

214. Brockson ME, Novotny LA, Mokrzan EM, Malhotra S, Jurcisek JA, Akbar R, et al. Evaluation of the kinetics and mechanism of action of anti-integration host factor- mediated disruption of bacterial biofilms. Mol Microbiol. 2014;93(6):1246-58.

215. Devaraj A, Buzzo J, Rocco CJ, Bakaletz LO, Goodman SD. The DNABII family of proteins is comprised of the only nucleoid associated proteins required for nontypeable Haemophilus influenzae biofilm structure. Microbiologyopen. 2018;7(3):e00563.

216. Hengge R. Principles of c-di-GMP signalling in bacteria. Nat Rev Microbiol. 2009;7(4):263-73.

217. Chambers JR, Sauer K. Small RNAs and their role in biofilm formation. Trends Microbiol. 2013;21(1):39-49.

218. Irie Y, Starkey M, Edwards AN, Wozniak DJ, Romeo T, Parsek MR. Pseudomonas aeruginosa biofilm matrix polysaccharide Psl is regulated transcriptionally by RpoS and post-transcriptionally by RsmA. Mol Microbiol. 2010;78(1):158-72.

219. Hufnagel DA, Evans ML, Greene SE, Pinkner JS, Hultgren SJ, Chapman MR. The Catabolite Repressor Protein-Cyclic AMP Complex Regulates csgD and Biofilm Formation in Uropathogenic Escherichia coli. J Bacteriol. 2016;198(24):3329-34.

220. Brockman KL, Jurcisek JA, Atack JM, Srikhanta YN, Jennings MP, Bakaletz LO. ModA2 Phasevarion Switching in Nontypeable Haemophilus influenzae Increases the Severity of Experimental Otitis Media. J Infect Dis. 2016;214(5):817-24. 185

221. Hall-Stoodley L, Hu FZ, Gieseke A, Nistico L, Nguyen D, Hayes J, et al. Direct detection of bacterial biofilms on the middle-ear mucosa of children with chronic otitis media. JAMA. 2006;296(2):202-11.

222. Allegrucci M, Hu FZ, Shen K, Hayes J, Ehrlich GD, Post JC, et al. Phenotypic characterization of Streptococcus pneumoniae biofilm development. J Bacteriol. 2006;188(7):2325-35.

223. Murphy TF, Kirkham C. Biofilm formation by nontypeable Haemophilus influenzae: strain variability, outer membrane antigen expression and role of pili. BMC Microbiol. 2002;2:7.

224. Pearson MM, Laurence CA, Guinn SE, Hansen EJ. Biofilm formation by Moraxella catarrhalis in vitro: roles of the UspA1 adhesin and the Hag hemagglutinin. Infect Immun. 2006;74(3):1588-96.

225. Ehrlich GD, Veeh R, Wang X, Costerton JW, Hayes JD, Hu FZ, et al. Mucosal biofilm formation on middle-ear mucosa in the chinchilla model of otitis media. JAMA. 2002;287(13):1710-5.

226. Bakaletz LO. Bacterial biofilms in the upper airway - evidence for role in pathology and implications for treatment of otitis media. Paediatr Respir Rev. 2012;13(3):154-9.

227. Post JC, Hiller NL, Nistico L, Stoodley P, Ehrlich GD. The role of biofilms in otolaryngologic infections: update 2007. Curr Opin Otolaryngol Head Neck Surg. 2007;15(5):347-51.

228. Marti S, Puig C, Merlos A, Vinas M, de Jonge MI, Linares J, et al. Bacterial Lysis through Interference with Peptidoglycan Synthesis Increases Biofilm Formation by Nontypeable Haemophilus influenzae. mSphere. 2017;2(1).

229. King PT, Sharma R. The Lung Immune Response to Nontypeable Haemophilus influenzae (Lung Immunity to NTHi). J Immunol Res. 2015;2015:706376.

230. Hernandez M, Leichtle A, Pak K, Webster NJ, Wasserman SI, Ryan AF. The transcriptome of a complete episode of acute otitis media. BMC Genomics. 2015;16:259.

231. Juneau RA, Pang B, Weimer KE, Armbruster CE, Swords WE. Nontypeable Haemophilus influenzae initiates formation of neutrophil extracellular traps. Infect Immun. 2011;79(1):431-8.

232. Skotnicka B, Stasiak-Barmuta A, Hassmann-Poznanska E, Kasprzycka E. Lymphocyte subpopulations in middle ear effusions: flow cytometry analysis. Otol Neurotol. 2005;26(4):567-71.

233. Jecker P, Pabst R, Westermann J. Proliferating macrophages, dendritic cells, natural killer cells, T and B lymphocytes in the middle ear and Eustachian tube

186

mucosa during experimental acute otitis media in the rat. Clin Exp Immunol. 2001;126(3):421-5.

234. Forseni M, Bagger-Sjoback D, Hultcrantz M. A study of inflammatory mediators in the human tympanosclerotic middle ear. Arch Otolaryngol Head Neck Surg. 2001;127(5):559-64.

235. Vasselon T, Detmers PA. Toll receptors: a central element in innate immune responses. Infect Immun. 2002;70(3):1033-41.

236. Li JD. Exploitation of host epithelial signaling networks by respiratory bacterial pathogens. J Pharmacol Sci. 2003;91(1):1-7.

237. Shuto T, Xu H, Wang B, Han J, Kai H, Gu XX, et al. Activation of NF-kappa B by nontypeable Hemophilus influenzae is mediated by toll-like receptor 2-TAK1- dependent NIK-IKK alpha /beta-I kappa B alpha and MKK3/6-p38 MAP kinase signaling pathways in epithelial cells. Proc Natl Acad Sci U S A. 2001;98(15):8774-9.

238. Clemans DL, Bauer RJ, Hanson JA, Hobbs MV, St Geme JW, 3rd, Marrs CF, et al. Induction of proinflammatory cytokines from human respiratory epithelial cells after stimulation by nontypeable Haemophilus influenzae. Infect Immun. 2000;68(8):4430-40.

239. Lin J, Caye-Thomasen P, Tono T, Zhang QA, Nakamura Y, Feng L, et al. Mucin production and mucous cell metaplasia in otitis media. Int J Otolaryngol. 2012;2012:745325.

240. Mittal R, Kodiyan J, Gerring R, Mathee K, Li JD, Grati M, et al. Role of innate immunity in the pathogenesis of otitis media. Int J Infect Dis. 2014;29:259-67.

241. Kawano H, Paparella MM, Ho SB, Schachern PA, Morizono N, Le CT, et al. Identification of MUC5B mucin gene in human middle ear with chronic otitis media. Laryngoscope. 2000;110(4):668-73.

242. Ubell ML, Kerschner JE, Wackym PA, Burrows A. MUC2 expression in human middle ear epithelium of patients with otitis media. Arch Otolaryngol Head Neck Surg. 2008;134(1):39-44.

243. Kerschner JE, Tripathi S, Khampang P, Papsin BC. MUC5AC expression in human middle ear epithelium of patients with otitis media. Arch Otolaryngol Head Neck Surg. 2010;136(8):819-24.

244. Chen R, Lim JH, Jono H, Gu XX, Kim YS, Basbaum CB, et al. Nontypeable Haemophilus influenzae lipoprotein P6 induces MUC5AC mucin transcription via TLR2-TAK1-dependent p38 MAPK-AP1 and IKKbeta-IkappaBalpha-NF-kappaB signaling pathways. Biochem Biophys Res Commun. 2004;324(3):1087-94.

187

245. Jono H, Shuto T, Xu H, Kai H, Lim DJ, Gum JR, Jr., et al. Transforming growth factor-beta -Smad signaling pathway cooperates with NF-kappa B to mediate nontypeable Haemophilus influenzae-induced MUC2 mucin transcription. J Biol Chem. 2002;277(47):45547-57.

246. Shen H, Yoshida H, Yan F, Li W, Xu F, Huang H, et al. Synergistic induction of MUC5AC mucin by nontypeable Haemophilus influenzae and Streptococcus pneumoniae. Biochem Biophys Res Commun. 2008;365(4):795-800.

247. Solzbacher D, Hanisch FG, van Alphen L, Gilsdorf JR, Schroten H. Mucin in middle ear effusions inhibits attachment of Haemophilus influenzae to mucosal epithelial cells. Eur Arch Otorhinolaryngol. 2003;260(3):141-7.

248. Brogden KA. Antimicrobial peptides: pore formers or metabolic inhibitors in bacteria? Nat Rev Microbiol. 2005;3(3):238-50.

249. Sitaram N, Nagaraj R. Host-defense antimicrobial peptides: importance of structure for activity. Curr Pharm Des. 2002;8(9):727-42.

250. Park CB, Kim HS, Kim SC. Mechanism of action of the antimicrobial peptide buforin II: buforin II kills microorganisms by penetrating the cell membrane and inhibiting cellular functions. Biochem Biophys Res Commun. 1998;244(1):253-7.

251. Skerlavaj B, Romeo D, Gennaro R. Rapid membrane permeabilization and inhibition of vital functions of gram-negative bacteria by bactenecins. Infect Immun. 1990;58(11):3724-30.

252. Lehrer RI, Barton A, Daher KA, Harwig SS, Ganz T, Selsted ME. Interaction of human defensins with Escherichia coli. Mechanism of bactericidal activity. J Clin Invest. 1989;84(2):553-61.

253. Scharf S, Zahlten J, Szymanski K, Hippenstiel S, Suttorp N, N'Guessan PD. Streptococcus pneumoniae induces human beta-defensin-2 and -3 in human lung epithelium. Exp Lung Res. 2012;38(2):100-10.

254. Jones EA, Kananurak A, Bevins CL, Hollox EJ, Bakaletz LO. Copy number variation of the beta defensin gene cluster on chromosome 8p influences the bacterial microbiota within the nasopharynx of otitis-prone children. PLoS One. 2014;9(5):e98269.

255. Lysenko ES, Gould J, Bals R, Wilson JM, Weiser JN. Bacterial phosphorylcholine decreases susceptibility to the antimicrobial peptide LL-37/hCAP18 expressed in the upper respiratory tract. Infect Immun. 2000;68(3):1664-71.

256. Walport MJ. Complement. First of two parts. N Engl J Med. 2001;344(14):1058- 66.

257. Walport MJ. Complement. Second of two parts. N Engl J Med. 2001;344(15):1140-4.

188

258. Zhang XL, Ali MA. Ficolins: structure, function and associated diseases. Adv Exp Med Biol. 2008;632:105-15.

259. Hallstrom T, Riesbeck K. Haemophilus influenzae and the complement system. Trends Microbiol. 2010;18(6):258-65.

260. Langereis JD, de Jonge MI, Weiser JN. Binding of human factor H to outer membrane protein P5 of non-typeable Haemophilus influenzae contributes to complement resistance. Mol Microbiol. 2014;94(1):89-106.

261. Rosadini CV, Ram S, Akerley BJ. Outer membrane protein P5 is required for resistance of nontypeable Haemophilus influenzae to both the classical and alternative complement pathways. Infect Immun. 2014;82(2):640-9.

262. Su YC, Jalalvand F, Morgelin M, Blom AM, Singh B, Riesbeck K. Haemophilus influenzae acquires vitronectin via the ubiquitous Protein F to subvert host innate immunity. Mol Microbiol. 2013;87(6):1245-66.

263. Hallstrom T, Blom AM, Zipfel PF, Riesbeck K. Nontypeable Haemophilus influenzae protein E binds vitronectin and is important for serum resistance. J Immunol. 2009;183(4):2593-601.

264. Ho DK, Ram S, Nelson KL, Bonthuis PJ, Smith AL. lgtC expression modulates resistance to C4b deposition on an invasive nontypeable Haemophilus influenzae. J Immunol. 2007;178(2):1002-12.

265. McConnell MJ, Rumbo C, Bou G, Pachon J. Outer membrane vesicles as an acellular vaccine against Acinetobacter baumannii. Vaccine. 2011;29(34):5705- 10.

266. Mitra S, Sinha R, Mitobe J, Koley H. Development of a cost-effective vaccine candidate with outer membrane vesicles of a tolA-disrupted Shigella boydii strain. Vaccine. 2016;34(15):1839-46.

267. Bottero D, Gaillard ME, Errea A, Moreno G, Zurita E, Pianciola L, et al. Outer membrane vesicles derived from Bordetella parapertussis as an acellular vaccine against Bordetella parapertussis and Bordetella pertussis infection. Vaccine. 2013;31(45):5262-8.

268. Nieves W, Petersen H, Judy BM, Blumentritt CA, Russell-Lodrigue K, Roy CJ, et al. A Burkholderia pseudomallei outer membrane vesicle vaccine provides protection against lethal sepsis. Clin Vaccine Immunol. 2014;21(5):747-54.

269. Roier S, Leitner DR, Iwashkiw J, Schild-Prufert K, Feldman MF, Krohne G, et al. Intranasal immunization with nontypeable Haemophilus influenzae outer membrane vesicles induces cross-protective immunity in mice. PLoS One. 2012;7(8):e42664.

189

270. Roier S, Blume T, Klug L, Wagner GE, Elhenawy W, Zangger K, et al. A basis for vaccine development: Comparative characterization of Haemophilus influenzae outer membrane vesicles. Int J Med Microbiol. 2015;305(3):298-309.

271. Sharpe SW, Kuehn MJ, Mason KM. Elicitation of epithelial cell-derived immune effectors by outer membrane vesicles of nontypeable Haemophilus influenzae. Infect Immun. 2011;79(11):4361-9.

272. Deknuydt F, Nordstrom T, Riesbeck K. Diversion of the host humoral response: a novel virulence mechanism of Haemophilus influenzae mediated via outer membrane vesicles. J Leukoc Biol. 2014;95(6):983-91.

273. Weiser JN, Pan N. Adaptation of Haemophilus influenzae to acquired and innate humoral immunity based on phase variation of lipopolysaccharide. Mol Microbiol. 1998;30(4):767-75.

274. Isaacson RE, Argyilan C, Kwan L, Patterson S, Yoshinaga K. Phase variable switching of in vivo and environmental phenotypes of Salmonella typhimurium. Adv Exp Med Biol. 1999;473:281-9.

275. Alamro M, Bidmos FA, Chan H, Oldfield NJ, Newton E, Bai X, et al. Phase variation mediates reductions in expression of surface proteins during persistent meningococcal carriage. Infect Immun. 2014;82(6):2472-84.

276. Dawid S, Barenkamp SJ, St Geme JW, 3rd. Variation in expression of the Haemophilus influenzae HMW adhesins: a prokaryotic system reminiscent of eukaryotes. Proc Natl Acad Sci U S A. 1999;96(3):1077-82.

277. Davis GS, Marino S, Marrs CF, Gilsdorf JR, Dawid S, Kirschner DE. Phase variation and host immunity against high molecular weight (HMW) adhesins shape population dynamics of nontypeable Haemophilus influenzae within human hosts. J Theor Biol. 2014;355:208-18.

278. Fox KL, Yildirim HH, Deadman ME, Schweda EK, Moxon ER, Hood DW. Novel lipopolysaccharide biosynthetic genes containing tetranucleotide repeats in Haemophilus influenzae, identification of a gene for adding O-acetyl groups. Mol Microbiol. 2005;58(1):207-16.

279. Fox KL, Atack JM, Srikhanta YN, Eckert A, Novotny LA, Bakaletz LO, et al. Selection for phase variation of LOS biosynthetic genes frequently occurs in progression of non-typeable Haemophilus influenzae infection from the nasopharynx to the middle ear of human patients. PLoS One. 2014;9(2):e90505.

280. Slauch JM, Mahan MJ, Michetti P, Neutra MR, Mekalanos JJ. Acetylation (O- factor 5) affects the structural and immunological properties of Salmonella typhimurium lipopolysaccharide O antigen. Infect Immun. 1995;63(2):437-41.

190

281. Berry DS, Lynn F, Lee CH, Frasch CE, Bash MC. Effect of O acetylation of Neisseria meningitidis serogroup A capsular polysaccharide on development of functional immune responses. Infect Immun. 2002;70(7):3707-13.

282. Cerutti A, Chen K, Chorny A. Immunoglobulin responses at the mucosal interface. Annu Rev Immunol. 2011;29:273-93.

283. Mehta SK, Plaut AG, Calvanico NJ, Tomasi TB, Jr. Human immunoglobulin A: production of an Fc fragment by an enteric microbial proteolytic enzyme. J Immunol. 1973;111(4):1274-6.

284. Plaut AG, Gilbert JV, Artenstein MS, Capra JD. Neisseria gonorrhoeae and neisseria meningitidis: extracellular enzyme cleaves human immunoglobulin A. Science. 1975;190(4219):1103-5.

285. Kilian M, Mestecky J, Schrohenloher RE. Pathogenic species of the genus Haemophilus and Streptococcus pneumoniae produce immunoglobulin A1 protease. Infect Immun. 1979;26(1):143-9.

286. Koomey JM, Falkow S. Nucleotide sequence homology between the immunoglobulin A1 protease genes of Neisseria gonorrhoeae, Neisseria meningitidis, and Haemophilus influenzae. Infect Immun. 1984;43(1):101-7.

287. Murphy TF, Kirkham C, Jones MM, Sethi S, Kong Y, Pettigrew MM. Expression of IgA Proteases by Haemophilus influenzae in the Respiratory Tract of Adults With Chronic Obstructive Pulmonary Disease. J Infect Dis. 2015;212(11):1798- 805.

288. Fernaays MM, Lesse AJ, Cai X, Murphy TF. Characterization of igaB, a second immunoglobulin A1 protease gene in nontypeable Haemophilus influenzae. Infect Immun. 2006;74(10):5860-70.

289. Poole J, Foster E, Chaloner K, Hunt J, Jennings MP, Bair T, et al. Analysis of nontypeable haemophilus influenzae phase-variable genes during experimental human nasopharyngeal colonization. J Infect Dis. 2013;208(5):720-7.

290. Vitovski S, Dunkin KT, Howard AJ, Sayers JR. Nontypeable Haemophilus influenzae in carriage and disease: a difference in IgA1 protease activity levels. JAMA. 2002;287(13):1699-705.

291. Parra-Lopez C, Baer MT, Groisman EA. Molecular genetic analysis of a locus required for resistance to antimicrobial peptides in Salmonella typhimurium. EMBO J. 1993;12(11):4053-62.

292. Lopez-Solanilla E, Garcia-Olmedo F, Rodriguez-Palenzuela P. Inactivation of the sapA to sapF locus of Erwinia chrysanthemi reveals common features in plant and animal bacterial pathogenesis. Plant Cell. 1998;10(6):917-24.

191

293. Hiles ID, Gallagher MP, Jamieson DJ, Higgins CF. Molecular characterization of the oligopeptide permease of Salmonella typhimurium. J Mol Biol. 1987;195(1):125-42.

294. Mason KM, Munson RS, Jr., Bakaletz LO. Nontypeable Haemophilus influenzae gene expression induced in vivo in a chinchilla model of otitis media. Infect Immun. 2003;71(6):3454-62.

295. Singh NK, Kunde DA, Tristram SG. Effect of epithelial cell type on in vitro invasion of non-typeable Haemophilus influenzae. J Microbiol Methods. 2016;129:66-9.

296. Egile C, Loisel TP, Laurent V, Li R, Pantaloni D, Sansonetti PJ, et al. Activation of the CDC42 effector N-WASP by the Shigella flexneri IcsA protein promotes actin nucleation by Arp2/3 complex and bacterial actin-based motility. J Cell Biol. 1999;146(6):1319-32.

297. Loisel TP, Boujemaa R, Pantaloni D, Carlier MF. Reconstitution of actin-based motility of Listeria and Shigella using pure proteins. Nature. 1999;401(6753):613- 6.

298. Harrison A, Dubois LG, St John-Williams L, Moseley MA, Hardison RL, Heimlich DR, et al. Comprehensive Proteomic and Metabolomic Signatures of Nontypeable Haemophilus influenzae-Induced Acute Otitis Media Reveal Bacterial Aerobic Respiration in an Immunosuppressed Environment. Mol Cell Proteomics. 2016;15(3):1117-38.

299. St Geme JW, 3rd, Falkow S. Haemophilus influenzae adheres to and enters cultured human epithelial cells. Infect Immun. 1990;58(12):4036-44.

300. Holmes KA, Bakaletz LO. Adherence of non-typeable Haemophilus influenzae promotes reorganization of the actin cytoskeleton in human or chinchilla epithelial cells in vitro. Microb Pathog. 1997;23(3):157-66.

301. Yoshida S, Handa Y, Suzuki T, Ogawa M, Suzuki M, Tamai A, et al. Microtubule- severing activity of Shigella is pivotal for intercellular spreading. Science. 2006;314(5801):985-9.

302. Hardwidge PR, Deng W, Vallance BA, Rodriguez-Escudero I, Cid VJ, Molina M, et al. Modulation of host cytoskeleton function by the enteropathogenic Escherichia coli and Citrobacter rodentium effector protein EspG. Infect Immun. 2005;73(5):2586-94.

303. Morey P, Cano V, Marti-Lliteras P, Lopez-Gomez A, Regueiro V, Saus C, et al. Evidence for a non-replicative intracellular stage of nontypable Haemophilus influenzae in epithelial cells. Microbiology. 2011;157(Pt 1):234-50.

304. Lopez-Gomez A, Cano V, Moranta D, Morey P, Garcia del Portillo F, Bengoechea JA, et al. Host cell kinases, alpha5 and beta1 integrins, and Rac1 192

signalling on the microtubule cytoskeleton are important for non-typable Haemophilus influenzae invasion of respiratory epithelial cells. Microbiology. 2012;158(Pt 9):2384-98.

305. Ahren IL, Williams DL, Rice PJ, Forsgren A, Riesbeck K. The importance of a beta-glucan receptor in the nonopsonic entry of nontypeable Haemophilus influenzae into human monocytic and epithelial cells. J Infect Dis. 2001;184(2):150-8.

306. Lajoie P, Nabi IR. Lipid rafts, caveolae, and their endocytosis. Int Rev Cell Mol Biol. 2010;282:135-63.

307. Simons K, Gerl MJ. Revitalizing membrane rafts: new tools and insights. Nat Rev Mol Cell Biol. 2010;11(10):688-99.

308. Duncan MJ, Shin JS, Abraham SN. Microbial entry through caveolae: variations on a theme. Cell Microbiol. 2002;4(12):783-91.

309. Conner SD, Schmid SL. Regulated portals of entry into the cell. Nature. 2003;422(6927):37-44.

310. Kaksonen M, Toret CP, Drubin DG. Harnessing actin dynamics for clathrin- mediated endocytosis. Nat Rev Mol Cell Biol. 2006;7(6):404-14.

311. Veiga E, Guttman JA, Bonazzi M, Boucrot E, Toledo-Arana A, Lin AE, et al. Invasive and adherent bacterial pathogens co-Opt host clathrin for infection. Cell Host Microbe. 2007;2(5):340-51.

312. Swords WE, Buscher BA, Ver Steeg Ii K, Preston A, Nichols WA, Weiser JN, et al. Non-typeable Haemophilus influenzae adhere to and invade human bronchial epithelial cells via an interaction of lipooligosaccharide with the PAF receptor. Mol Microbiol. 2000;37(1):13-27.

313. Swords WE, Ketterer MR, Shao J, Campbell CA, Weiser JN, Apicella MA. Binding of the non-typeable Haemophilus influenzae lipooligosaccharide to the PAF receptor initiates host cell signalling. Cell Microbiol. 2001;3(8):525-36.

314. Kerr MC, Teasdale RD. Defining macropinocytosis. Traffic. 2009;10(4):364-71.

315. Schnatwinkel C, Christoforidis S, Lindsay MR, Uttenweiler-Joseph S, Wilm M, Parton RG, et al. The Rab5 effector Rabankyrin-5 regulates and coordinates different endocytic mechanisms. PLoS Biol. 2004;2(9):E261.

316. Hewlett LJ, Prescott AR, Watts C. The coated pit and macropinocytic pathways serve distinct endosome populations. J Cell Biol. 1994;124(5):689-703.

317. Gould GW, Lippincott-Schwartz J. New roles for endosomes: from vesicular carriers to multi-purpose platforms. Nat Rev Mol Cell Biol. 2009;10(4):287-92.

193

318. Taguchi T. Emerging roles of recycling endosomes. J Biochem. 2013;153(6):505- 10.

319. Stoorvogel W, Strous GJ, Geuze HJ, Oorschot V, Schwartz AL. Late endosomes derive from early endosomes by maturation. Cell. 1991;65(3):417-27.

320. Jovic M, Sharma M, Rahajeng J, Caplan S. The early endosome: a busy sorting station for proteins at the crossroads. Histol Histopathol. 2010;25(1):99-112.

321. Sachse M, Urbe S, Oorschot V, Strous GJ, Klumperman J. Bilayered clathrin coats on endosomal vacuoles are involved in protein sorting toward lysosomes. Mol Biol Cell. 2002;13(4):1313-28.

322. De Maziere AM, Muehlethaler K, van Donselaar E, Salvi S, Davoust J, Cerottini JC, et al. The melanocytic protein Melan-A/MART-1 has a subcellular localization distinct from typical melanosomal proteins. Traffic. 2002;3(9):678-93.

323. Giordano F, Simoes S, Raposo G. The ocular albinism type 1 (OA1) GPCR is ubiquitinated and its traffic requires endosomal sorting complex responsible for transport (ESCRT) function. Proc Natl Acad Sci U S A. 2011;108(29):11906-11.

324. Raiborg C, Bache KG, Gillooly DJ, Madshus IH, Stang E, Stenmark H. Hrs sorts ubiquitinated proteins into clathrin-coated microdomains of early endosomes. Nat Cell Biol. 2002;4(5):394-8.

325. Rink J, Ghigo E, Kalaidzidis Y, Zerial M. Rab conversion as a mechanism of progression from early to late endosomes. Cell. 2005;122(5):735-49.

326. Luzio JP, Gray SR, Bright NA. Endosome-lysosome fusion. Biochem Soc Trans. 2010;38(6):1413-6.

327. Bright NA, Gratian MJ, Luzio JP. Endocytic delivery to lysosomes mediated by concurrent fusion and kissing events in living cells. Curr Biol. 2005;15(4):360-5.

328. Luzio JP, Bright NA, Pryor PR. The role of calcium and other ions in sorting and delivery in the late endocytic pathway. Biochem Soc Trans. 2007;35(Pt 5):1088- 91.

329. Weber T, Zemelman BV, McNew JA, Westermann B, Gmachl M, Parlati F, et al. SNAREpins: minimal machinery for membrane fusion. Cell. 1998;92(6):759-72.

330. Pryor PR, Mullock BM, Bright NA, Gray SR, Luzio JP. The role of intraorganellar Ca(2+) in late endosome-lysosome heterotypic fusion and in the reformation of lysosomes from hybrid organelles. J Cell Biol. 2000;149(5):1053-62.

331. Boya P. Lysosomal function and dysfunction: mechanism and disease. Antioxid Redox Signal. 2012;17(5):766-74.

194

332. Saftig P, Klumperman J. Lysosome biogenesis and lysosomal membrane proteins: trafficking meets function. Nat Rev Mol Cell Biol. 2009;10(9):623-35.

333. Fehrenbacher N, Bastholm L, Kirkegaard-Sorensen T, Rafn B, Bottzauw T, Nielsen C, et al. Sensitization to the lysosomal cell death pathway by oncogene- induced down-regulation of lysosome-associated membrane proteins 1 and 2. Cancer Res. 2008;68(16):6623-33.

334. Xu H, Ren D. Lysosomal physiology. Annu Rev Physiol. 2015;77:57-80.

335. Belland RJ, Nelson DE, Virok D, Crane DD, Hogan D, Sturdevant D, et al. Transcriptome analysis of chlamydial growth during IFN-gamma-mediated persistence and reactivation. Proc Natl Acad Sci U S A. 2003;100(26):15971-6.

336. Stein MP, Muller MP, Wandinger-Ness A. Bacterial pathogens commandeer Rab GTPases to establish intracellular niches. Traffic. 2012;13(12):1565-88.

337. Malik-Kale P, Jolly CE, Lathrop S, Winfree S, Luterbach C, Steele-Mortimer O. Salmonella - at home in the host cell. Front Microbiol. 2011;2:125.

338. Harrison RE, Brumell JH, Khandani A, Bucci C, Scott CC, Jiang X, et al. Salmonella impairs RILP recruitment to Rab7 during maturation of invasion vacuoles. Mol Biol Cell. 2004;15(7):3146-54.

339. Hybiske K, Stephens RS. Mechanisms of host cell exit by the intracellular bacterium Chlamydia. Proc Natl Acad Sci U S A. 2007;104(27):11430-5.

340. Fink SL, Cookson BT. Apoptosis, pyroptosis, and necrosis: mechanistic description of dead and dying eukaryotic cells. Infect Immun. 2005;73(4):1907- 16.

341. Ray K, Marteyn B, Sansonetti PJ, Tang CM. Life on the inside: the intracellular lifestyle of cytosolic bacteria. Nat Rev Microbiol. 2009;7(5):333-40.

342. Singh R, Jamieson A, Cresswell P. GILT is a critical host factor for Listeria monocytogenes infection. Nature. 2008;455(7217):1244-7.

343. Poussin MA, Goldfine H. Involvement of Listeria monocytogenes phosphatidylinositol-specific phospholipase C and host protein kinase C in permeabilization of the macrophage phagosome. Infect Immun. 2005;73(7):4410- 3.

344. Golovliov I, Baranov V, Krocova Z, Kovarova H, Sjostedt A. An attenuated strain of the facultative intracellular bacterium Francisella tularensis can escape the phagosome of monocytic cells. Infect Immun. 2003;71(10):5940-50.

345. Clemens DL, Lee BY, Horwitz MA. Virulent and avirulent strains of Francisella tularensis prevent acidification and maturation of their phagosomes and escape into the cytoplasm in human macrophages. Infect Immun. 2004;72(6):3204-17.

195

346. Santic M, Asare R, Skrobonja I, Jones S, Abu Kwaik Y. Acquisition of the vacuolar ATPase proton pump and phagosome acidification are essential for escape of Francisella tularensis into the macrophage cytosol. Infect Immun. 2008;76(6):2671-7.

347. Santic M, Molmeret M, Klose KE, Jones S, Kwaik YA. The Francisella tularensis pathogenicity island protein IglC and its regulator MglA are essential for modulating phagosome biogenesis and subsequent bacterial escape into the cytoplasm. Cell Microbiol. 2005;7(7):969-79.

348. Qin A, Scott DW, Thompson JA, Mann BJ. Identification of an essential Francisella tularensis subsp. tularensis virulence factor. Infect Immun. 2009;77(1):152-61.

349. Checroun C, Wehrly TD, Fischer ER, Hayes SF, Celli J. Autophagy-mediated reentry of Francisella tularensis into the endocytic compartment after cytoplasmic replication. Proc Natl Acad Sci U S A. 2006;103(39):14578-83.

350. Marti-Lliteras P, Regueiro V, Morey P, Hood DW, Saus C, Sauleda J, et al. Nontypeable Haemophilus influenzae clearance by alveolar macrophages is impaired by exposure to cigarette smoke. Infect Immun. 2009;77(10):4232-42.

351. Mulvey MA, Lopez-Boado YS, Wilson CL, Roth R, Parks WC, Heuser J, et al. Induction and evasion of host defenses by type 1-piliated uropathogenic Escherichia coli. Science. 1998;282(5393):1494-7.

352. Anderson GG, Dodson KW, Hooton TM, Hultgren SJ. Intracellular bacterial communities of uropathogenic Escherichia coli in urinary tract pathogenesis. Trends Microbiol. 2004;12(9):424-30.

353. Justice SS, Hung C, Theriot JA, Fletcher DA, Anderson GG, Footer MJ, et al. Differentiation and developmental pathways of uropathogenic Escherichia coli in urinary tract pathogenesis. Proc Natl Acad Sci U S A. 2004;101(5):1333-8.

354. Mysorekar IU, Hultgren SJ. Mechanisms of uropathogenic Escherichia coli persistence and eradication from the urinary tract. Proc Natl Acad Sci U S A. 2006;103(38):14170-5.

355. Justice SS, Lauer SR, Hultgren SJ, Hunstad DA. Maturation of intracellular Escherichia coli communities requires SurA. Infect Immun. 2006;74(8):4793-800.

356. Hunstad DA, Justice SS. Intracellular lifestyles and immune evasion strategies of uropathogenic Escherichia coli. Annu Rev Microbiol. 2010;64:203-21.

357. Schilling JD, Hultgren SJ. Recent advances into the pathogenesis of recurrent urinary tract infections: the bladder as a reservoir for uropathogenic Escherichia coli. Int J Antimicrob Agents. 2002;19(6):457-60.

196

358. Blango MG, Mulvey MA. Persistence of uropathogenic Escherichia coli in the face of multiple antibiotics. Antimicrob Agents Chemother. 2010;54(5):1855-63.

359. Horvath DJ, Jr., Li B, Casper T, Partida-Sanchez S, Hunstad DA, Hultgren SJ, et al. Morphological plasticity promotes resistance to phagocyte killing of uropathogenic Escherichia coli. Microbes Infect. 2011;13(5):426-37.

360. Conover MS, Hadjifrangiskou M, Palermo JJ, Hibbing ME, Dodson KW, Hultgren SJ. Metabolic Requirements of Escherichia coli in Intracellular Bacterial Communities during Urinary Tract Infection Pathogenesis. mBio. 2016;7(2):e00104-16.

361. Justice SS, Harrison A, Becknell B, Mason KM. Bacterial differentiation, development, and disease: mechanisms for survival. FEMS Microbiol Lett. 2014;360(1):1-8.

362. Loughman JA, Hunstad DA. Induction of indoleamine 2,3-dioxygenase by uropathogenic bacteria attenuates innate responses to epithelial infection. J Infect Dis. 2012;205(12):1830-9.

363. Rosen DA, Pinkner JS, Jones JM, Walker JN, Clegg S, Hultgren SJ. Utilization of an intracellular bacterial community pathway in Klebsiella pneumoniae urinary tract infection and the effects of FimK on type 1 pilus expression. Infect Immun. 2008;76(7):3337-45.

364. Schaffer JN, Norsworthy AN, Sun TT, Pearson MM. Proteus mirabilis fimbriae- and urease-dependent clusters assemble in an extracellular niche to initiate bladder stone formation. Proc Natl Acad Sci U S A. 2016;113(16):4494-9.

365. Oh JD, Karam SM, Gordon JI. Intracellular Helicobacter pylori in gastric epithelial progenitors. Proc Natl Acad Sci U S A. 2005;102(14):5186-91.

366. Oh JD, Kling-Backhed H, Giannakis M, Engstrand LG, Gordon JI. Interactions between gastric epithelial stem cells and Helicobacter pylori in the setting of chronic atrophic gastritis. Curr Opin Microbiol. 2006;9(1):21-7.

367. Mulvey MA, Schilling JD, Hultgren SJ. Establishment of a persistent Escherichia coli reservoir during the acute phase of a bladder infection. Infect Immun. 2001;69(7):4572-9.

368. Schwartz DJ, Chen SL, Hultgren SJ, Seed PC. Population dynamics and niche distribution of uropathogenic Escherichia coli during acute and chronic urinary tract infection. Infect Immun. 2011;79(10):4250-9.

369. Murphy TF. Vaccines for Nontypeable Haemophilus influenzae: the Future Is Now. Clin Vaccine Immunol. 2015;22(5):459-66.

370. Ostberg KL, Russell MW, Murphy TF. Mucosal immunization of mice with recombinant OMP P2 induces antibodies that bind to surface epitopes of multiple

197

strains of nontypeable Haemophilus influenzae. Mucosal Immunol. 2009;2(1):63- 73.

371. Forsgren A, Riesbeck K, Janson H. Protein D of Haemophilus influenzae: a protective nontypeable H. influenzae antigen and a carrier for pneumococcal conjugate vaccines. Clin Infect Dis. 2008;46(5):726-31.

372. Novotny LA, Jurcisek JA, Godfroid F, Poolman JT, Denoel PA, Bakaletz LO. Passive immunization with human anti-protein D antibodies induced by polysaccharide protein D conjugates protects chinchillas against otitis media after intranasal challenge with Haemophilus influenzae. Vaccine. 2006;24(22):4804- 11.

373. Novotny LA, Clements JD, Bakaletz LO. Transcutaneous immunization as preventative and therapeutic regimens to protect against experimental otitis media due to nontypeable Haemophilus influenzae. Mucosal Immunol. 2011;4(4):456-67.

374. Kyd JM, Cripps AW, Novotny LA, Bakaletz LO. Efficacy of the 26-kilodalton outer membrane protein and two P5 fimbrin-derived immunogens to induce clearance of nontypeable Haemophilus influenzae from the rat middle ear and lungs as well as from the chinchilla middle ear and nasopharynx. Infect Immun. 2003;71(8):4691-9.

375. Noda K, Kodama S, Umemoto S, Abe N, Hirano T, Suzuki M. Nasal vaccination with P6 outer membrane protein and alpha-galactosylceramide induces nontypeable Haemophilus influenzae-specific protective immunity associated with NKT cell activation and dendritic cell expansion in nasopharynx. Vaccine. 2010;28(31):5068-74.

376. Bakaletz LO, Baker BD, Jurcisek JA, Harrison A, Novotny LA, Bookwalter JE, et al. Demonstration of Type IV pilus expression and a twitching phenotype by Haemophilus influenzae. Infect Immun. 2005;73(3):1635-43.

377. Novotny LA, Adams LD, Kang DR, Wiet GJ, Cai X, Sethi S, et al. Epitope mapping immunodominant regions of the PilA protein of nontypeable Haemophilus influenzae (NTHI) to facilitate the design of two novel chimeric vaccine candidates. Vaccine. 2009;28(1):279-89.

378. Novotny LA, Clements JD, Bakaletz LO. Therapeutic Transcutaneous Immunization with a Band-Aid Vaccine Resolves Experimental Otitis Media. Clin Vaccine Immunol. 2015;22(8):867-74.

379. Novotny LA, Clements JD, Goodman SD, Bakaletz LO. Transcutaneous Immunization with a Band-Aid Prevents Experimental Otitis Media in a Polymicrobial Model. Clin Vaccine Immunol. 2017;24(6).

380. Palmer LD, Skaar EP. Transition Metals and Virulence in Bacteria. Annu Rev Genet. 2016;50:67-91. 198

381. Stojiljkovic I, Perkins-Balding D. Processing of heme and heme-containing proteins by bacteria. DNA and cell biology. 2002;21(4):281-95.

382. Pynnonen M, Stephenson RE, Schwartz K, Hernandez M, Boles BR. Hemoglobin promotes Staphylococcus aureus nasal colonization. PLoS Pathog. 2011;7(7):e1002104.

383. Hariadi NI, Zhang L, Patel M, Sandstedt SA, Davis GS, Marrs CF, et al. Comparative Profile of Heme Acquisition Genes in Disease-Causing and Colonizing Nontypeable Haemophilus influenzae and Haemophilus haemolyticus. J Clin Microbiol. 2015;53(7):2132-7.

384. Whitby PW, VanWagoner TM, Seale TW, Morton DJ, Stull TL. Comparison of transcription of the Haemophilus influenzae iron/heme modulon genes in vitro and in vivo in the chinchilla middle ear. BMC Genomics. 2013;14:925.

385. Duell BL, Su YC, Riesbeck K. Host-pathogen interactions of nontypeable Haemophilus influenzae: from commensal to pathogen. FEBS Lett. 2016;590(21):3840-53.

386. Coates H, Thornton R, Langlands J, Filion P, Keil AD, Vijayasekaran S, et al. The role of chronic infection in children with otitis media with effusion: evidence for intracellular persistence of bacteria. Otolaryngol Head Neck Surg. 2008;138(6):778-81.

387. Forsgren J, Samuelson A, Borrelli S, Christensson B, Jonasson J, Lindberg AA. Persistence of nontypeable Haemophilus influenzae in adenoid macrophages: a putative colonization mechanism. Acta Otolaryngol. 1996;116(5):766-73.

388. Scott VC, Haake DA, Churchill BM, Justice SS, Kim JH. Intracellular Bacterial Communities: A Potential Etiology for Chronic Lower Urinary Tract Symptoms. Urology. 2015;86(3):425-31.

389. Daly KA, Brown JE, Lindgren BR, Meland MH, Le CT, Giebink GS. Epidemiology of otitis media onset by six months of age. Pediatrics. 1999;103(6 Pt 1):1158-66.

390. McCaig LF, Besser RE, Hughes JM. Trends in antimicrobial prescribing rates for children and adolescents. JAMA. 2002;287(23):3096-102.

391. Teele DW, Klein JO, Rosner B. Epidemiology of otitis media during the first seven years of life in children in greater Boston: a prospective, cohort study. J Infect Dis. 1989;160(1):83-94.

392. Novotny LA, Bakaletz LO. The fourth surface-exposed region of the outer membrane protein P5-homologous adhesin of nontypable Haemophilus influenzae is an immunodominant but nonprotective decoying epitope. J Immunol. 2003;171(4):1978-83.

199

393. Novotny LA, Clements JD, Bakaletz LO. Kinetic analysis and evaluation of the mechanisms involved in the resolution of experimental nontypeable Haemophilus influenzae-induced otitis media after transcutaneous immunization. Vaccine. 2013;31(34):3417-26.

394. Tracy E, Ye F, Baker BD, Munson RS, Jr. Construction of non-polar mutants in Haemophilus influenzae using FLP recombinase technology. BMC Mol Biol. 2008;9:101.

395. Barcak GJ, Chandler MS, Redfield RJ, Tomb JF. Genetic systems in Haemophilus influenzae. Methods Enzymol. 1991;204:321-42.

396. Kelly BJ, Fitch JR, Hu Y, Corsmeier DJ, Zhong H, Wetzel AN, et al. Churchill: an ultra-fast, deterministic, highly scalable and balanced parallelization strategy for the discovery of human genetic variation in clinical and population-scale genomics. Genome Biol. 2015;16:6.

397. Biasini M, Bienert S, Waterhouse A, Arnold K, Studer G, Schmidt T, et al. SWISS-MODEL: modelling protein tertiary and quaternary structure using evolutionary information. Nucleic Acids Res. 2014;42(Web Server issue):W252- 8.

398. Laskowski RA, MacArthur MW, Moss DS, Thornton JM. Procheck - A program to check the stereochemicalqQuality of protein structures. J Appl Crystallogr. 1993;26:283-91.

399. Colovos C, Yeates TO. Verification of protein structures: patterns of nonbonded atomic interactions. Protein Sci. 1993;2(9):1511-9.

400. Eisenberg D, Luthy R, Bowie JU. VERIFY3D: assessment of protein models with three-dimensional profiles. Methods Enzymol. 1997;277:396-404.

401. Pontius J, Richelle J, Wodak SJ. Deviations from standard atomic volumes as a quality measure for protein crystal structures. J Mol Biol. 1996;264(1):121-36.

402. Fisher RA, Gollan B, Helaine S. Persistent bacterial infections and persister cells. Nat Rev Microbiol. 2017;15(8):453-64.

403. Finkel SE. Long-term survival during stationary phase: evolution and the GASP phenotype. Nat Rev Microbiol. 2006;4(2):113-20.

404. Weiser JN, Love JM, Moxon ER. The molecular mechanism of phase variation of H. influenzae lipopolysaccharide. Cell. 1989;59(4):657-65.

405. Jackson CJ, Hadler KS, Carr PD, Oakley AJ, Yip S, Schenk G, et al. Malonate- bound structure of the glycerophosphodiesterase from Enterobacter aerogenes (GpdQ) and characterization of the native Fe2+ metal-ion preference. Acta Crystallogr Sect F Struct Biol Cryst Commun. 2008;64(Pt 8):681-5.

200

406. Ramakrishnan K, Sparks RA, Berryhill WE. Diagnosis and treatment of otitis media. Am Fam Physician. 2007;76(11):1650-8.

407. Thornton RB, Rigby PJ, Wiertsema SP, Filion P, Langlands J, Coates HL, et al. Multi-species bacterial biofilm and intracellular infection in otitis media. BMC pediatrics. 2011;11:94.

408. Li B, Smith P, Horvath DJ, Jr., Romesberg FE, Justice SS. SOS regulatory elements are essential for UPEC pathogenesis. Microbes Infect. 2010;12(8- 9):662-8.

409. O'Brien VP, Hannan TJ, Nielsen HV, Hultgren SJ. Drug and Vaccine Development for the Treatment and Prevention of Urinary Tract Infections. Microbiol Spectr. 2016;4(1).

410. Hannan TJ, Roberts PL, Riehl TE, van der Post S, Binkley JM, Schwartz DJ, et al. Inhibition of Cyclooxygenase-2 Prevents Chronic and Recurrent Cystitis. EBioMedicine. 2014;1(1):46-57.

411. Casey JR, Pichichero ME. Changes in frequency and pathogens causing acute otitis media in 1995-2003. Pediatr Infect Dis J. 2004;23(9):824-8.

412. Vega NM, Allison KR, Khalil AS, Collins JJ. Signaling-mediated bacterial persister formation. Nat Chem Biol. 2012;8(5):431-3.

413. Maisonneuve E, Castro-Camargo M, Gerdes K. (p)ppGpp controls bacterial persistence by stochastic induction of toxin-antitoxin activity. Cell. 2013;154(5):1140-50.

414. Amato SM, Orman MA, Brynildsen MP. Metabolic control of persister formation in Escherichia coli. Mol Cell. 2013;50(4):475-87.

415. Helaine S, Cheverton AM, Watson KG, Faure LM, Matthews SA, Holden DW. Internalization of Salmonella by macrophages induces formation of nonreplicating persisters. Science. 2014;343(6167):204-8.

416. Hu Y, Kwan BW, Osbourne DO, Benedik MJ, Wood TK. Toxin YafQ increases persister cell formation by reducing indole signalling. Environ Microbiol. 2015;17(4):1275-85.

417. Blokesch M. In and out-contribution of natural transformation to the shuffling of large genomic regions. Curr Opin Microbiol. 2017;38:22-9.

418. Ibanez de Aldecoa AL, Zafra O, Gonzalez-Pastor JE. Mechanisms and Regulation of Extracellular DNA Release and Its Biological Roles in Microbial Communities. Front Microbiol. 2017;8:1390.

419. Kahn ME, Smith HO. Transformation in Haemophilus: a problem in membrane biology. J Membr Biol. 1984;81(2):89-103.

201

420. Lethem MI, James SL, Marriott C, Burke JF. The origin of DNA associated with mucus glycoproteins in cystic fibrosis sputum. Eur Respir J. 1990;3(1):19-23.

421. Matthews LW, Spector S, Lemm J, Potter JL. Studies on pulmonary secretions. I. The over-all chemical composition of pulmonary secretions from patients with cystic fibrosis, bronchiectasis, and laryngectomy. Am Rev Respir Dis. 1963;88:199-204.

422. Mell JC, Redfield RJ. Natural competence and the evolution of DNA uptake specificity. J Bacteriol. 2014;196(8):1471-83.

423. Chandler MS. The gene encoding cAMP receptor protein is required for competence development in Haemophilus influenzae Rd. Proc Natl Acad Sci U S A. 1992;89(5):1626-30.

424. Cameron AD, Redfield RJ. Non-canonical CRP sites control competence regulons in Escherichia coli and many other gamma-proteobacteria. Nucleic Acids Res. 2006;34(20):6001-14.

425. Cameron AD, Volar M, Bannister LA, Redfield RJ. RNA secondary structure regulates the translation of sxy and competence development in Haemophilus influenzae. Nucleic Acids Res. 2008;36(1):10-20.

426. Molina-Quiroz RC, Silva-Valenzuela C, Brewster J, Castro-Nallar E, Levy SB, Camilli A. Cyclic AMP Regulates Bacterial Persistence through Repression of the Oxidative Stress Response and SOS-Dependent DNA Repair in Uropathogenic Escherichia coli. mBio. 2018;9(1).

427. Kwan BW, Osbourne DO, Hu Y, Benedik MJ, Wood TK. Phosphodiesterase DosP increases persistence by reducing cAMP which reduces the signal indole. Biotechnol Bioeng. 2015;112(3):588-600.

428. Duraiswamy S, Chee JLY, Chen S, Yang E, Lees K, Chen SL. Purification of intracellular bacterial communities during experimental urinary tract infection reveals an abundant and viable bacterial reservoir. Infect Immun. 2018.

429. Murphy TF. Respiratory infections caused by non-typeable Haemophilus influenzae. Curr Opin Infect Dis. 2003;16(2):129-34.

430. Teweldemedhin M, Gebreyesus H, Atsbaha AH, Asgedom SW, Saravanan M. Bacterial profile of ocular infections: a systematic review. BMC Ophthalmol. 2017;17(1):212.

431. Leibovitz E, Jacobs MR, Dagan R. Haemophilus influenzae: a significant pathogen in acute otitis media. Pediatr Infect Dis J. 2004;23(12):1142-52.

432. Murphy TF, Faden H, Bakaletz LO, Kyd JM, Forsgren A, Campos J, et al. Nontypeable Haemophilus influenzae as a pathogen in children. Pediatr Infect Dis J. 2009;28(1):43-8.

202

433. Groeneveld K, Eijk PP, van Alphen L, Jansen HM, Zanen HC. Haemophilus influenzae infections in patients with chronic obstructive pulmonary disease despite specific antibodies in serum and sputum. Am Rev Respir Dis. 1990;141(5 Pt 1):1316-21.

434. Moller LV, Regelink AG, Grasselier H, Dankert-Roelse JE, Dankert J, van Alphen L. Multiple Haemophilus influenzae strains and strain variants coexist in the respiratory tract of patients with cystic fibrosis. J Infect Dis. 1995;172(5):1388-92.

435. Torres VJ, Attia AS, Mason WJ, Hood MI, Corbin BD, Beasley FC, et al. Staphylococcus aureus fur regulates the expression of virulence factors that contribute to the pathogenesis of pneumonia. Infect Immun. 2010;78(4):1618-28.

436. Mittal R, Sharma S, Chhibber S, Harjai K. Iron dictates the virulence of Pseudomonas aeruginosa in urinary tract infections. J Biomed Sci. 2008;15(6):731-41.

437. Hotomi M, Arai J, Billal DS, Takei S, Ikeda Y, Ogami M, et al. Nontypeable Haemophilus influenzae isolated from intractable acute otitis media internalized into cultured human epithelial cells. Auris Nasus Larynx. 2010;37(2):137-44.

438. Nakamura A, DeMaria TF, Lim DJ, van Blitterswijk CA. Primary culture of chinchilla middle ear epithelium. Ann Otol Rhinol Laryngol. 1991;100(9 Pt 1):774- 82.

439. Edwards JL. Neisseria gonorrhoeae survival during primary human cervical epithelial cell infection requires nitric oxide and is augmented by progesterone. Infect Immun. 2010;78(3):1202-13.

440. Shoji K, Ohashi K, Sampei K, Oikawa M, Mizuno K. Cytochalasin D acts as an inhibitor of the actin-cofilin interaction. Biochem Biophys Res Commun. 2012;424(1):52-7.

441. Nesvizhskii AI, Keller A, Kolker E, Aebersold R. A statistical model for identifying proteins by tandem mass spectrometry. Anal Chem. 2003;75(17):4646-58.

442. Simonsen A, Lippe R, Christoforidis S, Gaullier JM, Brech A, Callaghan J, et al. EEA1 links PI(3)K function to Rab5 regulation of endosome fusion. Nature. 1998;394(6692):494-8.

443. Gruenberg J. The endocytic pathway: a mosaic of domains. Nat Rev Mol Cell Biol. 2001;2(10):721-30.

444. Euba B, Lopez-Lopez N, Rodriguez-Arce I, Fernandez-Calvet A, Barberan M, Caturla N, et al. Resveratrol therapeutics combines both antimicrobial and immunomodulatory properties against respiratory infection by nontypeable Haemophilus influenzae. Sci Rep. 2017;7(1):12860.

203

445. Robino L, Scavone P, Araujo L, Algorta G, Zunino P, Pirez MC, et al. Intracellular bacteria in the pathogenesis of Escherichia coli urinary tract infection in children. Clin Infect Dis. 2014;59(11):e158-64.

446. Robino L, Scavone P, Araujo L, Algorta G, Zunino P, Vignoli R. Detection of intracellular bacterial communities in a child with Escherichia coli recurrent urinary tract infections. Pathog Dis. 2013;68(3):78-81.

447. Cheng Y, Chen Z, Gawthorne JA, Mukerjee C, Varettas K, Mansfield KJ, et al. Detection of intracellular bacteria in exfoliated urothelial cells from women with urge incontinence. Pathog Dis. 2016;74(7).

448. Rosen DA, Hooton TM, Stamm WE, Humphrey PA, Hultgren SJ. Detection of intracellular bacterial communities in human urinary tract infection. PLoS Med. 2007;4(12):e329.

449. Kim JH, Singh A, Del Poeta M, Brown DA, London E. The effect of sterol structure upon clathrin-mediated and clathrin-independent endocytosis. J Cell Sci. 2017;130(16):2682-95.

450. O'Riordan M, Moors MA, Portnoy DA. Listeria intracellular growth and virulence require host-derived lipoic acid. Science. 2003;302(5644):462-4.

451. Keeney K, Colosi L, Weber W, O'Riordan M. Generation of branched-chain fatty acids through lipoate-dependent metabolism facilitates intracellular growth of Listeria monocytogenes. J Bacteriol. 2009;191(7):2187-96.

452. Pilatz S, Breitbach K, Hein N, Fehlhaber B, Schulze J, Brenneke B, et al. Identification of Burkholderia pseudomallei genes required for the intracellular life cycle and in vivo virulence. Infect Immun. 2006;74(6):3576-86.

453. Ramaswamy AV, Maurelli AT. Chlamydia trachomatis serovar L2 can utilize exogenous lipoic acid through the action of the lipoic acid ligase LplA1. J Bacteriol. 2010;192(23):6172-81.

454. Wang WY, Lim JH, Li JD. Synergistic and feedback signaling mechanisms in the regulation of inflammation in respiratory infections. Cell Mol Immunol. 2012;9(2):131-5.

455. Kagnoff MF, Eckmann L. Epithelial cells as sensors for microbial infection. J Clin Invest. 1997;100(1):6-10.

456. Moon SK, Lee HY, Pan H, Takeshita T, Park R, Cha K, et al. Synergistic effect of interleukin 1 alpha on nontypeable Haemophilus influenzae-induced upregulation of human beta-defensin 2 in middle ear epithelial cells. BMC Infect Dis. 2006;6:12.

204

457. Watanabe T, Jono H, Han J, Lim DJ, Li JD. Synergistic activation of NF-kappaB by nontypeable Haemophilus influenzae and tumor necrosis factor alpha. Proc Natl Acad Sci U S A. 2004;101(10):3563-8.

458. Hamilton JA. Colony-stimulating factors in inflammation and autoimmunity. Nat Rev Immunol. 2008;8(7):533-44.

459. Yellon RF, Leonard G, Marucha P, Sidman J, Carpenter R, Burleson J, et al. Demonstration of interleukin 6 in middle ear effusions. Arch Otolaryngol Head Neck Surg. 1992;118(7):745-8.

460. Yellon RF, Leonard G, Marucha PT, Craven R, Carpenter RJ, Lehmann WB, et al. Characterization of cytokines present in middle ear effusions. Laryngoscope. 1991;101(2):165-9.

461. Fleetwood AJ, Cook AD, Hamilton JA. Functions of granulocyte-macrophage colony-stimulating factor. Crit Rev Immunol. 2005;25(5):405-28.

462. Saba S, Soong G, Greenberg S, Prince A. Bacterial stimulation of epithelial G- CSF and GM-CSF expression promotes PMN survival in CF airways. Am J Respir Cell Mol Biol. 2002;27(5):561-7.

463. Kariya S, Okano M, Higaki T, Makihara S, Haruna T, Eguchi M, et al. Neutralizing antibody against granulocyte/macrophage colony-stimulating factor inhibits inflammatory response in experimental otitis media. Laryngoscope. 2013;123(6):1514-8.

464. Dorer MS, Cohen IE, Sessler TH, Fero J, Salama NR. Natural competence promotes Helicobacter pylori chronic infection. Infect Immun. 2013;81(1):209-15.

465. Chan WT, Moreno-Cordoba I, Yeo CC, Espinosa M. Toxin-antitoxin genes of the Gram-positive pathogen Streptococcus pneumoniae: so few and yet so many. Microbiol Mol Biol Rev. 2012;76(4):773-91.

466. Spencer HT, Herriott RM. Development of competence of Haemophilus influenzae. J Bacteriol. 1965;90(4):911-20.

467. Justice SS, Hunstad DA, Seed PC, Hultgren SJ. Filamentation by Escherichia coli subverts innate defenses during urinary tract infection. Proc Natl Acad Sci U S A. 2006;103(52):19884-9.

468. Tan K, Moreno-Hagelsieb G, Collado-Vides J, Stormo GD. A comparative genomics approach to prediction of new members of regulons. Genome Res. 2001;11(4):566-84.

469. Hamilton S, Bongaerts RJ, Mulholland F, Cochrane B, Porter J, Lucchini S, et al. The transcriptional programme of Salmonella enterica serovar Typhimurium reveals a key role for tryptophan metabolism in biofilms. BMC Genomics. 2009;10:599.

205

470. Domka J, Lee J, Bansal T, Wood TK. Temporal gene-expression in Escherichia coli K-12 biofilms. Environ Microbiol. 2007;9(2):332-46.

471. Gong F, Yanofsky C. Reproducing tna operon regulation in vitro in an S-30 system. Tryptophan induction inhibits cleavage of TnaC peptidyl-tRNA. J Biol Chem. 2001;276(3):1974-83.

472. Schroeder GN, Hilbi H. Molecular pathogenesis of Shigella spp.: controlling host cell signaling, invasion, and death by type III secretion. Clin Microbiol Rev. 2008;21(1):134-56.

473. Shin S, Roy CR. Host cell processes that influence the intracellular survival of Legionella pneumophila. Cell Microbiol. 2008;10(6):1209-20.

474. Haraga A, Ohlson MB, Miller SI. Salmonellae interplay with host cells. Nat Rev Microbiol. 2008;6(1):53-66.

475. Garcia-Perez BE, Hernandez-Gonzalez JC, Garcia-Nieto S, Luna-Herrera J. Internalization of a non-pathogenic mycobacteria by macropinocytosis in human alveolar epithelial A549 cells. Microb Pathog. 2008;45(1):1-6.

476. Mercer J, Helenius A. Gulping rather than sipping: macropinocytosis as a way of virus entry. Curr Opin Microbiol. 2012;15(4):490-9.

477. Krzyzaniak MA, Zumstein MT, Gerez JA, Picotti P, Helenius A. Host cell entry of respiratory syncytial virus involves macropinocytosis followed by proteolytic activation of the F protein. PLoS Pathog. 2013;9(4):e1003309.

478. Mercer J, Helenius A. Vaccinia virus uses macropinocytosis and apoptotic mimicry to enter host cells. Science. 2008;320(5875):531-5.

479. Freeman MC, Peek CT, Becker MM, Smith EC, Denison MR. Coronaviruses induce entry-independent, continuous macropinocytosis. MBio. 2014;5(4):e01340-14.

480. Wang JT, Kerr MC, Karunaratne S, Jeanes A, Yap AS, Teasdale RD. The SNX- PX-BAR family in macropinocytosis: the regulation of macropinosome formation by SNX-PX-BAR proteins. PLoS One. 2010;5(10):e13763.

481. Malyukova I, Murray KF, Zhu C, Boedeker E, Kane A, Patterson K, et al. Macropinocytosis in Shiga toxin 1 uptake by human intestinal epithelial cells and transcellular transcytosis. Am J Physiol Gastrointest Liver Physiol. 2009;296(1):G78-92.

482. Yoshida S, Gaeta I, Pacitto R, Krienke L, Alge O, Gregorka B, et al. Differential signaling during macropinocytosis in response to M-CSF and PMA in macrophages. Front Physiol. 2015;6:8.

206

483. Hunstad DA, Justice SS, Hung CS, Lauer SR, Hultgren SJ. Suppression of bladder epithelial cytokine responses by uropathogenic Escherichia coli. Infect Immun. 2005;73(7):3999-4006.

484. Baxt LA, Garza-Mayers AC, Goldberg MB. Bacterial subversion of host innate immune pathways. Science. 2013;340(6133):697-701.

485. Diacovich L, Gorvel JP. Bacterial manipulation of innate immunity to promote infection. Nat Rev Microbiol. 2010;8(2):117-28.

486. Liss V, Hensel M. Take the tube: remodelling of the endosomal system by intracellular Salmonella enterica. Cell Microbiol. 2015;17(5):639-47.

487. Pu J, Guardia CM, Keren-Kaplan T, Bonifacino JS. Mechanisms and functions of lysosome positioning. J Cell Sci. 2016;129(23):4329-39.

488. Bucci C, Thomsen P, Nicoziani P, McCarthy J, van Deurs B. Rab7: a key to lysosome biogenesis. Mol Biol Cell. 2000;11(2):467-80.

489. Chen L, Hu J, Yun Y, Wang T. Rab36 regulates the spatial distribution of late endosomes and lysosomes through a similar mechanism to Rab34. Mol Membr Biol. 2010;27(1):23-30.

490. Kirkegaard K, Taylor MP, Jackson WT. Cellular autophagy: surrender, avoidance and subversion by microorganisms. Nat Rev Microbiol. 2004;2(4):301-14.

491. Deretic V, Levine B. Autophagy, immunity, and microbial adaptations. Cell Host Microbe. 2009;5(6):527-49.

492. Espinoza-Mellado Mdel R, Reyes-Picaso C, Garces-Perez MS, Jardon-Serrano CV, Lopez-Villegas EO, Giono-Cerezo S. Haemophilus influenzae triggers autophagy in HEp-2 cells. Arch Microbiol. 2016;198(2):199-204.

493. Goetz M, Bubert A, Wang G, Chico-Calero I, Vazquez-Boland JA, Beck M, et al. Microinjection and growth of bacteria in the cytosol of mammalian host cells. Proc Natl Acad Sci U S A. 2001;98(21):12221-6.

494. Joseph B, Mertins S, Stoll R, Schar J, Umesha KR, Luo Q, et al. Glycerol metabolism and PrfA activity in Listeria monocytogenes. J Bacteriol. 2008;190(15):5412-30.

495. Austin FE, Turco J, Winkler HH. Rickettsia prowazekii requires host cell serine and glycine for growth. Infect Immun. 1987;55(1):240-4.

496. Austin FE, Winkler HH. Proline incorporation into protein by Rickettsia prowazekii during growth in Chinese hamster ovary (CHO-K1) cells. Infect Immun. 1988;56(12):3167-72.

207

497. Renesto P, Ogata H, Audic S, Claverie JM, Raoult D. Some lessons from Rickettsia genomics. FEMS Microbiol Rev. 2005;29(1):99-117.

498. Noriega FR, Losonsky G, Lauderbaugh C, Liao FM, Wang JY, Levine MM. Engineered deltaguaB-A deltavirG Shigella flexneri 2a strain CVD 1205: construction, safety, immunogenicity, and potential efficacy as a mucosal vaccine. Infect Immun. 1996;64(8):3055-61.

499. Cersini A, Salvia AM, Bernardini ML. Intracellular multiplication and virulence of Shigella flexneri auxotrophic mutants. Infect Immun. 1998;66(2):549-57.

500. Runyen-Janecky LJ, Payne SM. Identification of chromosomal Shigella flexneri genes induced by the eukaryotic intracellular environment. Infect Immun. 2002;70(8):4379-88.

501. Chico-Calero I, Suarez M, Gonzalez-Zorn B, Scortti M, Slaghuis J, Goebel W, et al. Hpt, a bacterial homolog of the microsomal glucose- 6-phosphate translocase, mediates rapid intracellular proliferation in Listeria. Proc Natl Acad Sci U S A. 2002;99(1):431-6.

502. Eylert E, Schar J, Mertins S, Stoll R, Bacher A, Goebel W, et al. Carbon metabolism of Listeria monocytogenes growing inside macrophages. Mol Microbiol. 2008;69(4):1008-17.

503. Joseph B, Przybilla K, Stuhler C, Schauer K, Slaghuis J, Fuchs TM, et al. Identification of Listeria monocytogenes genes contributing to intracellular replication by expression profiling and mutant screening. J Bacteriol. 2006;188(2):556-68.

504. Cecchini G, Schroder I, Gunsalus RP, Maklashina E. Succinate dehydrogenase and fumarate reductase from Escherichia coli. Biochim Biophys Acta. 2002;1553(1-2):140-57.

505. Maklashina E, Cecchini G. The quinone-binding and catalytic site of complex II. Biochim Biophys Acta. 2010;1797(12):1877-82.

506. Tang H, Rothery RA, Voss JE, Weiner JH. Correct assembly of iron-sulfur cluster FS0 into Escherichia coli dimethyl sulfoxide reductase (DmsABC) is a prerequisite for molybdenum cofactor insertion. J Biol Chem. 2011;286(17):15147-54.

507. Darwin A, Hussain H, Griffiths L, Grove J, Sambongi Y, Busby S, et al. Regulation and sequence of the structural gene for cytochrome c552 from Escherichia coli: not a hexahaem but a 50 kDa tetrahaem nitrite reductase. Mol Microbiol. 1993;9(6):1255-65.

508. Whitby PW, Seale TW, Morton DJ, VanWagoner TM, Stull TL. Characterization of the Haemophilus influenzae tehB gene and its role in virulence. Microbiology. 2010;156(Pt 4):1188-200. 208

509. Santos LS, Antunes CA, Santos CS, Pereira JA, Sabbadini PS, Luna M, et al. Corynebacterium diphtheriae putative tellurite-resistance protein (CDCE8392_0813) contributes to the intracellular survival in human epithelial cells and lethality of Caenorhabditis elegans. Mem Inst Oswaldo Cruz. 2015;110(5):662-8.

510. Ponnusamy D, Hartson SD, Clinkenbeard KD. Intracellular Yersinia pestis expresses general stress response and tellurite resistance proteins in mouse macrophages. Vet Microbiol. 2011;150(1-2):146-51.

511. Fani R, Mori E, Tamburini E, Lazcano A. Evolution of the structure and chromosomal distribution of histidine biosynthetic genes. Orig Life Evol Biosph. 1998;28(4-6):555-70.

512. Juliao PC, Marrs CF, Xie J, Gilsdorf JR. Histidine auxotrophy in commensal and disease-causing nontypeable Haemophilus influenzae. J Bacteriol. 2007;189(14):4994-5001.

513. Brenner M, Ames BN. Histidine regulation in Salmonella typhimurium. IX. Histidine transfer ribonucleic acid of the regulatory mutants. J Biol Chem. 1972;247(4):1080-8.

514. Sporn LA, Sahni SK, Lerner NB, Marder VJ, Silverman DJ, Turpin LC, et al. Rickettsia rickettsii infection of cultured human endothelial cells induces NF- kappaB activation. Infect Immun. 1997;65(7):2786-91.

515. Hauf N, Goebel W, Fiedler F, Sokolovic Z, Kuhn M. Listeria monocytogenes infection of P388D1 macrophages results in a biphasic NF-kappaB (RelA/p50) activation induced by lipoteichoic acid and bacterial phospholipases and mediated by IkappaBalpha and IkappaBbeta degradation. Proc Natl Acad Sci U S A. 1997;94(17):9394-9.

516. Clifton DR, Goss RA, Sahni SK, van Antwerp D, Baggs RB, Marder VJ, et al. NF- kappa B-dependent inhibition of apoptosis is essential for host cellsurvival during Rickettsia rickettsii infection. Proc Natl Acad Sci U S A. 1998;95(8):4646-51.

517. Garmendia J, Viadas C, Calatayud L, Mell JC, Marti-Lliteras P, Euba B, et al. Characterization of nontypable Haemophilus influenzae isolates recovered from adult patients with underlying chronic lung disease reveals genotypic and phenotypic traits associated with persistent infection. PLoS One. 2014;9(5):e97020.

518. Oliver A. Mutators in cystic fibrosis chronic lung infection: Prevalence, mechanisms, and consequences for antimicrobial therapy. Int J Med Microbiol. 2010;300(8):563-72.

519. Pettigrew MM, Ahearn CP, Gent JF, Kong Y, Gallo MC, Munro JB, et al. Haemophilus influenzae genome evolution during persistence in the human

209

airways in chronic obstructive pulmonary disease. Proc Natl Acad Sci U S A. 2018;115(14):E3256-E65.

520. Hogardt M, Heesemann J. Microevolution of Pseudomonas aeruginosa to a chronic pathogen of the cystic fibrosis lung. Curr Top Microbiol Immunol. 2013;358:91-118.

521. Hogardt M, Heesemann J. Adaptation of Pseudomonas aeruginosa during persistence in the cystic fibrosis lung. Int J Med Microbiol. 2010;300(8):557-62.

522. Bianconi I, Milani A, Cigana C, Paroni M, Levesque RC, Bertoni G, et al. Positive signature-tagged mutagenesis in Pseudomonas aeruginosa: tracking patho- adaptive mutations promoting airways chronic infection. PLoS Pathog. 2011;7(2):e1001270.

210

APPENDIX A. Significantly altered proteins in RM33 biofilms.

NTHI gene # Acronym Function Fold p-value Change NTHI1768 TrpE anthranilate synthase 9.1 0.0000003 component I NTHI1763 TrpC bifunctional indole-3-glycerol 6.9 0.00012 phosphate synthase/phosphoribosylanthran ilate isomerase NTHI1702 TrpB tryptophan synthase subunit 6.9 0.00004 beta NTHI1810 MalQ alpha-glucanotransferase 5.3 0.00001 NTHI1701 TrpA tryptophan synthase subunit 4.8 0.0002 alpha

NTHI1920 Mao2 malic enzyme 4.3 0.000003 NTHI0202 HemR hemin receptor 4.3 0.0004 NTHI0623 AphA acid phosphatase/ 4.0 0.000003 phosphotransferase

NTHI0632 RbsB D-ribose transporter subunit 3.8 0.000002 RbsB

NTHI0232 sialic acid transporter, TRAP- 3.6 0.00003 type C4-dicarboxylate transport system, periplasmic component

NTHI1707 ABC transporter periplasmic 3.5 0.00001 protein

NTHI0398 SdaC serine transporter 3.1 0.00002 NTHI1964 SucA oxoglutarate dehydrogenase E1 3.0 0.00001

NTHI0397 SdaA L-serine dehydratase 3.0 0.000005

Table A.1. List of proteins that are significantly altered in biofilms formed by RM33 compared to the parent 86-028NP at 48 hours (continued on page 212).

211

NTHI gene # Acronym Function Fold p-value Change

NTHI1809 GlgB glycogen branching enzyme 3.0 0.00005

NTHI0235 acetylneuraminic acid 2.8 0.00003 mutarotase

NTHI1764 TrpD anthranilate 2.8 0.0001 phosphoribosyltransferase

NTHI1295 carbon starvation protein, 2.7 0.0001 membrane protein

NTHI1806 GlgA glycogen synthase 2.5 0.0002 NTHI0647 hypothetical protein 2.4 0.00005 NTHI0660 aspartate ammonia-lyase 2.3 0.00004 NTHI1808 GlgX glycogen operon protein 2.2 0.00004 NTHI1816 Cdd cytidine deaminase 2.2 0.0004 NTHI1903 folylpolyglutamate synthase 2.1 0.0006 NTHI0610 AtpG ATP synthase F0F1 subunit 2.1 0.0002 gamma

NTHI0973 phosphoenolpyruvate 2.0 0.0002 carboxykinase

NTHI0088 NrdD anaerobic ribonucleoside 2.0 0.0002 triphosphate reductase NTHI1028 ClpB ClpB chaperone -2.1 0.00003 NTHI0306 AroB 3-dehydroquinate synthase -3.4 0.00006

Table A.1 (cont’d). List of proteins that are significantly altered in biofilms formed by RM33 compared to the parent 86-028NP at 48 hours (continued from page 211).

212

APPENDIX B. Mutations acquired by NTHI following passaging through the chinchilla.

NTHI Gene Function Mutation Gene # Name NTHI1668 Unannotated 23SrRNA Transcriptional phase methyltransferase variation NTHI1450 hmw2A High molecular weight adhesin 2

NTHI0365 lgtC UDP-galactose—lipooligosaccharide Translational phase galactosyltransferase variation NTHI0585 lav Autotransported protein

NTHI0677 lic2A UDP-galactose-lipooligosaccharide galactosyltransferase NTHI1034 lic3A2 CMP-neu5Ac-lipooligosaccharide alpha 2-3 sialyltransferase NTHI0782 hgpB Hemoglobin-haptoglobin binding protein B NTHI0736 hgpD Hemoglobin-haptoglobin binding protein D NTHI0512 acyltransferase

NTHI1983 hmw1A High molecular weight adhesin 1

NTHI1750 Glycosyl transferase family protein

NTHI0840 hgpC Hemoglobin-haptoglobin binding Intragenic snp protein C NTHI0926 lpsA Lipooligosaccharide glycosyl Intragenic snp; transferase insertion NTHI1474 lgtD UPD-GlcNAc—lipooligosaccharide N- Intragenic snp acetylglycosamine glycosyltransferase

NTHI0580 Hypothetical protein Promoter region snp

NTHI1667 Hypothetical protein Deletion

Table B.1. List of mutations identified in NTHI after four subsequent passages through the chinchilla.

213

APPENDIX C. Additional Methods

Long-term culture of transiently restricted or continuously exposed RM33 (Figure

5.1)

NTHI strain RM33 or the parental strain 86-028NP were transiently restricted or continuously exposed to heme-iron as described in Chapter 2, Section 2.2 (Methods).

Transiently restricted or continuously exposed NTHI cultures were then adjusted to an

OD490 of 0.05 in 5 ml DIS medium containing 2 µg/mL heme for growth at 37°C under static conditions. To evaluate the role of co-infection on long-term survival in stationary phase, co-cultures were established at a goal ratio of 1:1. To distinguish the two strains in co-culture, 86-028NP and RM33 were engineered to carry antibiotic resistance markers. Both 86-028NP and RM33 were transformed with pGZRS-39A, a

Haemophilus-Actinobacillus pleuropneumoniae shuttle vector that contains a kanamycin resistance gene; or pSPEC1, a variant of pGZRS-39A in which the kanamycin resistance gene was replaced by a spectinomycin resistance gene. Every 24 hours, viability was determined by serially diluting and plating in triplicate on chocolate agar plates with or without antibiotic for selection.

Biofilm growth on abiotic surface (Figure 5.2)

NTHI strains 86-028NP and RM33 were grown overnight on chocolate II agar and then subcultured into pre-warmed brain heart infusion broth supplemented with 2 μg heme/mL and 1 μg NAD/mL (Becton Dickinson; sBHI). Cultures were normalized to an optical density at 490 nm (OD490) of 0.65, diluted 1:6 in 5mL sBHI in round bottom, 15mL tubes, and incubated for 3 hours, statically, at 37°C and 5% CO2. When in exponential phase (OD490 ~0.65), cultures were diluted 1:2500 in sBHI and 8 well chamber slides

(Thermo Fisher Scientific, Waltham MA) were inoculated with 200 µL of bacteria added 214 per well and incubated at 37°C, 5% C02. After 24 hours, spent medium was aspirated from each well and replaced with 200 uL/well fresh, pre-warmed sBHI. At 48 hours, biofilms were washed twice with 200 uL/well 1X DPBS and stained with BacLiteTM

LIVE/DEAD stain (Molecular Probes, Grand Island, NY). Biofilm structure and organization were imaged with an Axiovert 200M inverted epifluorescence microscope equipped with the Apotome attachment for improved fluorescence resolution and an

Axiocam MRM CCD camera (Carl Zeiss Inc. Thornwood, NY). Three-dimensional renderings were performed to generate orthogonal views and surface plots. Biofilm heights were measured from twenty random fields of view in five independent experiments.

For co-culture of 86-028NP and RM33 in a biofilm, green fluorescent protein

(GFP) and mCherry reporter strains were constructed by electroporation of pGM1.1 or pKM1.1, respectively, as published previously (60). Exponential phase cultures of 86-

028NP/pGM1.1 and RM33/pKM1.1 were mixed at a 1:1 ratio and 200 uL of the mixed culture was added to each well of a Nunc Lak-Tek 8-well chamber slide. The chamber slide was incubated for a total of 48 hours, fixed, and imaged as described for single strain biofilms.

Growth and preparation of biofilms for proteomic analysis (Appendix A)

48 hour biofilms of strains NTHI 86-028NP and RM33 were grown in 8 well chamber slides as described above. At 24 hours post-inoculation, supernatants were collected from each well and pooled for each strain. The 24-hour supernatants were centrifuged at 16,000 rpm for 5 minutes to remove any bacterial cells and the supernatants were transferred into new microcentrifuge tubes and stored at -80°C. 48

215 hours post-inoculation, the biofilm supernatants were collected from each well and pooled for each strain. Supernatants were centrifuged, transferred to new microcentrifuge tubes, and transferred to -80°C. 48-hour biofilms were then washed once with 200 uL of 50 mM ammonium bicarbonate (pH 8.0) before being resuspended in 200 uL 50 mM ammonium bicarbonate. Mini cell scrapers were used to ensure collection of entire biofilm from each well. Biofilms from all 8 wells were pooled for each strain and each sample was subjected to two passages through a high-pressure cell

(20,000 psi; One Shot Model, Constant Systems, Ltd., Kennesaw, GA). Protein concentrations for each sample were then determined using a Bradford’s assay using a

Pierce Coomassie (Bradford) Protein Assay Kit (Thermo Fisher Scientific) and samples were stored at -80oC prior to analyses. Three biological replicates were processed for each biofilm.

Proteomics of NTHI biofilms (Appendix A)

Label-free Quantitative Analysis of NTHI biofilm proteins. Quantitative liquid chromatography – tandem mass spectrometry (LC-MS/MS) was performed in singlicate on 0.3 µg of protein digest per sample, and the pool was analyzed 4 times with 0.3 µg injections spaced evenly across the run queue using the acquisition method immediately described below. The method used liquid chromatography on a nanoACQUITY UPLC system (Waters Corp) coupled to a Q-Exactive Plus high resolution accurate mass tandem mass spectrometer (Thermo) with nanoelectrospray ionization. Mobile phase A and B consisted of 0.1% formic acid (v/v) in water and acetonitrile, respectively.

Peptides were first trapped at 5 µl/min, 99.9%A, on a 5 µm Symmetry C18 180 µm × 20 mm trapping column (Waters). Analytical separations were performed on a 1.8 µm

Acquity HSS T3 C18 75 µm × 250 mm column (Waters) using a linear gradient of 5 to 216

40% B over 90 minutes, at a flow rate of 0.4 µl/min and column temperature of 55 °C.

Data collection on the Q-Exactive Plus mass spectrometer were performed in data- dependent acquisition (DDA) mode of acquisition with Rs = 70,000 (@ m/z 200) full MS scan from m/z 375 to 1600 with a target AGC value of 1e6 ions followed by 10 MS/MS scans at Rs = 17,500 (@ m/z 200) at a target AGC value of 5e4 ions. A 20 s dynamic exclusion was employed.

Analysis of proteins. Following the 16 analyses, data were imported into Rosetta

Elucidator v3.3 (Rosetta Biosoftware, Inc), and all LC-MS runs were aligned based on the accurate mass and retention time of detected ions (“features”) using PeakTeller algorithm (Elucidator). The relative peptide abundance was calculated based on area- under-the-curve (AUC) of aligned features across all runs. The MS/MS data was searched against a custom database containing NCBI Haemophilus influenzae sequences; the database also contained a reversed-sequence “decoy” database for false positive rate determination as well as several proteins which were surrogate standards or common contaminants (database is available at https://discovery.genome.duke.edu/express/resources/3697/chinch_86028np_mix1_041

614_reverse.fasta). Included in the database searches were variable modifications on M

(oxidation) and N/Q (deamination), and fixed modification on C (carbamidomethyl). After individual peptide scoring using PeptideProphet algorithm (Elucidator), the data was annotated at a 1% peptide false discovery rate. This analysis yielded identifications for

6,213 peptides and 1,207 proteins across all samples. 769 proteins were identified and quantified with 2+ peptides to match. With respect to species specificity, only 284 peptides (to 245 unique proteins) out of 6213 total peptides (<5%) were identified to chinchilla, which would be expected in this cultured NTHI system. For quantitative

217 processing, filters were applied to ensure peptides had good chromatographic peak shape and good MS spectral integrity. For this dataset, no peptides were removed and the final quantitative dataset contained 6,213 peptides and 1,207 proteins.

Quantification of indole produced by NTHI biofilms (Figure 5.3)

Extracellular indole concentration was measured using a modified version of an

Kovac’s assay for indoles. Bacterial strains were grown to mid-logarithmic phase (OD490) in sBHI medium at 37C with 5% CO2. Cultures were diluted 1:2500 and 200 uL placed in each well of three 8-well chamber slides per strain. Following incubation at 37C with 5%

CO2 for 24 or 48 hours, biofilm supernatants from each well were collected and pooled for each strain. A media change was performed at 24 hours for biofilms being incubated

48 hours. To determine indole concentration, 2.5 mL biofilm supernatant was mixed with

1 mL of HCl:n-Butanol (50 mL HCl + 150 mL n-Butanol) and allowed to become biphasic. 150 uL of the organic phase (containing extracted indole) was removed and placed in quadruplicate in a clear, flat bottom 96 well plate. 60 uL Kovac’s Reagent

(Sigma Aldrich) was then added to each well and incubated for 2 minutes. A replicate without Kovac’s reagent was included as a control for each sample. Absorbance at 570 nm was immediately read using a spectrophotometer. To quantify indole concentration, a standard curve was created as follows. 1 mg Indole (Sigma Aldrich) was dissolved in 1 mL HCl:n-Butanol in glass tubes and serially diluted to generate a standard curve ranging from 1 to 0 mg Indole. 2.5 mL freshly prepared sBHI was added to each tube and the mixture was allowed to become biphasic. 150 uL of the organic phase was removed and placed in quadruplicate in a clear, flat bottom 96 well plate. 60 uL Kovac’s reagent was added to each well and incubated for 2 minutes. A replicate without Kovac’s reagent was included as a control for each standard concentration. Absorbance at 570 218 was immediately read using a spectrophotometer. Indole concentration in experimental samples was determined using a linear trend line based on the standard curve.

Experiments were performed in biological and technical triplicate.

Animal study: Passaging of NTHI in an animal model of OM (Appendix B)

Ethics statement. All animal experiments were carried out in strict accordance with the accredited conditions in the Guide for the Care and Use of Laboratory Animals of the

National Institutes of Health. The protocol was approved by the Institutional Animal Care and Use Committee at The Research Institute at Nationwide Children’s Hospital. All experimental procedures were performed under xylazine and ketamine anesthesia, and all efforts were made to minimize suffering.

Sequential passaging of NTHI in the chinchilla. NTHI strain 86-028NP was grown on chocolate agar overnight. NTHI were resuspended in 0.9% (w/v) sodium chloride in non- pyrogenic sterile water (Pfizer, New York NY) to an optical density of 0.65 measured at

490 nm and diluted for inoculation. One cohort of three chinchillas each were transbullarly inoculated in each ear with 300μL containing ~5000 CFUs of bacteria.

Seven days post-inoculation, middle ear effusions were collected by epitympanic tap from the inferior bullae of each right ear. Chinchillas were then sacrificed and the right middle ear mucosa tissues were removed, weighed and homogenized. Effusions and mucosal homogenates were serially diluted and plated on chocolate agar for enumeration. The colonies from the enumeration plates were then separately collected.

Genomic DNA was purified from an aliquot of cells while the remained were resuspended in 0.9% (w/v) sodium chloride and diluted for inoculation as detailed for the initial inoculum. Three hundred microliters (~5000 CFU) of each inoculum was then

219 transbullarly inoculated into the left and the right ear of a chinchilla. After a further seven days the process of harvesting effusions and mucosal homogenates and inoculating chinchillas was repeated. In total bacteria were used to infect chinchillas for four consecutive seven day periods with genomic DNA being isolated at each step (D+7,

D+15, D+23, D+31). Genomic DNA from D+31 was sequenced as detailed below

Whole genome sequencing of recovered NTHI. Genomic DNA was purified using a

Gentra Puregene Yeast/Bact. Kit (QIAGEN, Germantown, MD) and then sequencing libraries were prepared and sequenced using paired end 300bp chemistry on the MiSeq platform to high coverage on a single flow cell (Illumina, San Diego, CA). Nucleotide variants were identified using the Churchill algorithm (396) and the percentage of discordant reads as compared to the 86-028NP reference genome was calculated.

Single nucleotide variants were independently validated by Sanger sequencing. Primers used for amplification of each gene by PCR are listed in Table C.1. The amplicons were then purified using a QIAquick PCR Purification Kit (QIAGEN) and sequenced by

Eurofins Genomics (Louisville KY), using the amplification primers as sequencing primers. Sequence data was assembled using SeqMan Pro (DNASTAR, Madison, WI) and nucleotide variants identified through comparison with the parental gene sequence.

Visualization of IBCs in chinchilla middle ear mucosal tissue (Figure 5.4). Excised middle ears were fixed in 4% paraformaldehyde, decalcified in 0.35 M Tris/EDTA solution and embedded in paraffin. Samples were thin sectioned onto microscope slides, and immunohistochemistry and imaging of IBCs were performed as described in

Chapter 2, Section 2.3 (Materials and Methods).

220

Name Sequence Tm Size Note AH0607 AACGGCTCGCATTATAGGA 50.1 F primer to sequence hgpB 538bp AH0608 ATATTAATCGCTACACGGTTTTC 49.5 R primer to sequence hgpB AH0609 CGAGTGTTGTGGCGTGAT 49.5 F primer to sequence lgtC 424bp AH0610 TTTATTATTGTCTTATTTTCCTGATTTA 49.6 R primer to sequence lgtC AH0611 ATATTTGGTCATTAGCCGTAGAA 49.8 F primer to sequence NTHI0512 529bp AH0612 GCAAACGCAATAAAAATCCA 50.7 R primer to sequence NTHI0512 AH0613 GCTTACGTGTACGCAATAATGA 50.2 F primer to sequence NTHI0545 519bp AH0614 TCGTGCCTTCTTCCATTTT 49.7 R primer to sequence NTHI0545 AH0615 TCAAACTGGCTTGACAAAATAG 49.3 F primer to sequence lav 530bp AH0616 CGAATGGAACGGAAGAATC 49.4 R primer to sequence lav AH0617 CACCCTACGAGCTTAACCTTAT 49.3 F primer to sequence NTHI0580 555bp AH0618 GCCCACGAAAAAAAAGAATA 49.2 R primer to sequence NTHI0580 AH0619 TGTTGATTTATTCTTTTCGTATAGTAGA 50.1 F primer to sequence lic2A 519bp AH0620 AAGGTGGAATTTTAGTTTGTTTCT 49.8 R primer to sequence lic2A AH0621 GCTAAAGCCCACTCTACAACTACT 50.4 F primer to sequence hgpD 571bp AH0622 ATTTTCATCTACCCCTCTAATCG 50.5 R primer to sequence hgpD AH0623 TAAACACCCATACCCAACAAA 49.8 F primer to sequence hgpC 677bp AH0624 GGTTGGCGCACGAAAA 50.7 R primer to sequence hgpC AH0625 CCCGACATTATTGAAGAAACC 50.5 F primer to sequence lpsA 627bp AH0626 TCCAAACATTTTTATTTTTACTCG 49.9 R primer to sequence lpsA AH0627 TAAAACGAAGCAATACAAACCTAA 50.4 F primer to sequence lic3A2 412bp AH0628 ATTGCAACGGAAAACATCATA 50.0 R primer to sequence lic3A2 AH0629 ATCGGGATAATCATAATAGGTGTT 50.1 F primer to sequence hmw2A 530bp AH0630 AATTACTTTATCATTTGCGTTCATT 50.8 R primer to sequence hmw2A AH0631 AGGTTAGTAGGCTTGGTTGTTTAG 50.4 F primer to sequence lgtD 619bp AH0632 GCCGAGTTCGTATTGTTGG 50.5 R primer to sequence lgtD AH0633 AGTCGCATCAGAAACAAAAGAT 50.1 F primer to sequence NTHI1667 594bp AH0634 GCAAAATATCAAGTGCTACAAAGT 49.7 R primer to sequence NTHI1667 AH0635 AACGACCACCAATTTTCTTTTA 50.3 F primer to sequence NTHI1669 602bp AH0636 GCTATACGCCCCAAACTCTT 50.5 R primer to sequence NTHI1669 AH0637 ATTTATTGAATATTTATGGATGGAAC 49.5 F primer to sequence NTHI1750 472bp AH0638 TACGAAATGCCATAAAATACTCC 50.2 R primer to sequence NTHI1750 AH0639 CATTCGTCAGTGTTATTGCTTAGT 50.2 F primer to sequence hmw1A 461bp AH0640 GCGTTTGCTGAATTTGAGAC 50.0 R primer to sequence hmw1A AH0675 TGAGCGTGGCATTAGATGA 49.7 F2 primer to sequence hgpC 577bp AH0676 TCACCATATTTTTTGCCATTTA 49.7 R2 primer to sequence hgpC AH0677 TCACTAAAATCAGCAGAAGGTAAT 49.6 F3 primer to sequence hgpC 511bp AH0678 TGTCATAACTAGGCAAATA 49.5 R3 primer to sequence hgpC AH0679 CTAGATTTAAATTATCGTTATGACCAC 49.7 F4 primer to sequence hgpC 662bp AH0680 TTATAGCCTAAATGGAAACCTTG 49.7 R4 primer to sequence hgpC

Table C.1. Primers used for validation of mutations acquired by NTHI after passaging in the chinchilla.

221