Neural Mechanisms Involved in Enterotoxin- Induced Intestinal Hypersecretion

NEURAL MECHANISMS INVOLVED IN ENTEROTOXIN- INDUCED INTESTINAL HYPERSECRETION

Katerina Koussoulas BSc, MSc UMelb.

Submitted in total fulfillment of the requirements of the degree of Doctor of Philosophy

November 2017

Department of Physiology

The University of Melbourne

ABSTRACT

Exotoxins of the bacteria Vibrio cholerae (cholera toxin, CT) and Clostridium difficile (C. diff., TcdA) induce rampant disease in the form of irresolvable diarrhoea causing rapid dehydration and potential death if left untreated. Both bacterial toxins affect the nervous system of the gut, the enteric nervous system (ENS), but the types of enteric neurons involved are still indistinct. Additionally, preliminary work from a collaborator showed that specific bacterial metabolites, particularly GABA produced by the microorganisms residing in the gut, exacerbate pathophysiological effects of C. diff. GABA and its receptors are expressed in several parts of the gut wall, including enteric neurons. While studies have proposed GABA to be a putative neurotransmitter in the ENS, physiological roles of GABA in the gut remain unclear. It is unknown how enterotoxins and increased GABA at the level of the gut mucosa activate underlying enteric circuitry; my PhD aimed to elucidate these mechanisms.

In Chapter 3, I investigated the enteric neural pathways underlying CT effects via in vitro incubations of CT in guinea pig jejunum. Previous work highlighted the impacts of CT on secretomotor neurons; I endeavoured to expand this by examining other key neuronal subtypes. I recorded neuronal activity in the myenteric plexus (MP) up to 6 hours after CT incubation via intracellular electrophysiology. A colleague undertook similar recordings in the submucosal plexus (SMP). We found that CT induced hyperexcitability in myenteric, but not submucosal, sensory neurons. The effect was neurally mediated and required activation of NK3 tachykinin receptors, but was independent of activation of 5-HT3 receptors or NK1 tachykinin receptors, suggesting that the effects of CT on myenteric sensory neurons are likely to be indirect and via a pathway independent of 5-HT release.

In Chapter 4, I determined the effects of luminal incubations TcdA and GABA on myenteric sensory neurons via electrophysiology. I found that in vitro incubations of guinea pig jejuna with TcdA or GABA also increased the excitability of myenteric sensory neurons, highlighting the key role of these neurons as a common point through which enterotoxins and GABA operate. The GABA-induced effects were inhibited by

GABAB and GABAC receptor antagonists, but enhanced by a GABAA antagonist, i

indicating involvement of at least two distinct GABA activated pathways. The GABAA antagonist enhanced excitability on its own suggesting that tonic release of endogenous GABA may play a role in suppressing the excitability of these neurons.

In Chapter 5, I explored the role of endogenous GABA in the ENS of mouse small intestine. I employed Wnt1-Cre;R26R-GCaMP3 mice, which express a fluorescent calcium indicator in the ENS, for use in Ca2+ imaging. Neurons responded to GABA exposure via activation of GABAA, GABAB and GABAC receptors in myenteric ganglia. Further, I showed that the effects of GABA were neuronal subtype specific, for example neurons immunoreactive for neuronal nitric oxide synthase rarely responded to GABA. I also demonstrated that endogenous release of GABA may inhibit activation of myenteric neurons by activation of GABAC receptors, despite such receptors exciting myenteric neurons when activated by exogenous GABA. My data also suggest that neither GABAA nor GABAB receptors contribute to synaptic transmission in this system. Further I also demonstrated the expression of GABA in neurons and varicosities surrounding specific enteric neurons within the MP. This study clarifies the complex nature of GABAergic transmission in the ENS.

In Chapter 6 to further examine the effects of enterotoxins on the enteric circuitry, I made intracellular recordings from myenteric neurons following in vivo incubations of CT in mouse ileal loops. A lab member previously showed that CT increases calcium responses in the submucosal but not myenteric, neurons. In undertaking electrophysiological recordings, a striking sampling bias was revealed with exclusion of largely descending interneurons and inhibitory motor neurons being markedly underrepresented in the data set and low sampling from sensory neurons meant that significant effects in excitability may have been missed. Nevertheless in concordance with the results from Ca2+ imaging, no significant changes in excitability of myenteric neurons were found at their resting membrane potential. However, CT induced spontaneous synaptic activity in specific myenteric neurons, but the sources of this input could not be identified due to the technical difficulty of maintaining impalements, the relative rarity of myenteric sensory neurons and the sampling bias. The data suggest a minor role for myenteric neurons in CT-induced hypersecretion in vivo.

In Chapter 7 I employed the high throughput assay of Ca2+ imaging to perform a more extensive examination of the effects of TcdA on the ENS. I utilized the well-established ileal loop mouse model and incubated TcdA in vivo. Spontaneous and neurally- stimulated calcium responses were reduced in submucosal neurons, and myenteric neuronal activity was unchanged. However enteric neurons in regions of the gastrointestinal tract off-target from the site of acute toxin exposure were activated during the incubation as indicated by expression of activity dependent markers. This off target effect could possibly be due to release of inflammatory cytokines into the circulation or extrinsic neural pathways.

In all, I have demonstrated a generality in the actions of enterotoxins and GABA as the pathways they activate converge to excite myenteric sensory neurons which may lead to activation of submucosal secretomotor neurons. I have extended our understanding of the role of GABA in the ENS as a means to elucidate the mechanisms through which microbial metabolites act and contribute to disease. Using the mouse ileal loop model, I further defined the effects of enterotoxins on the enteric circuitry.

In this way, my thesis highlights neural elements involved in the mechanisms underlying enterotoxin-induced hypersecretion and identifies potential avenues for future research.

iii

DECLARATION

This is to certify that:

• this thesis comprises only my original work except where indicated in the preface;

• due acknowledgement has been made in the text to all other material used; and

• the thesis is less than 100,000 words in length, inclusive of footnotes, but exclusive of tables, maps, appendices and bibliography.

Katerina Koussoulas

November 2017

PREFACE

Under my direction, the following people have made the stated contributions to this work:

Chapter 3: I undertook 50% of the intracellular recording experiments, about half of these were conducted during my Masters (MSc). Ms Rachel Gwynne performed 50% of the intracellular recording experiments including some myenteric and all submucosal neurons. I performed 60% of the data analysis. Assistance in preparing the manuscript for this chapter for publication was provided by Prof Joel Bornstein, Dr Jaime Foong and Ms Rachel Gwynne.

Chapter 4: I undertook 90% of the intracellular recordings and all data analysis. Dr Jaime Foong performed 2 intracellular recording experiments, Ms Rachel Gwynne performed 3 intracellular recording experiments. Technical assistance and advice on use of TcdA was provided by A/Prof. Tor Savidge.

Chapter 5: I undertook 75% of the calcium imaging experiments, immunohistochemistry and data analysis. Dr Jaime Foong performed some calcium imaging experiments with drug antagonists and some post-hoc immunohistochemical experiments. Calcium imaging preparations were rung by Ms Candice Fung. Technical assistance and advice on calcium imaging experiments and immunohistochemistry was provided by Dr Jaime Foong. Dr Jaime Foong and Ms Mathusi Swaminathan imaged and analysed GABAergic varicosities.

Chapter 6: I performed 80% of the ileal loop surgeries and undertook 100% of the intracellular recordings and data analysis. Technical assistance with mouse surgeries was provided by Prof Andrew Allen and Ms Candice Fung. Ms Petra Unterweger performed some surgeries.

Chapter 7: I performed all ileal loop surgeries and undertook 100% of the experiments and data analysis for calcium imaging and immunohistochemistry. Ms Candice Fung rung preparations for calcium imaging and performed all Ussing chamber experiments and associated data analysis. Technical assistance and advice on cell culture for TcdA cytotoxicity assays was provided by Dr Marissa Caldow.

Publications arising from this thesis include:

Koussoulas, K., Gwynne, R.M., Foong, J.P.P., and Bornstein, J.C. (2017). Cholera Toxin Induces Sustained Hyperexcitability in Myenteric, but Not Submucosal, AH Neurons in Guinea Pig Jejunum. Frontiers in Physiology. 254 (8).

Chambers, J.D., Bornstein, J.C., Gwynne, R.M., Koussoulas, K., and Thomas, E.A. (2014). A detailed, conductance-based computer model of intrinsic sensory neurons of the gastrointestinal tract. Am J Physiol 307(5) G517-G532.

Abstracts (* refereed):

Koussoulas, K., Gwynne, R.M., Foong, J.P.P., Ross, C., Savidge, T.C., and Bornstein, J.C. Clostridium difficile toxin and microbial-derived GABA signals converge to hyperexcite myenteric intrinsic sensory neurons (2015). Gastroenterology 148, S-21*

Dann, S.M., Aitken, S., Ross, C., Tessier, M.E.M., Loeffelholz, M., Koussoulas, K., Joel C. Bornstein, J.C., Versalovic, J., Pothoulakis, C., Britton, R., Garey, K.W., and Savidge, T. (2015). Zolpidem confers disease susceptibility to Clostridium difficile infection. Gastroenterology 148, S727–S728. *

Koussoulas, K., Gwynne, R.M., Foong, J.P.P., Savidge, T.C., and Bornstein, J.C. (2014). Clostridium difficile toxin A and GABA increase the excitability of myenteric intrinsic sensory neurons in the guinea-pig jejunum. 33rd Annual Meeting of the Australian Neuroscience Society, Melbourne, AU.

Koussoulas, K., Gwynne, R.M., Unterwerger, P., Foong, J.P.P., and Bornstein, J.C. (2014). Mucosal GABA excites local inhibitory reflexes and increases the excitability of myenteric intrinsic sensory neurons in the guinea pig jejunum via a GABAA receptor independent mechanism. Neurogastroenterology & Motility Society, Melbourne, AU.

Koussoulas, K., Gwynne, R.M., and Bornstein, J.C. (2012). Cholera toxin increases excitability of myenteric, but not submucosal, intrinsic sensory neurons in guinea-pig jejunum, Proceedings of the 32nd Annual Meeting of the Australian Neuroscience Society, Gold Coast, AU.

Of published material in this thesis, the named authors made the following contributions:

Koussoulas, K., Gwynne, R.M., Foong, J.P., and Bornstein, J.C. (2017). Cholera Toxin Induces Sustained Hyperexcitability in Myenteric, but Not Submucosal, AH Neurons in Guinea Pig Jejunum. Frontiers in physiology, 8. (Chapter 3)

K. Koussoulas undertook recordings from myenteric neurons, analysed data and wrote the manuscript draft, R. M. Gwynne undertook recordings from myenteric and submucosal neurons, analysed data and helped revise manuscript, J.P.P. Foong assisted with experimental design and figure construction, also revised manuscript, J.C. Bornstein designed experiments, revised manuscript.

J.D. Chambers, J.C. Bornstein and E.A. Thomas conception and design of research; J.D. Chambers, R.M. Gwynne and K. Koussoulas performed experiments; J.D. Chambers, R.M. Gwynne and K. Koussoulas analysed data; J.D. Chambers, J.C. Bornstein and E.A. Thomas interpreted results of experiments; J.D. Chambers prepared figures; J.D. Chambers drafted manuscript; .D. Chambers, J.C. Bornstein and E.A. Thomas edited and revised manuscript; J.D. Chambers, J.C. Bornstein, R.M. Gwynne, K. Koussoulas and E.A. Thomas approved final version of manuscript.

vii

ACKNOWLEDGEMENTS

Undertaking my Ph.D. has been at many times a challenging, but ultimately rewarding experience. I have several people to thank for their knowledge, support and assistance, without which, this thesis would not have come to completion.

My primary supervisor Prof Joel Bornstein. Joel, I am grateful that you opened your office door to me all those years ago. Undoubtedly this thesis wouldn’t have been possible without your knowledge and guidance. I have particularly enjoyed our lengthy conversations often about everything but science. My supervisor Dr Jaime Foong. Jaime thank you for everything, from project design, career advice and all the ‘3.30 pm’ meetings and phone calls. I am so very appreciative of your hard work and the sacrifices you have made on my behalf. I am thankful to you both for shaping my career as a scientist and for your friendship.

I would like to thank A/Prof Tor Savidge who walked into our lab with a vial of TcdA and ultimately shaped the course of my Ph.D. Thank you Tor for all your assistance with Chapter 7 and dealing with ‘lost’ Fed-Ex shipments of toxin.

Rachel Gwynne has assisted and supported me from my very first day in a laboratory and taught me everything I needed to know about electrophysiology. Without her, Chapter 3 would not be possible. Thanks Rach.

I owe thanks to Prof Andrew Allen for his technical assistance in setting up the animal surgeries used in Chapters 6 and 7. I am grateful to Dr Marissa Caldow for her expertise in cell culture, without which I would have been fumbling about. I would like to acknowledge Dr Jordan Chambers for his assistance in helping me understand computer modelling and Dr Parvin Zarei Eskikand for her computer modelling work in Chapter 5. I thank Angela Huf for her help with Cryosectioning and H&E. Thank you to Pavitha Parathan for genotyping my Wnt1-Cre;R26R-GCaMP3 mice.

I would like to acknowledge the members of the Department of Physiology including the Physiology Student’s Society for our friendly and lively work environment and for valuable scientific feedback.

viii

I am grateful to members of the Bornstein Laboratory, past and present, for their invaluable assistance, company and laughter, all of which extended well beyond 5 pm. I owe special thanks to:

Lin Hung, always willing to complain about immunohistochemistry with me and for our vending machine Twix trips.

Candice Fung, my Masters buddy. We have been through our fair share of trials together especially during the first year of our Ph.D. I am appreciative of all of your help with the calcium imaging preparations.

Mathusi Swaminathan, the immuno whiz. Thank you for your help with the GABA varicosity imaging and analysis. We have experienced just about everything science can throw at a person and I am glad we did it together. Your friendship gave me the sanity to complete this Ph.D.

Petra Unterweger, what can I say? From your first day as an RA it was clear to me that the next three years would be some of my fondest in the lab. Thank you for your assistance with ileal loop surgeries and labelling just about every piece of glassware in the lab with my initials!

Lastly, I am indebted to my family; Mum, Dad, Jeannie, George (and Ajax) for their patience, encouragement, humour, and always putting life into perspective for me in the best possible way. Thank you to my partner Jimmy who endured many weekends and holidays alone when I was experimenting and writing. I know your guidance and support will remain long after the Ph.D. I love you all.

Katerina Koussoulas

November 2017

TABLE OF CONTENTS

ABSTRACT ...... i

DECLARATION ...... iv

PREFACE ...... v

ACKNOWLEDGEMENTS ...... viii

TABLE OF CONTENTS ...... x

LIST OF TABLES ...... xix

LIST OF FIGURES ...... xx

LIST OF ABBREVIATIONS ...... xxiii

CHAPTER 1: INTRODUCTION ...... 1

1. THE ENTERIC NERVOUS SYSTEM ...... 2

1.1 Involvement of the ENS in hypersecretion ...... 2

1.2 ENS anatomy and function...... 3

2. SECRETION ...... 5

2.1 Secretion at the mucosa ...... 5

2.2 Secretion under the control of the ENS ...... 5

3. CONSTITUENT NEURONS OF THE ENS ...... 7

3.1 Morphology ...... 8

3.2 Electrophysiology ...... 10

3.3 Neurochemical coding of neurons ...... 16

4. FUNCTIONAL CLASSIFICATION OF NEURONS ...... 16

4.1 Intrinsic sensory neurons ...... 17

4.2 interneurons ...... 20

4.3 Motor and secretomotor neurons ...... 21

5. SYNAPTIC TRANSMISSION IN AH AND S NEURONS ...... 24

5.1 Synaptic transmission in AH neurons ...... 24

5.2 Synaptic transmission in S neurons ...... 25

6. REFLEX CIRCUITRY OF SECRETION ...... 27

7. MECHANISMS FOR PATHOGENIC CT AND TCDA-INDUCED HYPERSECRETION ...... 28

7.1 Direct effect of CT on enterocytes (non-neural mechanism)...... 28

7.2 Direct effect of TcdA (non-neural mechanism) ...... 29

7.3 Neural mechanism for CT-induced hypersecretion ...... 31

7.4 Sustained CT-induced hyperexcitability of reflex ‘output’ ...... 32

7.5 Neural and inflammatory mechanism for TcdA-induced hypersecretion . 35

7.6 Interaction of C.diff and the ‘microbiome’ ...... 40

8. THE PHYSIOLOGICAL ROLE OF GABA IN THE ENS ...... 44

8.1 GABA receptors and their function ...... 45

8.2 GABA in secretion and as an endocrine mediator ...... 48

9. AIMS OF THESIS ...... 50

CHAPTER 2: MATERIALS AND METHODS ...... 51

1. ELECTROPHYSIOLOGY ...... 51

1.1 Tissue preparation for intracellular recording ...... 51

1.2 Impaling and recording from neurons ...... 54

1.3 Morphology of biocytin-filled neurons ...... 56

1.4 Analysis and statistics ...... 57

2. ILEAL LOOP SURGERY ...... 57

2.1 Surgical preparation ...... 57

2.2 Surgical procedure ...... 58

3. CRYOSECTIONING ...... 59

4. CELL CULTURE ASSAY ...... 61 xi

5. IMMUNOHISTOCHEMISTRY ...... 62

5.1 Tissue preparation...... 62

5.2 Imaging and analysis ...... 62

6. CALCIUM IMAGING ...... 63

6.1 Tissue preparation for Ca2+ imaging ...... 63

6.2 Experimental protocol ...... 63

6.3 Analysis ...... 64

7. MEASUREMENT OF SHORT-CIRCUIT CURRENT IN VITRO USING USSING CHAMBERS ...... 64

7.1 Tissue preparation...... 64

7.2 Electrical measurements ...... 65

7.3 Data analysis ...... 65

8. DRUGS AND TOXINS USED ...... 66

CHAPTER 3: CHOLERA TOXIN INDUCES SUSTAINED HYPEREXCITABILITY IN MYENTERIC, BUT NOT SUBMUCOSAL, AH NEURONS IN GUINEA PIG JEJUNUM ...... 67

ABSTRACT ...... 67

1. INTRODUCTION ...... 67

2. METHODS ...... 69

2.1 Tissue preparation and electrophysiology ...... 69

2.2 Immunohistochemistry for biocytin-filled neurons ...... 70

2.3 Drugs and toxins ...... 70

2.4 Analysis and statistics ...... 71

3. RESULTS ...... 73

3.1 CT increases excitability of myenteric AH neurons and the effect is neurally mediated ...... 73

xii

3.2 Effect of CT on the membrane properties and synaptic potentials of myenteric AH neurons ...... 75

3.3 Firing of submucosal AH neurons is unaltered by CT ...... 77

3.4 Excitability of myenteric AH neurons is affected by proximity to intact mucosa ...... 79

3.5 CT’s effect on myenteric AH neurons is mediated by NK3, but not NK1 or

5-HT3 receptors ...... 81

3.6 Myenteric S neurons show no change in firing following CT incubation .. 83

4. DISCUSSION ...... 86

4.1 CT increases the excitability of myenteric AH neurons, but not submucosal AH neurons or myenteric S neurons ...... 86

4.2 CT increases the excitability of myenteric AH neurons via a mechanism requiring neural activity ...... 87

4.3 CT probably acts to reduce IK channel activity in AH myenteric neurons ...... 88

4.4 Role of mucosal mediators in CT-induced effects on enteric neurons ...... 89

4.5 Role of tachykinin receptors...... 90

4.6 New proposed circuit ...... 91

4.7 Conclusions and future directions ...... 94

CHAPTER 4: CLOSTRIDIUM DIFFICILE TOXIN A AND GABA SIGNALS CONVERGE TO MAKE MYENTERIC INTRINSIC SENSORY NEURONS HYPEREXCITABLE ...... 95

ABSTRACT ...... 95

1. INTRODUCTION ...... 95

2. METHODS ...... 96

2.1 Tissue preparation and electrophysiology ...... 96

2.2 Drugs and Toxins ...... 97

3. RESULTS ...... 98

xiii

3.1 TcdA increases excitability of myenteric AH neurons ...... 98

3.2 GABA increases excitability of myenteric AH neurons ...... 100

3.3 Co-incubating GABA with bicuculline enhances the excitability, while bicuculline alone produces hyperexcitability in myenteric AH neurons ...... 100

3.4 The GABAB antagonist CGP54626 and GABAC antagonist TPMPA inhibit the GABA induced excitability ...... 102

4. DISCUSSION ...... 107

4.1 TcdA makes myenteric AH neurons hyperexcitable ...... 107

4.2 Pathways activated by TcdA ...... 108

4.3 GABA makes myenteric AH neurons hyperexcitable ...... 109

4.4 GABA acts via GABA receptor subtypes in separate pathways ...... 110

4.5 AH neurons, enterotoxins and GABA in the context of the circuit ...... 112

4.6 Conclusions and future directions ...... 113

CHAPTER 5: GABAERGIC TRANSMISSION IN THE MP OF THE MURINE ILEUM ...... 115

ABSTRACT ...... 115

1. INTRODUCTION ...... 115

2. METHODS ...... 118

2.1 Mice ...... 118

2.2 Immunohistochemical analysis of the expression of GABA in neurons and its co-expression with other neuronal subtype markers ...... 118

2.3 Imaging, analysis and statistics ...... 118

2.4 Tissue preparation for Ca2+ imaging ...... 119

2.5 Post-hoc immunohistochemistry for Ca2+ imaging experiments ...... 120

2.6 Drugs ...... 120

2.7 Analysis and statistics ...... 120

3. RESULTS ...... 122

xiv

3.1 Expression of GABA in the distal ileal myenteric plexus ...... 122

3.2 Some myenteric neurons including GABA-IR neurons respond to 2+ exogenous GABA with [Ca ]i transients ...... 125

2+ 3.3 Half the GABA-induced [Ca ]i transients are due to synaptic activation ...... 127

3.4 GABA-evoked effects are mediated through GABA A, B and C receptors ...... 127

3.5 Involvement of endogenous GABA in synaptic transmission ...... 132

4. DISCUSSION ...... 136

4.1 GABA colocalises with some calretinin and nNOS neurons in the MP of the mouse ileum ...... 136

4.2 GABA activates myenteric neurons in the mouse ileum ...... 137

4.3 GABAA GABAB and GABAC receptors are expressed in the myenteric circuitry ...... 138

4.4 The GABAC receptor has a synaptic function and an inhibitory role in neural circuits ...... 141

4.5 Future directions and conclusions ...... 146

CHAPTER 6: A MORPHOLOGICAL AND ELECTROPHYSIOLOGICAL STUDY OF MYENTERIC NEURONS IN THE MOUSE ILEUM – EFFECTS OF CT ON THEIR EXCITABILITY ...... 147

ABSTRACT ...... 147

1. INTRODUCTION ...... 147

2. METHODS ...... 149

2.1 Mice ...... 149

2.2 Ileal loop surgery ...... 149

2.3 Tissue preparation for electrophysiology ...... 149

2.4 Intracellular recording ...... 149

2.5 Analysis and statistics ...... 150

2.6 Drugs ...... 150

3. RESULTS ...... 150

3.1 CT in the ileum in vivo produces enhanced secretion...... 150

3.2 Electrophysiology and morphology of myenteric neurons in the mouse ileum ...... 151

3.3 CT induces an increase in spontaneous activity in S neurons ...... 153

4. DISCUSSION ...... 160

4.1 Considerations of the general electrophysiology and morphology of myenteric neurons in the mouse ileum ...... 160

4.2 CT enhances secretion and increases spontaneous activity in myenteric S neurons ...... 163

4.3 Effects of CT on AH neurons ...... 165

4.4 In context of the circuit ...... 166

4.5 Conclusions and future directions ...... 167

CHAPTER 7: EFFECTS OF IN VIVO LUMINAL INCUBATION OF TCDA ON THE MOUSE ILEUM ...... 169

ABSTRACT ...... 169

1. INTRODUCTION ...... 169

2. METHODS ...... 171

2.1 Mice ...... 171

2.2 Ileal loop surgery ...... 171

2.3 Validation of the viability of TcdA ...... 171

2.4 Histological processing and imaging of ileal loop and off-target tissue ... 173

2.5 Measurement of short-circuit current (ISC) in vitro using Ussing chambers ...... 173

2.6 Ca2+ imaging and analysis ...... 173

2.7 Immunohistochemistry for activity dependent markers ...... 174

xvi

2.8 Imaging and analysis of activity-dependent markers ...... 174

3. RESULTS ...... 176

3.1 Incubations of TcdA ...... 176

3.2 TcdA in the ileum in vivo produces gross mucosal damage and a sustained increase in basal short circuit current ...... 176

3.3 TcdA does not alter pCREB expression in submucosal neurons, but submucosal neurons display a sustained reduction in spontaneous Ca2+ transients ...... 178

3.4 TcdA does not induce changes in c-Fos expression or sustained changes in excitability in the MP ...... 181

3.5 TcdA induced epithelial damage and increased neuronal pCREB and c-Fos expression in off-target regions ...... 184

4. DISCUSSION ...... 188

4.1 TcdA produces secretion, mucosal damage and inflammation...... 188

4.2 TcdA evokes a sustained reduction in the excitability of the SMP and no change in excitability in the MP within the ileal loop ...... 189

4.3 Potential role for extrinsic neural activity within the ileal loop ...... 191

4.4 Some enteric neurons in off-target regions may be activated during TcdA incubation ...... 192

4.5 Conclusions and future directions ...... 193

CHAPTER 8: CONCLUSIONS ...... 195

1. CT INDUCES SUSTAINED HYPEREXCITABILITY IN MYENTERIC AH NEURONS ...... 196

2. MYENTERIC AH NEURONS ARE A COMMON PATHWAY THROUGH WHICH CT, TCDA AND GABA ACT ...... 197

3. ENDOGENOUS GABA IS INVOLVED IN SYNAPTIC TRANSMISSION 198

4. THE INVOLVEMENT OF THE ENS IN CT-INDUCED HYPERSECRETION IN VIVO ...... 199

xvii

5. THE INVOLVEMENT OF THE ENS IN TCDA-INDUCED HYPERSECRETION IN VIVO ...... 201

6. A FINAL WORD ...... 202

LIST OF REFERENCES ...... 204

xviii

LIST OF TABLES

Table 1. 1 Functional classes of neurons in the myenteric plexus of the guinea pig and mouse small intestine...... 23 Table 2. 1 Protocol for H&E staining...... 60 Table 3. 1 Membrane properties and synaptic potentials of myenteric AH neurons close to mucosa following CT treatment...... 76 Table 3. 2 Membrane properties and synaptic potentials of myenteric S neurons in control and CT-treated preparations...... 85 Table 4. 1 Membrane properties of myenteric AH neurons following TcdA and GABA pre-treatment...... 104 Table 4. 2 Membrane properties and synaptic potentials of myenteric AH neurons following TcdA and GABA pre-treatment...... 105 Table 4. 3 Membrane properties and synaptic potentials of myenteric AH neurons following various GABA receptor antagonist pre-treatments...... 106 Table 5. 1 Primary antisera used for immunostaining...... 121 Table 5. 2 Secondary antisera used for immunostaining...... 121 Table 5. 3 Number of neurons responding to GABA in control conditions and in the presence of drug antagonists...... 129 Table 5. 4 Number of neurons responding to electrical stimulation in control conditions and in the presence of antagonists...... 133 Table 6.1 Electrophysiological parameters of neurons recorded from control and CT-treated ileal loops...... 159 Table 7.1: Primary and secondary antisera used for immunostaining...... 175

xix

LIST OF FIGURES

Figure 1. 1 The layers of the guinea pig intestinal wall constituting the ENS...... 4 Figure 1. 2 Confocal micrograph showing the morphology of enteric neurons...... 9 Figure 1. 3 The action potential and after-potentials in AH neurons...... 15 Figure 1. 4 The models for the secretory reflex pathway in the ENS...... 34 Figure 1. 6 The microbiome and disease susceptibility in CDI...... 42 Figure 1. 7 Zolpidem promotes infection and disease severity in CDI...... 43 Figure 1. 8 GABA receptor activation and signalling underlying motor output in the rodent small intestine...... 48 Figure 2. 1 Tissue preparation for intracellular recording...... 53

Figure 3. 1 Measurements of Ih-induced rectification and ADP amplitude...... 72 Figure 3. 2 CT increases the excitability of myenteric AH neurons close to the mucosa...... 74 Figure 3. 3 CT did not affect the excitability of submucosal AH neurons close to the mucosa...... 78 Figure 3. 4 Presence of the mucosa reduces action potential firing in myenteric AH neurons...... 80

Figure 3. 5 Co-incubation of CT with a NK3 but not NK1 or 5-HT3 receptor antagonist prevents CT-induced hyperexcitability in myenteric AH neurons...... 82 Figure 3. 6 Effects of CT on the action potential firing properties of myenteric S neurons close to the mucosa...... 84 Figure 3. 7 Schematic of secretomotor pathways of the guinea pig ENS activated by CT...... 93 Figure 4. 1 TcdA makes myenteric AH neurons hyperexcitable...... 99

Figure 4. 2 The effects of GABA and the GABAA antagonist Bicuculline, on the action potential firing properties myenteric AH neurons...... 101

Figure 4. 3 Co-incubation with either a GABAB or a GABAC antagonist, prevents the GABA-induced hyperexcitability of myenteric AH neurons...... 103 Figure 5. 1 Immunoreactivities of Hu, GABA, nNOS and calretinin in the distal ileum of the mouse...... 123

Figure 5. 2 GABA varicosities form close contacts with neurons immunoreactive for calretinin and nNOS...... 124 Figure 5. 3 Identity of neurons in the mouse ileum responding to pressure injection of GABA (1mM)...... 126 2+ Figure 5. 4 GABA (1mM) - evoked [Ca ]i transient responses in the presence of antagonists...... 131 ...... 134 2+ Figure 5. 5 Electrically - evoked [Ca ]i transient responses in the presence of antagonists...... 135 Figure 5. 6 Schematic of a postulated circuit showing dis-inhibition of unknown inputs...... 140 Figure 5. 7 Preliminary computer modelling to interpret the possible inhibitory role of GABA in modulating synaptic transmission...... 145 Figure 6.1 Control versus CT -treated ileal loops following 3.5 h incubation in vivo...... 151 Figure 6.2 Correlations between electrophysiological and morphological features in myenteric neurons of the mouse ileum...... 155 Figure 6.3 Examples of cell bodies and projections of neurons in the mouse ileum...... 156 Figure 6.4 Myenteric neurons of CT-incubated ileal segments displayed increased spontaneous synaptic input...... 158 Figure 7.1 TcdA (1µg/ml) cytotoxicity assay...... 172 Figure 7.2 Control vs. TcdA-treated ileal loops following 2.5-3.5 h incubation in vivo...... 177 Figure ...... 179 7.3 Submucosal neurons do not show a change in pCREB expression following TcdA exposure...... 179 Figure 7.4 Submucosal neurons show an overall sustained reduction in excitability ...... 180 Figure 7.5 Myenteric neurons do not show a change in c-Fos expression following TcdA exposure...... 182 Figure 7.6 MP shows no sustained changes in excitability following TcdA exposure in vivo...... 183 xxi

Figure 7.7 TcdA-induced epithelial damage in the jejunum...... 185 Figure 7.8 Expression of activity-dependent markers in off-target regions of the GIT...... 187

xxii

LIST OF ABBREVIATIONS

5-HT: 5-hydroxytryptamine, serotonin ACh: Acetylcholine ADP: After-depolarising potential AH: Neurons that display slow after-hyperpolarising potentials AHP: After-hyperpolarising potential AP: Action potential ATP: Adenosine triphosphate BK: Large conductance potassium channel Ca2+: Calcium [Ca2+]i: Intracellular calcium cAMP: Cyclic adenosine monophosphate C.diff: Clostridium difficile CFTR: Cystic fibrosis transmembrane conductance regulator CGRP: Calcitonin gene-related peptide Cl-: Chloride CNS: Central nervous system CT: Cholera toxin DRG: Dorsal root ganglion EC: Enterochromaffin ENS: Enteric nervous system EPSP: Excitatory post synaptic potential GABA: γ-amino butyric acid gKCa: Ca²⁺- dependent K⁺ conductance GIT: Gastrointestinal tract GM1: Plasma membrane ganglioside HCN: Hyperpolarisation-activated non-specific cation conductance

xxiii

Ih: Hyperpolarisation-activated cation current

ISC: Short circuit current IK: Intermediate conductance potassium channel IPAN: Intrinsic primary afferent neuron IPSP: inhibitory postsynaptic potential IR: Input resistance ISN: Intrinsic sensory neuron KCl: Potassium chloride MMC: Migrating motor complex MP: Myenteric plexus nAChR: Nicotinic acetylcholine receptor NK: Neurokinin NPY: Neuropeptide Y NO: Nitric oxide NOS: Nitric oxide synthase PBS: Phosphate buffered saline PCAP: Pituitary adenylate cyclase activating peptide PKA: Protein kinase A PKC: Protein kinase C RMP: Resting membrane potential SEM: Standard error of the mean SMP: Submucosal plexus SP: Substance P SOM: Somatostatin SSPE: Sustained slow post synaptic potential TcdA: Clostridium difficile toxin A TcdB: Clostridium difficile toxin B TEA: Tetraethylammonium TK: Tachykinin xxiv

TTX: Tetrodotoxin VIP: Vasoactive intestinal peptide

xxv

CHAPTER 1: INTRODUCTION

Fluid secretion in the intestine in response to benign stimuli such as nutrient molecules occurs every day under physiological conditions. It is imperative for regulation of whole body water and electrolyte balance, digestion, and helps to maintain the osmolarity and pH of the microenvironment at the absorbing epithelium (Furness & Costa 1987). Under pathological circumstances; the focus of this thesis, secretion functions to flush the intestine of unwanted or noxious stimuli. Despite being a protective reflex, the secretion itself can become copious and deadly (Furness, 2006).

Worldwide, there are almost two billion cases of diarrhoeal disease every year (WHO, 2017). While there are many causes of diarrhoea, often the most severe cases are those due to bacterial infections. This occurs either due to a lack of clean water supply and basic sanitation, conditions that are prevalent in third-world nations and regions devastated by natural disasters, or by conditions commonly found in nosocomial environments. It is a leading cause of malnutrition and mortality in children in underdeveloped countries, responsible for the death of 0.5 million children annually (WHO, 2017). In the developed world, diarrhoeal disease places a huge burden on the healthcare system in many nations, costing over $1 billion every year in the United States alone (Kyne et al., 2002).

Many diarrhoea-causing bacteria release exotoxins that result in a harmful hypersecretion of water and electrolytes in the gut (Farthing, 2000; Lundgren, 2002), with the most widely studied of these toxins being cholera toxin (CT). Produced by the bacterium vibrio cholera, CT evokes an often fatal hypersecretory response. Similarly, Clostridium difficile (C.diff), the leading cause of antibiotic-associated diarrhoea and colitis, releases enterotoxins (Toxin A -TcdA and Toxin B - TcdB) that cause severe mucosal damage and inflammation in addition to harmful hypersecretion (Castagliuolo et al., 1994; Pothoulakis et al., 1998).

The primary line of treatment for diarrhoeal disease is an oral rehydration solution (ORS) of clean water and electrolytes, but given that the very nature of transmission of these bacterially- borne infections relies on a lack of clean water; this treatment is often

ineffective and not possible. Although oral cholera vaccines are available, they require multiple doses for full protection and only provide cover for restricted time periods. Moreover, in hospital settings, recurrence of C.diff infection (CDI) remains the key impediment to its treatment, and with a global rise in antibiotic-resistance, alternative therapies are needed (WHO, 2016, 2017).

These enterotoxins each act directly on the mucosal epithelium of the gastrointestinal tract and also indirectly by activating and perturbing the function of the intrinsic nervous system of the gut, the enteric nervous system (ENS) (Castagliuolo et al., 1994; Lundgren, 1998; Pothoulakis et al., 1998; Xia et al., 2000; Farthing, 2000, 2002). The responses triggered by the different enterotoxins, however, appear to involve different neural pathways, the exact elements of which remain surprisingly undefined (Pothoulakis and Lamont 2001).

This thesis aims to contribute to our understanding of the enteric neural pathways underlying enterotoxin-induced diarrhoeal disease. A deeper understanding of the complex enteric neural circuitry may assist efforts in the discovery of therapeutic targets for the treatment of diarrhoeal disease. Even though both enterotoxins induce conditions affecting humans, rodent models are used to investigate disease pathophysiology. In this chapter, details of the guinea pig model are predominately discussed as it is the most widely studied model of the ENS over the last three decades. Where available, data from the mouse is also presented as it has been the model of choice more recently for studies of CDI.

1. THE ENTERIC NERVOUS SYSTEM

1.1 Involvement of the ENS in hypersecretion

The fundamental functions of the of the gastrointestinal tract (GIT) such as motility and secretion are largely under the control of the intrinsic nervous system of the gut- the ENS, which is the only grouping of independently functioning peripheral neurons. For this reason it has been likened to a second ‘brain’ (Gershon, 1999). It does however work in concert with the central nervous system (CNS) and extrinsic neural pathways (Furness et al., 2014). It has been established that enterotoxins like CT and TcdA and 2

those produced by other diarrhoeal agents such as Escherichia coli (E coli) and viruses such as rotavirus, exert at least some of their hypersecretory effects via the ENS (Castagliuolo et al., 1994; Farthing, 2000; Lundgren, 2002).

1.2 ENS anatomy and function

The enteric nervous system is located entirely within the wall of the GIT with its neurons contained in many thousands of ganglia within two main ganglionated plexuses; the myenteric and submucosal plexus. The myenteric plexus (MP) lies between the outer longitudinal and inner circular muscle layers and forms a continuous network running from the upper oesophagus to the rectum. The submucosal plexus (SMP) is located between the mucosa and circular muscle and is primarily prominent in the small and large intestine (Furness and Costa, 1987). Figure 1.1 shows the partially separated layers of a segment of guinea pig small intestine. Many studies examining the different components of the ENS have been performed using guinea pig due to the ease with which the layers can be dissected away for experimentation (Brookes, 2001).

The GIT is exposed to a variety of physicochemical stimuli from the external environment and relies on the activity of coordinated smooth muscle contractions and relaxations, secretion, blood flow and immune processes that are under the control of the enteric plexuses (Costa et al., 2000; Hansen, 2003). These functions operate together to digest food, absorb nutrients and excrete wastes; essential processes in maintaining homeostasis and sustaining life. The primary functions of motility and secretion are regulated by intrinsic nerve circuits within the plexuses which form reflex pathways. Enteric reflex pathways in their simplest forms are comprised of three neuron types: intrinsic sensory neurons (ISNs) also known as intrinsic primary afferent neurons, IPANs (in this thesis they are referred to as ISNs), interneurons and motor neurons. In the guinea pig, motility-controlling circuits are largely contained within myenteric plexus and secretomotor reflexes are mediated by the submucosal plexus (Furness, 2006). Often these plexuses are treated as if they are independent, yet they are extensively interconnected, allowing each to play a role in both motility and secretion (Gwynne and Bornstein, 2007b). It is likely that secretion and motility are coupled, since the secretomotor and motility reflexes share common sensory neurons and therefore there is a level of simultaneous activity between changes in motility and 3

secretion (Greenwood and Davison, 1987; Mellander et al., 2000). It has been established across a number of species, including humans, that phases of the ‘housekeeping’ migrating motor complex (MMC), also involve secretory changes in the GIT, such as epithelial chloride secretion, pancreatic secretion and bile release (Vantrappen et al., 1979; Keane et al., 1980; Read, 1980; Mellander et al., 2000).

Figure 1. 1 The layers of the guinea pig intestinal wall constituting the ENS. The mucosal villi are shown projecting into the lumen of the small intestine. The SMP is situated between a layer of circular muscle and the mucosa. The MP exists between a layer of outer longitudinal muscle and the inner circular muscle layer; it is continuous around the circumference of the GIT and along its entire length. The SMP and the MP are the two main ganglionated plexuses of the ENS mediating the chief functions of the GIT. Figure adapted from (Furness and Costa, 1980).

2. SECRETION

2.1 Secretion at the mucosa

The epithelium of the gastrointestinal tract functions as the interface between the internal and external environment and is made up of various cell types including absorptive cells (villi and surface enterocytes) and crypt cells. In addition, enteroendocrine (EE) cells which can respond to luminal contents by releasing peptides and hormones (Cho et al., 2014; Jin and Blikslager, 2015) are located throughout the epithelium (Cooke, 1998). The villi are responsible for absorption, while crypt cells are effector cells responsible for electrogenic Cl- secretion; a driving force for accompanying fluid transport into the intestinal lumen. Crypt cells accumulate Cl- into their interior by an energy-requiring Na+/2Cl-/K+ cotransporter on the ‘blood-side’, the basolateral side of the cell (Cooke, 1998; Xue et al., 2007). This establishes an electrochemical gradient, by which Cl- can exit the cell into the lumen via diffusion through cAMP-activated, cystic fibrosis transmembrane regulator (CFTR), Cl- channels (Xue et al., 2007). This may also occur through other Cl- channels such as Ca2+ - activated CLCA Cl- channels or cAMP-mediated CIC-2 channels, however the exact contribution to intestinal secretion of the latter remains unclear (Xue et al., 2007; Inagaki et al., 2010; Jin and Blikslager, 2015).

2.2 Secretion under the control of the ENS

Absorption of fluid and electrolytes occurs via local transport processes at the intestinal villi, which are in close contact with the luminal content. Secretion occurs at the crypts, where CFTR expression is greatest (Ameen et al., 1995). Crypts are located further away from the lumen, thus rely on the actions of the ENS. Early studies have supported this, having established that secretion is under neural control (Caren et al., 1974; Hubel, 1985; Lundgren, 2002).

Chloride secretion is predominantly regulated by neurons in the submucosal plexus. Luminal stimulation acts primarily to evoke serotonin (5-HT) release from enterochromaffin (EC) cells; a type of enteroendocrine cell in the mucosa, (Grider et al., 1996; Sutherland et al., 2007). 5-HT, a key mediator in regulating secretion, excites

intrinsic sensory neurons in the SMP and MP (Kirchgessner et al., 1992; Bertrand et al., 2000). Within the SMP, intrinsic sensory neurons expressing tachykinins relay a signal that converges on the main classes of secretomotor neurons in the SMP: cholinergic and vasoactive intestinal peptide (VIP) neurons which release acetylcholine (ACh) and VIP as their neurotransmitters respectively (Xue et al., 2007). At the basolateral membrane, ACh binds to muscarinic (M3) receptors, resulting in Ca2+ mobilization, whereas VIP acting at VPAC1 receptors leads to an elevation of intracellular cAMP; subsequently Cl- leaves the cell via Cl-channels for both signalling pathways (Xue et al., 2007).

Extrinsic primary afferents containing substance P (SP) that have their cell bodies outside the wall of the gastrointestinal tract are another neural source of regulation in Cl- secretion. SP-containing extrinsic nerve fibres innervate submucosal ganglia in guinea pig small intestine (Costa et al., 1981) and in Ussing chamber experiments on guinea pig ileum, Cl- secretion is enhanced by activation of these afferents with capsaicin which stimulates release and depletion of SP from extrinsic afferents (Vanner and MacNaughton, 1995). This only occurs, however, if the submucosal plexus is intact and may be due to extrinsic activation of submucosal secretomotor neurons expressing the NK1 receptor (Vanner and MacNaughton, 1995). Further to this, extrinsic afferents release SP to stimulate secretomotor neurons in the SMP of the guinea pig colon that express tachykinin NK1 and NK3 receptors, as distention-evoked secretion is significantly reduced by capsaicin treatment (Weber et al., 2001). The capsaicin- insensitive response is inhibited by NK1 and 3 receptor antagonists highlighting that control of secretomotor neurons, and subsequently Cl- secretion, involves an extrinsic mechanism and intrinsic tachykinergic transmission within the SMP, with both sources of neural input converging onto common effectors (Weber et al., 2001).

While secretion is predominantly under the control of transmission within the SMP, there is substantial evidence that the MP has a significant function in the coordination of motility and secretion (Gwynne and Bornstein, 2007b). A small population of VIP neurons from the SMP project to myenteric ganglia (Costa et al., 2000), while submucosal afferents also project to the MP (Kirchgessner et al., 1992). Electrical stimulation of the SMP demonstrates that myenteric neurons receive synaptic input from the SMP (Monro et al., 2008). Further, the removal of the myenteric plexus in the

guinea pig small intestine reduces the number of excitatory inputs to the submucosal neurons largely responsible for secretion (Bornstein et al., 1987; Moore and Vanner, 2000; Lundgren, 2002). Therefore interconnections between both plexuses are necessary for the coordination of motility and secretion. Additionally, sensory neurons in the secretory reflex activated by toxins such as CT project to the myenteric plexus since ablating this plexus prevents net fluid secretion in the rat (Jodal et al., 1993). CT has also been shown in vivo to activate specific motility patterns in rat gut in response to distension (Kordasti et al., 2006). This CT-induced response is prevented by the serosal application of lidocaine; in other words, by the preferential anaesthetisation of the myenteric plexus. Additionally, an in vitro study has reported that CT enhances propulsive motor activity in guinea pig jejunum (Fung et al., 2010). This implies contribution from the myenteric plexus in CT-induced secretion (Kordasti et al., 2006). Thus, given the significant, yet often overlooked contribution of the MP in secretory reflexes, the majority of studies in this thesis primarily investigate neural mechanisms in the MP that underlie hypersecretion (Chapters 3-7).

3. CONSTITUENT NEURONS OF THE ENS

To date, the guinea pig ileum has been the most extensively studied model of the ENS, where at least 20 separate classes of enteric neurons have been categorised. The properties of individual enteric neurons have been investigated in some detail, but their organisation into functional circuits still requires analysis (Costa et al., 1996; Furness et al., 2000; Brookes, 2001; Gwynne & Bornstein 2007b). In order to gain a comprehensive understanding of secretomotor pathways underlying pathological secretory states, the basic elements of the enteric nervous system; enteric neurons and their propertiesth are covered in this section.

Enteric neurons are classified on a morphological, electrophysiological and neurochemical basis. Studies spanning over 30 years have correlated these properties so that summaries of functionally recognized neurons in the guinea pig small intestine exist (Brookes, 2001; Brookes and Costa, 2002; Bornstein et al., 2002; Furness, 2006). While the guinea pig is the most widely studied model, much of the data obtained is

applicable to other species. Information also exists for rats, mice and humans (Furness et al., 2004b), yet a substantial amount of detail is still unknown. Chapters 5-7 include work conducted using the mouse model and contribute to further classification of these neuronal properties in this species. Instances occur where functionally similar neurons in different species, exist in different locations in the ENS. Unlike the guinea pig where all innervation of circular muscle comes from the myenteric plexus (Wilson et al., 1987), in larger and more complex animals such as the dog and pig, circular muscle of the small intestine receives substantial innervation from the submucous plexus (Furness et al., 1990a; Timmermans et al., 1992). Additionally, the strong correlation in specific morphology and electrophysiological properties of ISNs (see below) established in the guinea pig small intestine, is not observed in the human ENS (Schemann and Neunlist, 2004), while in comparison to the guinea pig, morphologically analogous neurons in the porcine intestine also fail to exhibit typical electrophysiological properties seen in the guinea pig (Cornelissen et al., 2000; Cornelissen et al., 2001). Neurons from the intestine of the guinea pig, account for the majority of data collected over many years, but ISNs, which are common in the guinea pig intestine, are absent from the stomach corpus (Schemann and Wood, 1989a) and are rare in the antrum (Tack and Wood, 1992).

3.1 Morphology

Enteric neurons were initially classified by their morphology as first identified by the Russian anatomist Alexander Dogiel well over 100 years ago. He divided neurons of the ENS in human, guinea pig, rat, rabbit, cat and dog intestine (Dogiel, 1895b; Dogiel, 1899) into three main classes: types I, II and III. Dogiel type I neurons have short lamellar dendrites with a single axon. Dogiel type II neurons have several long processes arising from a large smooth cell body that project mainly in a circumferential direction (Bornstein et al., 1991b; Stebbing and Bornstein, 1996; Brookes, 2001). Dogiel type III neurons have relatively short dendritic processes that terminate within the ganglion of origin (Figure 1.2). While the majority of the extensive morphological data has been collected from the guinea pig, neurons that tend to serve the same functions across different species appear to have the same shape and projection patterns (Furness et al., 2004b). Small differences do occur, however, for instance smaller

mammalian species such as mice tend to have less intricate dendritic geometry than larger species. In the mouse, Dogiel type I neuronal dendrites are much less elaborate than those seen in humans (Dogiel, 1899; Nurgali et al., 2004).

Figure 1. 2 Confocal micrograph showing the morphology of enteric neurons. Two myenteric cells from the guinea pig ileum were impaled and filled with biocytin in the sharp microelectrode which allowed their morphology to be observed by post hoc staining, revealing the chief morphological neuron types. Dogiel type I are uniaxonal with short lamellar dendrites. Dogiel II cells have large, smooth cell bodies and several long processes. Figure adapted from (Gwynne and Bornstein, 2007b).

3.2 Electrophysiology

Intracellular recording studies during the 1970s classified cell types that can be differentiated based on their electrophysiological properties (Nishi and North, 1973a; Hirst et al., 1974). The classification scheme proposed by Hirst and colleagues remains the most commonly used. The scheme identifies two main cell types: AH-type neurons and S-type neurons. AH neurons are characterised by prolonged after-hyperpolarising potentials (AHP) following their action potentials under normal circumstances and are thus aptly named. S-type neurons, exhibit prominent fast excitatory post synaptic potentials (fEPSPs) following electrical stimulation of their inputs, they have a more depolarized resting membrane potential and their action potentials are not followed by prolonged hyperpolarisations (Hirst et al., 1974). A strong relationship linking Dogiel type I or II morphology with electrophysiological properties reveals that S neurons typically exhibit Dogiel type I morphology and sometimes filamentous morphology while AH neurons exhibit type II morphology (Bornstein et al., 1984, 1991b; Furness, 2006). This correlated electrophysiological classification and morphological association was first established for myenteric neurons of the guinea pig (Bornstein et al., 1984; Katayama et al., 1986) and later extended to the submucosal plexus (Bornstein et al., 1986; Bornstein et al., 1989) and has now been identified in the mouse myenteric plexus (Mao et al., 2006; Foong et al., 2012). Some exceptions do arise though, in the mouse colon a very small proportion (3/70) of Dogiel type I neurons have AH –type electrophysiology (Nurgali et al., 2004).

S neurons

S neurons fire action potentials (APs) that are followed by a only a very short after- hyperpolarisation lasting between 2-100 ms. The AP is carried by a voltage-gated Na+ conductance, as it is inhibited by the Na+ channel blocker tetrodotoxin (TTX). S neurons receive substantial fast synaptic input, where a fEPSP of sufficient amplitude will evoke an AP (Hirst et al., 1985a; Furness, 2006). In the mouse small intestine, S- type neurons display analogous electrophysiological properties (Foong et al., 2012) to those in other species along different regions of the GIT (Brookes et al., 1987; Bornstein et al., 1992; Tamura, 1992).

AH neurons

This thesis comprehensively examines the electrophysiological properties of AH neurons located in the myenteric plexus due to the important role that these neurons have in enteric neural processing and functional plasticity (Furness, 2000; Thomas and Bornstein, 2003). For this reason, their electrophysiological profile will be discussed in depth.

AH neurons exhibit several distinct electrophysiological features (Figure 1.3). An AP in an AH neuron is distinctively broad, with a duration of approximately 2.5 ms and has a large amplitude of roughly 70-90 mV (Iyer et al., 1988; Hirst et al., 1985a; Kunze et al., 1994), with analogous features observed in AH neurons of the murine small intestine (Mao et a., 2006; Foong et al., 2012). The AP itself is generated by inward sodium and calcium currents, with the opening of voltage-gated sodium and calcium channels. The calcium current is responsible for a characteristic inflection or ‘hump’ on the falling phase of the AP, since this inflection is reduced by an N-type calcium channel blocker ω-conotoxin GVIA and is therefore carried by N-type Ca²⁺ channels (Kunze et al., 1994). In the presence of TTX, APs in the somata of AH neurons will persist (Bertrand et al., 1997) as the inward currents underlying the AP include a TTX- sensitive Na⁺ current, a TTX-insensitive Ca ²⁺ current and a TTX-insensitive Nav1.9 current (Hirst and Spence, 1973; North, 1973; Hirst et al., 1985a; Rugiero et al., 2002b). Myenteric AH neurons in the mouse small intestine also exhibit a TTX-resistant Na+ current, with comparable basic parameters such as activation threshold and maximal current, and thus this is also likely carried by Nav1.9 (Mao et al., 2006). As the sodium and calcium currents diminish the AP comes to an end. The repolarisation of the AP is due to the activation of 3 outward K+ conductances that are also responsible for the early after hyperpolarisation following an AP (see below).

After-potentials and excitability in AH neurons

Following an AP, AH neurons exhibit a long after-hyperpolarising potential (AHP) as the terminology ‘AH’, coined by (Hirst et al., 1974), suggests (Figure 1.3). A large AHP greatly decreases the excitability of the soma of the AH neuron (Bertrand & Thomas 2004). The AHP consists of an early and late phase. The early phase lasts

approximately 50-100 ms, is continuous with the falling phase of the AP and is carried by three K+ currents. These conductances include a large voltage-sensitive current suppressed by external tetraethylammonium (TEA) applied to the recording bath solution which greatly reduces the amplitude of the early AHP (Hirst et al., 1985a). This current was later found to be carried by large conductance (BK) channels in whole cell recordings and was also blocked by charybdotoxin (Kunze et al., 2000). Additionally, the early AHP is carried by a transient TEA-resistant inactivating A-type K+ current

(KA), sensitive to 4-aminopyridine (4-AP) block (Hirst et al., 1985a), but not to ion substitution of external Ca2+ (Galligan et al., 1989b). A voltage-sensitive current observed in the presence of TEA carries the residual early AHP and is due to a

Hodgkin-Huxley delayed rectifier (Kdr) current (Hirst et al., 1985a).

A brief depolarisation occurs between the early AHP and the slow AHP, identified as an after-depolarising potential (ADP) (Vogalis et al., 2002b). Whole-cell patch clamp recordings from myenteric neurons have shown that the ADP is mediated by a Ca²⁺- activated depolarising cation conductance (Vogalis et al., 2002b). The relative contributions of Na⁺ and K⁺ to the ADP current are estimated to be approximately 1:5 using this technique. The amplitude of the ADP is greatly reduced in the presence of the N-type Ca²⁺ channel blockers ω-conotoxin GVIA and ω-conotoxin MVIIC and thus is largely attributable to AP-mediated Ca²⁺ entry through N-type Ca²⁺ channels (Vogalis et al., 2002b). Interestingly, ADPs are present in AH neurons in the mouse small intestine, but were only seen in the duodenum, not the ileum (Foong et al., 2012) and were blocked by the Ca2+ channel blocker cadmium chloride.

The late AHP is critical in determining the excitability of AH neurons, since the AHP has the capacity to limit firing rate and slow excitatory transmission as shown by computational modelling (Bertrand and Thomas, 2004). The hyperpolarisation can reduce the generation of action potentials to small deflections (proximal process potentials) despite an identical stimulation strength and frequency (North and Nishi, 1974). The AHP therefore acts as a ‘gate’ for APs travelling across the cell body and has a modulatory function in gating and driving outputs from AH sensory neurons (Wood and Mayer, 1979a; Bertrand and Thomas, 2004; Furness, 2006). It is known that AH sensory neurons connect together to form recurrent networks and interact through 12

slow excitatory transmission (Thomas et al., 2000) and AHPs are able to provide stability to such networks. Without the AHP, recurrent sensory networks are unstable and only small stimuli are required to change the network to a maximal firing state. With the AHP and slow excitatory transmission the network can encode graded ongoing stimuli (Thomas et al., 2000; Thomas and Bornstein, 2003; Chambers et al., 2005).

The late AHP has a time course of approximately 2-30 sec. The opening of a Ca2+ - activated potassium channel is the main underlying mechanism generating the late AHP and depends on calcium entry during AP production (Hirst et al., 1974; Hirst et al.,

1985b). The Ca²⁺ dependence of the K⁺ current (gKCa) is highlighted by the use of calcium channel blockers such as Co²⁺, Mn²⁺ or Mg², which have been shown to suppress the gKCa (Morita et al., 1982a; Hirst et al., 1985b; North and Tokimasa, 1987); a residual current remains however (Hirst et al., 1985b) which results from the release of Ca²⁺ from intracellular stores (Hillsey et al., 2000; Vogalis et al., 2001). Similarly, in the small intestine of the mouse, the slow AHP is present (Mao et al., 2006; Foong et al., 2012) and is carried by Ca2+ dependent K+ conductance, sensitive to the calcium- dependent potassium channel blocker charybdotoxin and to clotrimazole which blocks calcium-release–activated Ca2+channels (Mao et al., 2006).

Although the channels responsible for the AHP current were initially thought to be voltage-sensitive BK channels (Kunze et al., 1994), the late AHP has recently been shown to be mediated by IK channels (Vogalis et al., 2002a; Neylon et al., 2004). Binding of protein kinase A (PKA) and protein kinase C (PKC) to these IK channels and subsequent channel phosphorylation by either of these kinases is a major source of potassium channel regulation in AH neurons (Vogalis et al., 2003). An increased drive of the cAMP-pKA pathway inhibits the opening of these channels, suppresses the AHP and therefore increases cell excitability (Vogalis et al., 2003). A study using whole-cell patch-clamp recordings has indicated that the Na(v)1.9 sodium channel expressed on guinea pig AH neurons, has a role in mediating neuronal excitability (Copel et al., 2009). A specific NK3 receptor agonist was shown to increase the Nav1.9 current, lowering AP firing threshold and increasing cell excitability. The Nav1.9 current was also potentiated by PKC activation with phorbol ester (Copel et al., 2009).

In a very large proportion of myenteric neurons displaying a late AHP in both guinea pig and mouse, hyperpolarisation of the membrane triggers a non-specific depolarising cation current (Ih). This reduces the amplitude of the late AHP and is involved in re- establishing of the resting membrane potential as the late AHP decreases. Such currents are important for the regulation of cell excitability (Galligan et al., 1990; Rugiero et al., 2002b; Nguyen et al., 2005; Mao et al., 2006; Foong et al., 2012). Indeed computer modelling confirms the suppressive role Ih has on the AHP (Chambers et al., 2014).

The primary Ih channel type found in AH neurons of the guinea pig and mouse is the hyperpolarisation-activated cyclic nucleotide-gated (HCN) channel isoform HCN2, since HCN2 immunoreactivity is observed on the neuronal membranes of AH cells in both of these species (Xiao et al., 2004). Additionally, the HCN4 isoform is also localised in guinea pig myenteric AH neurons while the HCN3 isoform is localised in murine AH neurons (Xiao et al., 2004). During inflammatory states, it has been reported that the Ih is increased in AH neurons of the guinea pig distal colon and decreases the magnitude of the AHP. The Ih therefore has implications in inflammation-induced hyperexcitability of neurons (Linden et al., 2003).

Figure 1. 3 The action potential and after-potentials in AH neurons. The AP of an AH neuron is generated by an inward calcium current carried by N-type Ca²⁺ channels, a TTX-sensitive Na⁺ current and a TTX-insensitive Nav1.9 current. The AHP following an AP in the soma of an AH neuron consists of an early and late phase. The early phase is primarily carried by three outward K⁺ currents: a voltage-sensitive + BK current, an inactivating A-type K (KA) current and a voltage-sensitive delayed rectifier (Kdr) current. The temporary depolarisation between the early and late AHP, is mediated by a depolarising cation conductance activated by Ca²⁺ (gCAN). The late AHP can last up to 30 secs, and results predominately from a Ca²⁺- activated K⁺ current

(gKCa) through IK channels. As the late AHP declines, a large percentage of AH neurons exhibit a non-specific depolarising cation current (Ih) which contributes to restoration of the resting membrane potential. The inflection on the falling phase of the AP is also shown. Figure taken from Furness (2006).

3.3 Neurochemical coding of neurons

Enteric neurons express combinations of chemicals and neurotransmitters that can be localised via immunohistochemical techniques as an additional method of differentiating neuronal classes. This is known as “chemical coding”. Chemicals may include primary neurotransmitters, proteins and peptides that modulate the cellular function of other cells after their release such as Neuropeptide Y (NPY), somatostatin (SOM). Sometimes the exact physiological role of particular chemical may be unknown, for example calretinin and calbindin are calcium binding proteins that probably have a calcium buffering role and are found in neurons with a calcium current underlying APs in the guinea pig (Iyer et al., 1988; Bornstein et al., 1989; Clerc et al., 1998; Brookes, 2001). However calretinin is expressed in neurons that do not appear to have prominent calcium APs in both the guinea pig ileum (Pan and Gershon, 2000; Brookes, 2001) and murine neurons (Qu et al., 2008; Foong et al., 2012).

Occasionally neurochemical profiles of supposedly analogous neurons differ according to intestinal region and across species. SOM and NPY fibres in the small intestine of the guinea pig have opposite projection directions in the colon (Messenger and Furness, 1990), while in the mouse, myenteric Dogiel type II neurons express calcitonin gene- related peptide (CGRP), but this is not the case in the case in the guinea pig or in humans (Brehmer, 2007; Qu et al., 2008).

Immunohistochemistry combined with electrophysiological and morphological analyses of neurons, provides a powerful tool in identifying functional classes of neurons and the structure and organisation of circuits (see next).

4. FUNCTIONAL CLASSIFICATION OF NEURONS

The classification schemes outlined above when integrated together, present information pertinent to understanding function and reveal several broad functional categories of neurons which are discussed here. A full classification for each class is near complete in the guinea pig small intestine, but hasn’t been achieved to this extent in other species. The studies of this thesis examine the enteric circuitry with a focus on the MP; Chapters

3 and 4 examine elements of the myenteric circuitry in the guinea pig and chapters 5, 6 and 7 (in-part) focus on the MP in the mouse. Hence the unique functional definition, shape and neurochemical coding of neurons from the MP is summarised and compared in Table 1.1 and includes collated work from both the guinea pig and mouse small intestine. The neural elements of the SMP are not investigated to the same extent as the MP in this thesis. A main focus of this thesis includes examining the firing properties of ISNs (chapters 3 and 4), however this major neuronal subclass is not present in the SMP of the mouse. As such functional definitions of submucosal neurons have not been summarised in a table, but instead are discussed in-text only.

4.1 Intrinsic sensory neurons

AH/Dogiel type II neurons are found in both plexuses (Surprenant, 1984a; Bornstein et al., 1989; Evans et al., 1994; Reed and Vanner, 2001) and account for approximately one quarter of the neurons found in the myenteric ganglia of the guinea small intestine (Furness et al., 2004a). These neurons have been shown to project to the mucosa (Furness et al., 1990b; Song et al., 1991, 1994; Neunlist and Schemann, 1997) to other functional neuron types within the myenteric plexus (Pompolo and Furness, 1998; Furness et al., 2004a), and to neurons in the submucosal plexus (Furness et al., 1990b). In the mouse, myenteric Dogiel type II also neurons project within myenteric ganglia and to the mucosa in the colon and ileum (Furness et al., 2004b; Qu et al., 2008). AH neurons account for approximately 10% of neurons in ganglia of the submucosal plexus of the guinea pig (Furness et al., 2004a). Submucosal AH neurons in the guinea pig also project to the mucosa (Song et al., 1992), to the myenteric plexuses (Kirchgessner and Gershon, 1988a; Song et al., 1998) and make contact with other neurons in submucosal ganglia (Reed and Vanner, 2001; Furness et al., 2003a).

AH/Dogiel type II neurons have been found to have a ‘sensory’ function however this implies that they convey conscious sensation from the gut-which is untrue (Kirchgessner & Gershon, 1988). They can be, however, afferent neurons since they are the first neurons in nerve circuits and hence the first to detect the state of the intestine by responding to chemical and mechanical stimuli (Furness et al., 1998). However AH/Dogiel type II neurons aren’t always strictly ‘afferent’; they project circumferentially and synapse with other AH neurons (Pompolo and Furness, 1998; 17

Bornstein et al., 1991b; Kunze et al., 1993), where a slow EPSP evoked in one AH neuron arises from a slow EPSP in another AH neuron, allowing AH neurons to form interconnected networks (Thomas et al., 1999; Thomas et al., 2000; Thomas et al., 2004). This suggests that these neurons regulate activity around the circumference of the intestine. Since this transmission is excitatory, the networks are self-reinforcing. Given this ability to form networks, AH neurons are not exclusively afferent as they act as interneurons under these conditions.

Dogiel type II neurons with AH electrophysiology exist in the small intestine of other species such as rat (Browning and Lees, 1996) and mouse (Mao et al., 2006; Foong et al., 2012) and in the colon of the guinea pig (Lomax et al., 1999). They are also present in the mouse colon (Nurgali et al., 2004) and it is therefore likely that these neurons are ISNs in other intestinal regions and species.

AH neurons/ISNs in the MP of the guinea pig small intestine are typically immunoreactive for choline acetyltransferase (ChAT) the enzyme responsible for the synthesis of acetylcholine, the main excitatory neurotransmitter in the ENS. Additionally these neurons are immunoreactive for calbindin, tachykinin (TK) and the neurokinin (NK) 3 receptor. In the submucosal plexus, AH sensory neurons also possess ChAT/TK/calbindin immunoreactivity (Furness et al., 2004a).

In the MP of the mouse small intestine, while proportions of ISNs are similar to the guinea pig, these neurons are also immunoreactive for CGRP (Qu et al., 2008) and for neurofilament M (NF-M) which also labels for sensory neurons in the pig and human (Brehmer, 2007) (Table 1.1). Interestingly, the existence of submucosal AH/type II neurons in the mouse is yet to be convincingly confirmed. Neither immunohistochemical nor electrophysiological analyses have identified ISNs in the submucosal plexus of this species (Wong et al., 2008; Mongardi Fantaguzzi et al., 2009; Foong et al., 2014) thus a neurochemical profile is absent for these neurons.

Reflex circuits that regulate the main functions of the intestine, including secretion, are activated through the terminals of submucosal or myenteric AH neurons, found in the mucosa. The AH neurons are responsive to stretch, mechanical distortion and luminal chemicals (Gershon and Kirchgessner, 1991; Kirchgessner et al., 1992; Bertrand et al., 1997; Bertrand et al., 1998; Kunze et al., 1998; Kunze et al., 1999; Kunze et al., 2000). 18

In addition to these physiological stimuli, the terminals of AH neurons are thought to be sensitive to the actions of enterotoxins acting at the mucosa proposed to be an indirect effect mediated by 5-HT (Turvill et al., 1998; Farthing, 2000).

AH neurons in the MP of the guinea pig small intestine respond to chemicals such as acid, 5-HT short chain fatty acids and amino acids (Kunze et al., 1995; Bertrand et al., 1997; Bertrand et al., 2000; Gwynne and Bornstein, 2007a) applied to the luminal mucosa while mucosal application of glucose activates submucosal AH neurons (Kirchgessner et al., 1996).

It has been suggested that the neurons do not respond directly to chemical stimuli but rather to release of sensory mediators from the mucosa, thus activating their nerve terminals. Indeed, mechanical stimulation of the intestinal mucosa induces c-Fos expression in submucosal AH neurons, but this is blocked by a 5-HT antagonist suggesting a release of 5-HT from mucosal stores that activates sensory terminals (Kirchgessner et al., 1992). Additionally adenosine triphosphate (ATP) has been shown to stimulate the mucosal nerve terminals of myenteric AH neurons via P2X receptors, but also stimulates serotonin (5-HT) release (Bertrand and Bornstein, 2002). Further, 5- HT is known to be released from EC cells in response to glucose (Kim et al., 2001) as well as CT acting at the mucosa (Bearcroft et al., 1996).

Moreover EE cells contain other mediators including hormones such as cholecystokinin (CCK), secretin and neurotensin (Engelstoft et al., 2013; Fothergill et al., 2017) which are imperative in a variety of physiological responses. CCK is released from enteroendocrine cells by stimulation with fatty acids in humans, dogs, and rats (Liddle, 1997) and peptides induce CCK release from murine enteroendocrine cell lines (Foltz et al., 2008). CCK has been shown to activate enteric neurons in both plexuses and is a mediator in regulating enteric reflexes in the small intestine (Washington et al., 2011; Ellis et al., 2013).

AH neurons in the guinea pig myenteric plexus also possess mechanosensitive ion channels, allowing them to respond to direct distortion or stretch at their processes or through muscle movements (Kunze et al., 1998; Kunze et al., 1999; Kunze et al., 2000); while in the SMP, AH neurons have been shown to respond to distortion of the mucosal villi (Kirchgessner et al., 1992). Similarly mechanosensitive AH neurons also exist in

the mouse small intestine (Mao et al., 2006), thus it is likely most AH/type II neurons are mechanosensitive. There are also mechanosensitive myenteric S-type neurons in guinea pig and mouse intestine (Spencer and Smith, 2004; Mazzuoli and Schemann, 2009, 2012) and so, it might be that submucosal neurons with sensory functions within reflex pathways are present in mouse but this remains to be established

4.2 interneurons

Within the myenteric plexus of the guinea pig small intestine, there are interneurons in both orally-projecting/ascending pathways and anally- projecting/descending pathways. In the guinea pig ileum ascending interneurons are cholinergic and immunoreactive for calretinin, SP and enkephalin (ENK) with Dogiel type I morphology and corresponding S-type electrophysiology (Brookes, 2001; Bornstein et al., 1984). They are organised into chains running orally within the gut wall (Kunze and Furness, 1999) and are likely to be involved in propulsive gut reflexes. Descending interneurons can be subdivided into three types: nitric oxide synthase (NOS)/VIP immunoreactive neurons, ChAT/SOM neurons and ChAT/5-HT neurons (Furness et al., 1999b; Furness, 2000). 5-HT and SOM-immunoreactive descending interneurons project to the SMP (Costa et al., 1982; Song et al., 1997) and so are likely to have functions in both myenteric and submucosal ganglia. Equivalent neurochemical profiles exist in neurons in the mouse small intestine (Sang and Young, 1996; Sang and Young, 1998; Sang et al., 1997) (Table 1.1), but fully characterised summaries of the projections and roles of the different interneurons that exist for the guinea pig, remain to be determined in the mouse small intestine. However S-type/Dogiel I neurons in the myenteric plexus of mouse colon do project to other myenteric ganglia and probably function as interneurons (Nurgali et al., 2004). Interestingly, unlike the Dogiel type I morphology of the other descending interneurons in the guinea pig small intestine, ChAT/SOM interneurons have filamentous dendrites (Portbury et al., 1995; Song et al., 1996) (Table 1.1) and display electrophysiological characteristics sometimes seen in ISNs. Like most S-type neurons, SOM interneurons receive fEPSPs, which is not usually observed in AH/type II neurons in guinea pig ileum (Bornstein et al., 1994) but some also exhibit an after-hyperpolarisation following an AP and display the characteristic Ih-induced ‘sag’ during the injection of hyperpolarising current pulses (Song et al., 1997). For this reason they may account for

some descending interneurons being incorrectly classified as AH neurons/ISNs (Brookes, 2001). It is possible that these neurons may also include the S-type neurons that have been reported to comprise some descending interneurons that likely have mechanosensitive processes in the circular muscle (Spencer and Smith, 2004). The classification of interneurons within the submucosal plexus of the guinea pig is less well defined, however a small population of VIP neurons, project directly to myenteric ganglia, and not to the mucosa (where they usually innervate the epithelium and submucosal arterioles thus participating in secretomotor pathways), and are therefore likely to function as interneurons between the two plexuses (Song et al., 1998). Further, an electrophysiological study has described local interneurons within the SMP where stimulation of VIP and cholinergic neurons evokes fEPSPs in an adjacent VIP neuron (Reed and Vanner, 2001). Other electrophysiological studies show that submucosal interneurons do exist; the study of Monro et al. (2008) established that projections from the SMP to the MP are present when electrical stimulation of the SMP evoked synaptic responses in both myenteric S and AH neurons impaled over 1 mm away. Additionally Bornstein et al. (1987) demonstrated using lesioned preparations without the myenteric plexus, that fast synaptic input to submucosal neurons remained and that this was likely originating from local cholinergic interneurons within the SMP.

As previously discussed, ISNs do not appear to exist in the murine submucosal plexus, similarly, the existence of interneurons is not as clear-cut as the guinea pig, and thus the wiring of the murine circuitry may differ significantly to that of the guinea pig. In the colon of the mouse, electrophysiological studies have shown that Dogiel type I neurons do exist that receive fast synaptic input, but the exact role and hence definitive function of these neurons remains unknown (Wong et al., 2008; Foong et al., 2014).

4.3 Motor and secretomotor neurons

In mammals studied so far, excitatory and inhibitory neurons that innervate the longitudinal and circular musculature of the gastrointestinal tract (GIT) co-exist within the MP. In guinea pig, excitatory motor neurons are immunoreactive for ChAT and tachykinins, while inhibitory motor neurons contain NOS and release nitric oxide (NO) in addition to adenosine triphosphate (ATP), VIP and pituitary adenylyl cyclase activating peptide (PCAP) (Furness, 2000; Brookes, 2001). In the mouse small intestine, 21

inhibitory and excitatory motor neurons have comparable neurochemical coding, with the exception of excitatory motor neurons of the circular muscle, which are immunoreactive for calretinin in company with ACh/TK; this is not observed in the guinea pig small intestine (Qu et al., 2008) (Table 1.1).

Secretomotor or vasomotor neurons are the final neurons in the secretomotor reflex pathway that ultimately controls secretion of water and electrolytes into the lumen. They account for the large majority of cells in the submucosal plexus of the guinea pig small intestine. Secretomotor neurons include both cholinergic and non-cholinergic subtypes (Furness et al., 2003a). Non-cholinergic secretomotor neurons are VIP- immunoreactive and account for most of the neurons in this functional class. They are uniaxonal with several short irregular dendrites and project to the mucosa and submucosal arterioles (Furness et al., 2003a; Song et al., 1992). Cholinergic neurons include those that are immunoreactive for NPY, which innervate the mucosa and glands and are uniaxonal with a unique filamentous clustered dendritic structure (Song et al., 1992). The second population of cholinergic neurons are not as numerous, are immunoreactive for calretinin, project to the mucosa and submucosal arterioles and probably have stellate/unclustered dendrites and a single axon (Brookes et al., 1991a; Song et al., 1992). Each secretomotor subtype exhibits S-type electrophysiology, where prolonged AHPs are not observed in these cells and electrical stimulation of internodal strands evokes fEPSPs (Song et al., 1997; Furness et al., 2003a).

Similarly, in the SMP of the mouse small intestine, secretomotor neurons can be subdivided into cholinergic and non-cholinergic populations, but differences in neurochemical coding do exist. NPY colocalises with VIP, which does not occur in the guinea pig, while a population of secretomotor neurons that are neither cholinergic or VIP- immunoreactive are present in murine submucosal ganglia (Mongardi Fantaguzzi et al., 2009).

The MP of the guinea pig also contains a very small population of cholinergic secretomotor neurons akin to their submucosal counterparts in their neurochemical profile and projection to the mucosa (Furness et al., 1985) as well as noncholinergic, VIP- immunoreactive neurons (Costa et al., 1996).

Table 1. 1 Functional classes of neurons in the myenteric plexus of the guinea pig and mouse small intestine. Guinea pig Mouse Functional Proportion * Shape Chemical coding Proportion ^ Shape Chemical coding Definition Excitatory circular 12% Dogiel type I ChAT/TK 21% Small-medium, ChAT/TK ± calretinin muscle motor no apparent neurons dendrites Inhibitory circular 16% Dogiel type I NOS/VIP/ATP/PACAP 23% Dogiel type I NOS/VIP ± NPY muscle motor neurons Excitatory 25% Dogiel type I, ChAT/calretinin/TK 13% Small, no ChAT/calretinin ± TK longitudinal small apparent muscle motor dendrites neurons Inhibitory ~2% Dogiel type I, NOS/VIP/GABA 3% Small, no NOS/VIP longitudinal small apparent muscle motor dendrites neurons Ascending 5% Dogiel type I, ChAT/calretinin/TK ~4% Dogiel type I ChAT/TK ± calretinin interneurons large Descending 5% Dogiel type I NOS/VIP 3% Dogiel type I ChAT/NOS interneurons 2% Dogiel type I ChAT/5-HT 1% Dogiel type I ChAT/5-HT 4% Filamentous ChAT/SOM 4% Filamentous ChAT/SOM/calretinin Intrinsic sensory 26% Dogiel type II ChAT/TK/calbindin 26% Dogiel type II ChAT/NF- neurons M/CGRP/calbindin ± calretinin Table outlines functional neuron types found in the guinea pig and mouse ileum according to their defining characteristics. ATP, adenosine triphosphate; ChAT, choline acetyltransferase; CGRP, calcitonin gene-related peptide; GABA, gamma amino butyric acid; NOS, nitric oxide synthase; NPY, neuropeptide Y; PACAP, pituitary adenylyl cyclase activating peptide; SOM, somatostatin; TK, tachykinin; VIP, vasoactive intestinal peptide, 5-HT, 5-hydroxytryptamine. Table adapted from Furness (2006), Que et al. (2008) and Sang and Young (1996). * Proportion of total number of myenteric neurons labelled by a ‘nerve cell body’ antiserum (Costa et al., 1996). ^ Proportion of total number of myenteric neurons labelled by Hu (Qu et al., 2008). 23

5. SYNAPTIC TRANSMISSION IN AH AND S NEURONS

5.1 Synaptic transmission in AH neurons

A number of different synaptic potentials displayed by various types of enteric neurons have been recognized in the guinea pig small intestine. These classifications are made based on time course, (fast, slow or intermediate), changes in membrane potential (excitatory or inhibitory) and pharmacology (Gwynne and Bornstein, 2007b). Excitatory post synaptic potentials (EPSPs) can be placed into 3 broad categories: fast EPSPs (fEPSPs) lasting 30-50 ms, slow EPSPs (sEPSPs) lasting up to several minutes and intermediate EPSPs lasting 150 ms-2.5 s (Bornstein et al., 1986; Furness et al., 1998; Monro et al., 2004; Gwynne and Bornstein, 2007b; Gwynne and Bornstein, 2009).

A prolonged synaptic event, known as sustained slow post-synaptic excitation (SSPE) is also observed is some enteric neurons, predominately myenteric AH neurons. The SSPE is induced by a repeated low frequency train stimulation of synaptic inputs and involves increased soma excitability, which can outlast the nerve stimulation by minutes (Clerc et al., 1999; Alex et al., 2001). AH neurons typically exhibit slow excitatory post synaptic potentials triggered by a train of electrical stimuli applied to interganglionic fibres and associated with a K⁺ conductance decrease, a Cl¯ conductance increase and a membrane depolarisation of 3-30 mV (Katayama and North, 1978; Wood and Mayer, 1978; Johnson et al., 1981; Bornstein et al., 1984; Bornstein et al., 1986; Bertrand and Galligan, 1994; Starodub and Wood, 2000a). sEPSPs seem to be a highly conserved feature of AH cells and are evoked in AH neurons in the MP of the mouse small intestine (Mao et al., 2006; Foong et al., 2012) and colon (Nurgali et al., 2004), they have also been recorded in the rat and human (Browning and Lees, 1996; Brookes et al., 1987).

A number of substances found in nerve terminals within the ENS can imitate a slow EPSP, however tachykinins acting at neurokinin (NK)1 and NK3 receptors are the principal transmitters in AH neurons (Bertrand and Galligan, 1995; Alex et al., 2001; Johnson and Bornstein, 2004). Slow EPSPs evoked by stimulation of circumferential pathways are blocked by NK1 and NK3 antagonists (Hu et al., 2003). There is also considerable evidence that suggests 5-HT mediates slow EPSPs via a number of 5-HT receptor subtypes (Takaki et al., 1985; Mawe et al., 1986; Monro et al., 2005), where the specific receptors responsible

have proven difficult to define. Additionally, acetylcholine acting on M1 muscarinic receptors has also been shown to mediate some slow EPSPs (North et al., 1985). As mentioned previously, a characteristic feature of slow EPSPS in AH neurons is that they suppress the AHP (Grafe et al., 1980). Since AH neurons form recurrent networks and transmit to each other via slow EPSPs, the AHP is an obvious candidate for controlling slow EPSP-mediated excitation at the network level (Thomas et al., 2000; Thomas and Bornstein, 2003; Chambers et al., 2005). It should also be noted that during the SSPE, AHPs are similarly suppressed (Clerc et al., 1999).

Fast excitatory inputs to myenteric AH neurons are rare and when recorded, AH neurons exhibit fast EPSPs of low amplitude (Iyer et al., 1988; Bornstein et al., 1994; Furness, 2006). Since myenteric AH neurons are known to express nicotinic and purinergic P2X receptors, fast transmission in these neuron is believed to be nicotinic and purinergic (Kirchgessner and Liu, 1998; Castelucci et al., 2002), and may also be mediated by glutamate acting through AMPA receptors (Liu et al., 1997). In the murine MP, fast transmission in AH neurons also only rarely occurs (Mao et al., 2006). Similarly, fEPSPs are not often seen in submucosal AH neurons (Bornstein et al., 1989; Evans et al., 1994).

5.2 Synaptic transmission in S neurons

Slow synaptic transmission occurs in S-type uniaxonal neurons in myenteric and submucosal ganglia in both the guinea pig and mouse, but is not as well characterised as sEPSPs in AH/Dogiel type II neurons. Due to a diverse range of functional subtypes of S/uniaxonal neurons, comprised of sub-populations including inhibitory/excitatory neurons supplying the muscle coats, ascending/descending interneurons and secretomotor neurons that have outputs to number of myenteric and submucosal neurons (Pompolo and Furness, 1995; Pompolo and Furness, 1998; Moore and Vanner, 2000; Bornstein et al., 2002; Bornstein et al., 2004) the entire range of mediators facilitating slow transmission is not clear (Gwynne and Bornstein, 2007b). In the guinea pig, NK1 receptors mediate sEPSPs in NOS-immunoreactive inhibitory motor neurons in the MP, but not in descending interneurons (Thornton and Bornstein, 2002), while in submucosal ganglia, sEPSPs in non-cholinergic secretomotor neurons are mediated by P2Y receptors and metabotropic glutamate receptors (mGluR1) (Hu et al., 2003; Foong and Bornstein, 2009).

In the mouse, sEPSPs are exhibited by myenteric S neurons in the small intestine and submucosal S neurons in the colon, but the transmitters mediating these are even less well- defined than those in guinea pig (Foong et al., 2012; Foong et al., 2014).

As indicated by their nomenclature, S neurons typically receive fast synaptic input. The prevailing component of fast transmission is cholinergic (Nishi and North, 1973a; Hirst et al.,

1974) but it also involves ATP acting at P2X receptors and 5-HT at 5-HT3 receptors since resistant fEPSPs after nicotinic receptor blockade are reduced in amplitude or inhibited by the P2X receptor antagonist pyridoxalphosphate-6-azophenyl-2',4'-disulfonic acid (PPADS) or a

5HT3 receptor antagonist (LePard et al., 1997; Zhou and Galligan, 1999). Similarly, fast transmission is common in secretomotor neurons in the SMP and is attributable largely to nicotinic receptors, but also to P2X and 5-HT3 receptors (Bornstein et al., 1986; Monro et al., 2004). fEPSPs are observed in S neurons in both plexuses of the intestine of the mouse. In the MP of the mouse small intestine ACh and ATP mediate fast synaptic transmission as fEPSPs are only partly inhibited by the cholinergic receptor antagonist mecamylamine but are also reduced by PPADS (Bian et al., 2003; Ren et al., 2003) indicating that ACh and ATP mediate fast synaptic transmission in the murine MP as occurs in the guinea pig. Similarly in the colon of the mouse fEPSPs are blocked largely by nicotinic receptor antagonists with a small component of fast transmission reduced by PPADs (Nurgali et al., 2004). It appears, though, that fast transmission in the SMP of the mouse colon, is entirely cholinergic (Foong et al., 2014).

While inhibitory post synaptic potentials (IPSPs) do occur in a small population of myenteric neurons (Johnson et al., 1980; Hodgkiss and Lees, 1984; Johnson and Bornstein, 2004) they are more frequently encountered in submucosal S neurons in the guinea pig small intestine (Hirst and McKirdy, 1975; North and Surprenant, 1985; Bornstein et al., 1986; Bornstein et al., 1988; Foong et al., 2010b) and to a lesser extent in mouse colon (Wong et al., 2008; Foong et al., 2014). In the guinea pig, they are most prominently observed in VIP- immunoreactive secretomotor neurons and are associated with an increase in membrane K+ conductance and thus a decrease in input resistance (North and Surprenant, 1985; Evans et al., 1994; Bornstein et al., 1986). They are mediated by the activation of extrinsic sympathetic nerve terminals and the release of noradrenaline acting at α2-adrenoreceptors (Hirst and McKirdy, 1975; North and Surprenant, 1985) and by myenteric non-adrenergic 26

input (Mihara et al., 1987; Bornstein et al., 1988) as IPSPs cannot be evoked in myectomized preparations and agents that block the action of noradrenaline on α2-adrenoceptor have no effect on IPSPs in extrinsically denervated preparations (Bornstein et al., 1988). Intrinsic neurotransmitters mediating IPSPs have been found to be somatostatin (SOM) acting via

SST1 and SST2 receptor subtypes (Shen and Surprenant, 1993; Foong et al., 2010b) and 5-HT via 5-HT1A receptors (Foong et al., 2010b).

Electrical stimulation evokes intermediate EPSPs in myenteric S neurons (Gwynne and Bornstein, 2009) and some submucosal S neurons (Monro et al., 2004), these are blocked by a P2Y1 receptor antagonist.

6. REFLEX CIRCUITRY OF SECRETION

In conjunction with mechanical stimuli, such as distention (Weber et al., 2001) and mucosal stroking (Christofi et al., 2004; Cooke et al., 2004) as well as chemical stimulation (Hansen, 2003), secretomotor reflex pathways are activated by noxious stimuli like toxins and pathogens (Lundgren and Jodal, 1997).

The basic elements of such pathways include sensory AH neurons with their cell bodies in either the SMP or MP and their terminals in the mucosa, as the first (primary afferent) neurons activated during secretomotor reflexes. Secretomotor neurons in the SMP are the effector (efferent) neurons of secretomotor pathways (Furness, 2006). Reflex secretion can occur through an axon reflex in the soma of a sensory neuron, via a direct synapse between a submucosal sensory neuron and a secretomotor neuron in the SMP, or through a long reflex that passes through the myenteric plexus which impinges onto submucosal secretomotor neurons (Furness, 2006). It is proposed that secretomotor reflexes activated by CT project via the MP (Kirchgessner et al., 1992; Jodal et al., 1993). In the murine ENS, since in ISNs do not appear to exist in the SMP, it is likely that such long secretomotor reflexes passing through the MP also occur. Three types of secretomotor neurons exist in the guinea pig small intestine (as previously discussed) with comparable counterparts in the murine small intestine: noncholinergic VIP secretomotor/vasodilator neurons, cholinergic NPY secretomotor neurons and cholinergic calretinin vasodilator neurons. Vasodilator neurons

innervate small arterioles in the gut wall, facilitating fluid delivery to the mucosa, hence enhancing secretion. In this way, secretion is coupled to vasodilation (Furness et al., 2003a; Furness, 2006). Together these secretomotor neurons act at the epithelial villi to stimulate water and electrolyte transport across the mucosa (Furness et al., 2004a). How luminal exposure to CT and TcdA perturb this system is a key focus of this thesis. While both bacterial exotoxins have sites of action at the mucosal surface, they also operate through neural pathways that, while not fully defined, are distinctly different. Knowing in what way changes are mediated requires an understanding of what is currently known about the interactions of these enterotoxins with the ENS, the details of which are discussed below.

7. MECHANISMS FOR PATHOGENIC CT AND TCDA-INDUCED HYPERSECRETION

7.1 Direct effect of CT on enterocytes (non-neural mechanism)

While pathogenic hypersecretion occurs via activation of the ENS, CT binds to the mucosa to activate a local mechanism. CT is a protein with a molecular mass of approximately 84 000 Da, and is comprised of six subunits: 1 A subunit and 5 B subunits. CT binds to mucosal enterocytes where the B subunits of CT bind to a specific membrane ganglioside (GM1) on the apical membrane (Holmgren et al., 1975). The enzymatic A subunit is endocytosed into the enterocyte and, after being split in the lysosomes, the activated enzyme links adenosine diphosphate (ADP) ribose to a guanine nucleotide-binding protein (G protein), which activates adenylate cyclase. The result is an increase in intracellular cyclic AMP (Sharp and Hynie, 1971), which then opens cystic fibrosis transmembrane conductance regulator (CFTR) channels found on enterocytes through a cascade of steps and promotes chloride and fluid secretion (Lundgren and Jodal, 1997; Farthing, 2000; Burleigh and Banks, 2007). The ADP- ribosylation of the G protein prevents it being inactivated, thus effectively leading to irreversible activation of adenylate-cyclase. This alone, was thought to be the primary mechanism governing secretion- but the toxin acts mainly at the villus tips (Weiser and Quill, 1975) and never reaches the crypts during prolonged incubations. CFTR channels are not expressed at villus tips, but close to crypts (Ameen et al., 1995) thus secretion occurs at the crypts which rely on indirect mechanisms for their local control. Additionally the toxin does not penetrate the mucosa (Gwynne et al., 2009) to interact directly with the underlying 28

circuitry. It is therefore likely that indirect activation of enteric neural pathways mediates the pathogenic hypersecretion produced by CT (see below).

7.2 Direct effect of TcdA (non-neural mechanism)

C.diff produces two large exotoxins, toxin A (TcdA), one of the largest known bacterial toxins, and toxin B (TcdB) which are approximately 308 kDa and 275 kDa, respectively. CT produces watery diarrhoea without inflammation, but the hypersecretion produced by Clostridial toxins is distinctly different, in that it is accompanied by an inflammatory response with infiltration of neutrophils and tissue necrosis, which is highly characteristic of the bacterial infection (Kelly et al., 1994; Pothoulakis et al., 1998; Kelly and Kyne, 2011). Well established in vivo studies in rabbits, rats, hamsters and mice have shown that these effects appear to be mediated primarily by TcdA and hence it is the chief component responsible for the pathophysiology of C.diff in animal models, but this not the case in humans (Lyerly et al., 1985; Mitchell et al., 1986; Triadafilopoulos et al., 1987; Sun et al., 2010). TcdB is cytotoxic to mammalian cells in culture (Thelestam and Florin, 1984; Hecht et al., 1992), but does not possess enterotoxic effects in rodent models. In the rabbit ileal loop, TcdA, but not TcdB, elicits fluid secretion, tissue damage, an increase in mannitol permeability and a higher concentration of neutrophils in the perfusate (Mitchell et al., 1986; Triadafilopoulos et al., 1987). While in hamsters, rats and mice, TcdA but not TcdB elicits intestinal fluid secretion (Lyerly et al., 1982). Moreover, immunisation against TcdA enables protection from C. diff infection (CDI) in rodents (Kim et al., 1987; Corthier et al., 1991). Why TcdB does not produce enterotoxic effects in rodent models is not clear, it may, be due to an absence of the TcdB receptor site in rodent tissue. In the hamster, radiolabelled TcdB failed to show TcdB binding sites (Rolfe, 1991) in an animal model known to be insensitive to the effects of TcdB in vivo (Lyerly et al., 1985). In contrast, rabbits and hamsters possess TcdA receptors and are sensitive to the intestinal effects of TcdA (Lyerly et al., 1985; Mitchell et al., 1986; Rolfe, 1991). Further in rabbit and mouse, TcdB has been shown to preferentially bind to the basal membrane of the epithelium and so if TcdB receptors exist they may not be exposed to luminal contents (Sutton et al., 2008). It may be that once the mucosal barrier of the intestine is compromised by the effects of TcdA, TcdB may then gain access to its receptors (Sun et al., 2010).

The story is quite different when considering human data. In the human colon, it has been observed that TcdB plays an important role in the pathogenesis of infection. It was generally considered that the development of C.diff was mediated via strains producing both TcdA and TcdB however, cases of clinical isolates of TcdA- TcdB+ strains have been documented (Kato et al., 1999; Limaye et al., 2000) and more recently, TcdB+ strains have been responsible for numerous nosocomial outbreaks world-wide (Al-Barrak et al., 1999; Alfa et al., 2000; Kuijper et al., 2001; Goorhuis et al., 2009) with a tendency towards more severe disease in patients infected with TcdA- TcdB+ strains (Drudy et al., 2007). Indeed it has been demonstrated that TcdB is 10 times more potent than TcdA in damaging the human colonic epithelium in vitro, reducing epithelial resistance and increasing mannitol permeability (Riegler et al., 1995). Further, Savidge et al. (2003) examined the effects of TcdB on human intestine maintained as xenografts in mice in vivo. Like TcdA, TcdB induced epithelial damage, increased mucosal permeability and inflammation. Collectively this suggest that in man, TcdB serves a much more important role in C. diff pathogenesis than initially thought, particularly demonstrated by the rise of TcdB+ outbreaks. Since TcdA and TcdB are important in disease pathogenesis in humans, both should be considered in experimental models of C.diff and targeted in therapy development. However, the ENS of rodent animals is the most comprehensively understood model and an ideal one to examine the effects of TcdA. Consequently the effects of TcdA on the ENS in guinea pigs and mice have been investigated in this thesis and for this reason the actions of TcdA will be discussed in detail.

TcdA initiates the hypersecretory and inflammatory response by binding to carbohydrate cell surface receptors on the plasma membrane of enterocytes facing the intestinal lumen, where it is internalised (Rolfe and Song, 1993; Pothoulakis and Lamont 2001). This produces a disruption of the actin cytoskeletal framework via modification of GTP-binding Rho proteins, by glucosylation (Just et al., 1995; Aktories and Just, 1995). Rho proteins are involved in receptor-mediated regulation of the actin cytoskeleton, this inactivation of Rho function leads to disaggregation of actin filaments and impairment of tight junctions. In cultured cells this causes an increase on epithelial permeability and a decrease in transepithelial resistance (Hecht et al., 1988; Riegler et al., 1995). This does not occur following the binding of CT to enterocytes which stimulates secretion via an increase in adenylate cyclase.

While the toxin can affect epithelial barrier function and permeability in cell culture, these processes do not address the inflammatory hypersecretion observed in rodent models in vivo. Strong evidence exists that the inflammatory diarrhoeal response following the direct action of the toxin is neurally-mediated.

7.3 Neural mechanism for CT-induced hypersecretion

Early studies in the rat and cat showed that CT–induced hypersecretion in vivo is prevented when neural activity is inhibited by the neurotoxin TTX or lignocaine, a local anaesthetic (Cassuto et al., 1981a), a result confirmed in many subsequent studies (Lundgren, 2002; Burleigh and Banks, 2007). Blocking nicotinic receptors prevents CT-induced hypersecretion, while the hypersecretion is accompanied by a release of VIP, a known potent secretagogue, (Eklund et al., 1979) into the venous effluent (Cassuto et al., 1981b; Cassuto et al., 1982a), thus pathways must involve nicotinic synapses and the actions of VIP. These studies along with a range of in vitro observations have provided a basic model of CT pathogenesis, yet the pathways involved are still not entirely defined.

Figure 1.4 represents the standard model for CT action where it is proposed that CT produces its pathogenic effects via a neuronal reflex arc consisting of a sensory neuron with terminals in the mucosa, an interneuron in the MP, and a secretomotor efferent in the SMP (Farthing, 2000). CT binds to mucosal enterochromaffin (EC) cells which produce a large and ongoing release of 5-HT probably via the same irreversible activation of adenylate-cyclase that occurs during CT’s interaction with mucosal epithelium mentioned above (Bearcroft et al., 1996). The 5-HT released by EC cells excites intrinsic sensory neurons via 5-HT₃ receptors on their mucosal terminals (Bertrand et al., 2000). In vivo, the 5-HT₃ receptor antagonist granisetron, abolishes the hypersecretion induced by intraluminal CT (Mourad et al., 1995; Kordasti et al., 2006), suggesting that EC cells transmit the luminal signal to the 5-HT-sensitive nerve terminals. Intrinsic sensory neurons then excite interneurons in the MP. The MP appears to be essential in toxin-induced hypersecretion; chemical ablation of this plexus with benzalkonium chloride prevents the secretory reflex (Jodal et al., 1993), stripping off the MP- attenuates the maximal secretory response to CT (Carey and Cooke, 1986), while projections of the MP have been shown to pass through the SMP and reach the mucosa (Furness et al., 1990b; Song et al., 1991, 1994; Neunlist and Schemann, 1997). CT-induced hypersecretion is prevented by the nicotinic receptor antagonist hexamethonium (Cassuto et al., 1982a) and by tachykinin 31

antagonists (Turvill et al., 2000a), indicating the participation of nicotinic and tachykinergic synapses in this reflex pathway. The efferent limb of the reflex is likely to be mediated by VIP secretomotor neurons found in the SMP; CT evokes a massive release of VIP (Eklund et al., 1979; Cassuto et al., 1981b) and the hypersecretion is inhibited by VIP antagonists (Mourad and Nassar, 2000; Banks et al., 2005). Since VIP is known to be located exclusively in intestinal neurons (Cassuto et al., 1981a), particularly in a large portion of non-cholinergic secretomotor/ vasodilator neurons in the SMP (Furness, 2006), secretion induced by CT is probably mediated by VIP released from non-cholinergic secretomotor/vasodilator neurons. VIP binds to VPAC1 receptors on the basolateral membrane of epithelial cells and activates adenylate cyclase cAMP, this opens the apical CFTR channels, driving Cl¯ and fluid secretion (Farthing, 2000; Lundgren, 2002; Burleigh and Banks, 2007) (Figure 1.4 A).

7.4 Sustained CT-induced hyperexcitability of reflex ‘output’

An electrophysiological study has shown that CT incubated in vitro induces sustained hyperexcitability in both cholinergic and non-cholinergic submucosal secretomotor neurons of the guinea pig jejunum i.e. CT pre-treatment greatly enhances their firing (Gwynne et al., 2009). This is consistent with studies contributing to the standard model of CT action, in that the luminal co-incubation with TTX, hexamethonium, 5-HT₃ antagonists and NK antagonists prevent the toxin-induced hypersecretion, indicating that the effect is neurally mediated and specific synapses must be active to elicit the effect (Cassuto et al., 1981b, 1982a; Mourad et al., 1995, Bertrand et al., 2000; Turvill et al, 2000a). How this finding differs from the conventional model, however, is that it suggests a mechanism for a self-sustained excitation of the secretomotor neurons. The excitation is independent of the ongoing release of 5-HT due to the irreversible activation of EC cells, since the hyperactivity of the neurons persists for many hours after CT exposure during which time it is unaffected by 5-HT3 receptor blockade (Gwynne et al., 2009) (Figure 1.4 B). In agreement, Kordasti et al. (2006) demonstrated that ongoing 5-HT is not necessary for CT-induced effects; in the rat jejunum in vivo, in addition to hypersecretion, CT activates specific motility responses which are present well after toxin washout. Further, these motility responses are enhanced by antagonism of 5-HT3 receptors.

Interestingly, the hyperexcitability of the secretomotor neurons is not accompanied by increased spontaneous firing of APs, they may be considered hyperresponsive rather than hyperactive in the absence of an external stimulus. Furthermore, the hyperexcitability is confined to neurons close to mucosa (Gwynne et al., 2009), and so it requires an ongoing input from the mucosa of some kind, but it remains to be determined whether this is direct or via a neural pathway.

Sustained increases of firing occurring hours after initial toxin exposure suggests that there is an amplification of the secretomotor pathway and illustrate that this reflex is much more complex than the conventional pathway. While augmented firing is observed in the efferent arm, it is unknown whether other elements in the circuitry are changed- Chapter 3 addresses this question.

Figure 1. 4 The models for the secretory reflex pathway in the ENS. The conventional model of CT –induced hypersecretion is shown in (A). The main disturbance (circled) occurs at the ‘input’ to the pathway; where the overactivity of secretomotor neurons is due to an ongoing release of 5-HT by the irreversible activation of EC cells. The alternative model (B) suggests a mechanism for a self-sustained excitation of secretomotor neurons, independent of 5-HT release. The disturbance is at the ‘output’. Figure adapted from Bertrand et al. (2000).

7.5 Neural and inflammatory mechanism for TcdA-induced hypersecretion

TcdA has been shown in vivo to produce acute hypersecretion in company with inflammation and epithelial damage (Kelly et al., 1994; Castagliuolo et al., 1998a; Kelly and Kyne, 2011), when incubated in rabbit, rat, mouse and hamster ileal loops (Lyerly et al., 1985; Castagliuolo et al., 1994; Castagliuolo et al., 1998a; Pothoulakis et al., 1994; Savidge et al., 2011; de Araújo Junqueira et al., 2011) (Figure 1.5). The effect is neurally mediated, since the addition of lidocaine and hexamethonium prevents it (Castagliuolo et al., 1994). Functionally ablating primary afferent neurons thorough capsaicin treatment inhibits the TcdA-induced response (Castagliuolo et al., 1994) and prior exposure of the ileum to antagonists of calcitonin gene-related peptide (CGRP) (Keates et al., 1998) and substance P (SP) which are neurotransmitters found in extrinsic and some intrinsic sensory neurons, similarly prevents the response (Pothoulakis et al., 1994). Additionally, the neuropeptide Corticotropin- releasing hormone (CRH) is required for TcdA mediated intestinal inflammation, since CRH receptor-deficient mice demonstrate reduced inflammatory responses to ileal administration of TcdA (Kokkotou et al., 2006). Taken together, this suggests that TcdA requires activation of extrinsic neural mechanisms to elicit its full actions (Pothoulakis et al., 1998). In contrast, chemical ablation of extrinsic afferents or pretreatment with a SP antagonist in the same animal models does not prevent CT-induced effects (Castagliuolo et al., 1994; Pothoulakis et al., 1994). CT does not therefore, operate through extrinsic pathways, instead it relies on neurotransmitters such as 5-HT and VIP within the ENS to elicit its actions.

The pathway through which extrinsic neural afferents are activated probably involves release of cytokines from enterocytes following TcdA binding at the epithelium. These include leukotriene (LT) B4, prostaglandins and tumor necrosis factor (TNF)-α in animal models and interleukin (IL)-8 in humans; inhibition of cytokine synthesis reduces TcdA-evoked intestinal responses (Pothoulakis and Lamont 2001). Indeed, injection of TcdA into mouse ileal loops increases the local tissue production of cytokines such as TNF-α and interleukin-1β (IL-1β) (de Araújo Junqueira et al., 2011).

The cytokines diffuse into the mucosa and extrinsic afferents with their cell bodies in the dorsal root ganglia (DRG) are activated (Pothoulakis et al., 1998). Cytokines probably also activate enteric nerves; while denervation of extrinsic afferents in the rat attenuates the inflammatory response and tissue damage induced by TcdA, fluid secretion occurs when

extrinsic nerves are denervated, but is prevented by hexamethonium and not granisetron (Sörensson et al., 2001). Sörensson et al. (2001) attributed fluid secretion in the absence of extrinsic supply to activation of enteric nerves not by 5-HT released from the mucosa but possibly via cytokines released by TcdA which have been shown to excite enteric nerves (Dekkers et al., 1997a; Lakhan and Kirchgessner, 2010).

The activation of extrinsic afferents by cytokines occurs early on in the signalling cascade, since treatment of rats with SP antagonists 30 mins after initial TcdA exposure does not prevent TcdA-mediated responses (Pothoulakis et al., 1994).

Transmitters such as SP and CGRP released from extrinsic afferents trigger subsequent secretory and inflammatory responses. Mast cells are activated following in vivo administration of TcdA; significant mast cell degranulation in the rat as shown by electron microscopy occurs with increased mucosal levels of the mucosal mast cell enzyme- rat mast cell protease II (RMCP II) as early as 15 mins after exposure (Castagliuolo et al., 1994). Further to this, in response to TcdA, mast cell deficient mice show a reduction in fluid secretion (Wershil et al., 1998). In humans both TcdA and TcdB are known to trigger mast cell degranulation directly (Meyer et al., 2007). The products of mast cell degranulation, such as leukotriene B4, leukotriene C4 and platelet-activating factor (PAF) are thought to directly act on enterocytes themselves and contribute to secretion into the lumen (Pothoulakis et al., 1993). NO is known to regulate mast cell function by inhibiting histamine and PAF release (Kanwar et al., 1994), and likely plays a protective role against TcdA in the gut. In fact, pretreatment of rats with the nitric oxide synthase (NOS) inhibitor nitro-Larginine methyl ester (L-NAME ) increased TcdA-induced mucosal permeability and secretion while a NO donor inhibited TcdA-mediated mast cell degranulation and secretion (Qiu et al., 1996). More recently it has been demonstrated that a part of this nitric oxide based protection against TcdA is mediated by S-nitrosothiols, which are NO intermediates generated during CDI (Savidge et al., 2011).

In addition, TcdA stimulates extensive neutrophil recruitment and consequential tissue damage (Kelly et al., 1994; de Araújo Junqueira et al., 2011). Intestinal mast cells are critical in TcdA –induced neutrophil activation since mast cell deficient mice display reduced neutrophil infiltration in response to TcdA (Wershil et al., 1998; Sun et al., 2010). Antibodies 36

raised against the neutrophil adhesion molecule CD18, inhibit the infiltration in vivo in rabbit ileal loops and prevent mucosal tissue erosion and reduce fluid secretion (Kelly et al., 1994) while the neutrophil rolling inhibitor Fucoidin prevents issue injury and inflammation in mouse ileal loops (Barreto et al., 2008). Further in a clinical setting, an increase in neutrophils in the peripheral blood is also a marker of severe CDI and is used as a prognostic tool (Cloud et al., 2009; Pepin et al., 2009). It has been suggested that the enteric pathways activated by CT and TcdA are distinct with no commonalities or overlap, where CT produces diarrhoea via activation of cholinergic and VIPergic pathways and TcdA activates a separate pathway involving mucosal mast cells and release of SP and CGRP from extrinsic sensory neurons (Pothoulakis and Lamont 2001). However, while CT and TcdA operate via distinct neural pathways, it has also been suggested that TcdA operates in part via the ENS. Firstly, the TcdA-induced release of cytokines across the epithelium as discussed above and the products of mast degranulation can activate enteric nerves (Nemeth et al., 1984; Frieling et al., 1994; Xia et al., 2000; Liu et al., 2003; Wood, 2006). Secondly, any ENS contribution in the effects of TcdA may be secondary to the inflammation TcdA is known to produce. In models of trinitrobenzene sulfonic acid (TNBS ) –induced inflammation in the guinea pig, enteric neurons show signs of increased excitability including increased firing, spontaneous activity and supression of membrane properties such as the AHP in AH neurons (Linden et al., 2003; Lomax et al., 2005)). The effects of inflammation on enteric neuronal properties can persist for days or weeks after the initial insult and may form the foundation for post-inflammatory conditions such as irritable bowel syndrome (Mawe et al., 2009). Thirdly, the activation of extrinsic efferents may also induce enteric neuronal activation. SP activation of enteric neurons in response to intraluminal TcdA in the rat ileum was examined by immunocytochemical analysis of SP Receptor (SPR) endocytosis as an indication of SPR activation (Mantyh et al., 1996). TcdA induced SPR endocytosis in enteric neurons of both the SMP and MP and it was therefore postulated that TcdA stimulates SP release from extrinsic afferents, with the SP then acting act on SPR- immunoreactive enteric neurons to initiate an intrinsic neural reflex via an unknown mechanism.

Furthermore Xia et al. (2000) proposed that since both mast cell-deficient and SP receptor (NK1-R)-deficient mice show attenuated responses to TcdA (Castagliuolo et al., 1998a; Wershil et al., 1998), enteric nervous input may reinforce mast cell degranulation since SP is 37

a putative neurotransmitter for slow excitatory transmission in the ENS (Katayama and North, 1978; Surprenant, 1984a; Ren et al., 2000; Johnson and Bornstein, 2004; Gwynne and Bornstein, 2007b).

Subsequently, the direct action of TcdA on enteric neurons was examined and application of TcdA by micro-pressure injection was shown to enhance the excitability of enteric AH and S neurons in the SMP of the guinea pig small intestine (Xia et al., 2000). TcdA induced membrane depolarisation and spontaneous discharge for several seconds in both neuronal subtypes and a suppression of the AHP in AH neurons. The effects were due to direct excitation of TcdA on the cell body since TTX did not inhibit the response, further the excitability was not attributed to mast cell degranulation since the application of a H2 histamine receptor antagonist failed to reduce the effect. Additionally, TcdA supressed IPSPs evoked by stimulation of sympathetic postganglionic nerve fibres. TcdA could therefore be contributing to hyperstimulation of mucosal secretion via direct excitation of the secretomotor neurons responsible for driving secretion or by the suppression of IPSPs, potentially eliminating sympathetic inhibition of submucosal neurons and permitting increased firing within enteric microcircuits. Whether changes in firing activity in other elements of the enteric circuitry occur remains to be determined. This thesis investigates this by luminal in vitro and in vivo incubations with TcdA (Chapters 4 and 7 respectively).

Figure 1. 5 TcdA-induced epithelial damage and inflammation in the mouse.

Histological sections of mouse ileum following an in vivo incubation with either saline (A) or TcdA (B). Saline-treated tissue displays healthy epithelial architecture with intact villi. TcdA produces severe epithelial necrosis with a break down in villus architecture and an infiltration with inflammatory cells. Figure taken from Castagliuolo et al. (1994).

7.6 Interaction of C.diff and the ‘microbiome’

The GIT is host to a unique ecosystem of bacteria termed the microbiome. The human intestinal microbiome contains an extensive array of microbial taxonomic profiles (Methé et al., 2012). While the composition of the microbiome is important in normal gut physiology, contributing to metabolic functions and protecting against pathogens, it may establish risk factors for various diseases including, obesity, colorectal cancer, diabetes and inflammatory bowel disease (Dupont, 2014; Shreiner et al., 2015). Further, opportunistic C.diff colonization and infection primarily occurs after antibiotic use, producing a state of dysbiosis, and restoration of a normal microbial ecosystem via faecal microbial transplantation (FMT) has immense clinical benefit (CDC, 2013). Moreover, work from a collaborator demonstrates that CDI is associated with compositional and functional changes in the microbiota and stool metabolites distinctly associated with recurrent CDI are lost after successful FMT (Dann et al., 2015) (Figure 1.6). Thus the microbial community structure is inextricably linked with gut physiology in healthy and disease states.

Evidence of microbial interaction with the ENS in animal models suggests the presence of bacteria in the lumen of the gut involves signalling along a gut-microbiota axis. AH neurons in the myenteric plexus of germfree mice lacking microbiota show suppressed excitability, which is restored following colonisation with bacterial communities (McVey Neufeld et al., 2013). It is unknown which bacteria signal to the MP and whether this occurs directly or indirectly via epithelial intermediates; spore-forming bacteria from mouse and human microbiota have been shown to promote 5-HT biosynthesis from colonic EC cells (Yano et al., 2015; Savidge, 2016). Additionally, intestinal bacteria are necessary for the development and function of the ENS. Germfree mice show decreased proportions of myenteric nitrergic neurons without changes to other neuronal subtypes (Anitha et al., 2012). In germfree rats specific microbial species have been shown to promote or suppress migrating myoelectric complex (MMC) activity (Husebye et al., 2001). Taken together, the microbiome influences the development and activity of enteric neural networks. Further to this, in the human ENS, interactions can occur between pathogenic bacteria and glia surrounding enteric neurons where human enteric glial cells have been shown to discriminate between pathogens and probiotics via different receptor signalling mechanisms (Turco et al., 2014). Thus, microbial interactions with the ENS appear to occur in the human as they do in rodent models.

Shifts in microbial communities and their associated metabolites are thought to alter enteric function and progression of disease states. Of particular interest is the role bacterial metabolites play in recurrence of CDI. Recurrence of CDI remains a major hurdle in its treatment and occurs in up to 50% of patients often over several years, despite antibiotic treatment (Lessa et al., 2015). Data from our collaborator examining both human and animal models suggest that microbial-derived GABA may be associated with susceptibility and recurrence of CDI (Dann et al., 2015). Metabolome profiling of stool samples in patients who developed CDI revealed increased L-arginine conversion to GABA and an elevated concentration of stool GABA. Similarly in a mouse model of CDI, daily administration GABA-producing (hCRIB) bacteria prior to CDI, elevated stool GABA and mice experienced significant weight loss (Dann et al., 2015). Furthermore, Zolpidem (a GABAA- receptor α1 subunit agonist) use in patients was significantly associated with CDI onset and increased the risk of CDI by almost 5 fold, while in a mouse model of CDI, daily administration of zolpidem increased C. diff- induced inflammation and mucosal damage (Dann et al., 2015) (Figure 1.7).

Bacteria found within the human and animal microbiome are known to a be a source of luminal GABA (Barret et al., 2012; Dann et al., 2015) with Bifidobacterium and Lactobacillus species reported as being some of these GABA-producers (Barret et al., 2012; Pokusaeva et al., 2017). Luminal GABAergic signalling appears to present a risk factor in CDI pathogenesis; consequently an emphasis on the role of luminal GABA as a mechanism underlying disease recurrence could constitute a new approach in disease management and therapy, but little is known about how luminal GABA signals affect the ENS; Chapter 4 explores this question. In order to gain an understanding of GABAergic transmission via the gut-microbiota axis, the physiological role of GABA in the ENS is discussed the next section.

Figure 1. 6 The microbiome and disease susceptibility in CDI. A CDI patient-metabolite network. CDI is associated with compositional and functional changes in the microbiome, where global metabolomics show that a distinct fecal metabolome is associated with healthy (green squares) vs. CDI patients (red squares). Fecal microbiota transplantation (FMT) restores a healthy metabolome within weeks of treatment (see Pre-FMT vs Post-FMT). Analysis of the metabolomics and subject data network shows that stool metabolites distinctly associated with recurrent CDI are lost after successful FMT illustrating a linkage of functional pathways with clinical outcomes. Figure taken from Dann et al. (2015).

Figure 1. 7 Zolpidem promotes infection and disease severity in CDI. The bacterial metabolite GABA is associated with increased risk of infection- in human patients prescribed Zolpidem (GABAA-receptor α1 subunit agonist) the risk of developing CDI increases by almost five-fold (A). In a mouse model of CDI, zolpidem increases C. difficile-induced inflammation and mucosal damage. Representative caeca and their histological staining are shown (B). Figure adapted from Dann et al. (2015).

8. THE PHYSIOLOGICAL ROLE OF GABA IN THE ENS

GABA is a major inhibitory neurotransmitter in the CNS, but the role of GABAergic transmission in the ENS is much less well defined due to its multifactorial functions, multiple receptor subtypes and species and regional differences. Chapter 5 of this thesis aims to elucidate mechanisms of GABAergic transmission in the MP. The distribution of endogenous GABA throughout the gut in higher concentrations than those present in the brain, across a number of species including mouse, rat and human certainly suggests a role in gastrointestinal function (Krantis, 2000; Tedeschi et al., 2003). GABAergic neurons, fibres and receptors are expressed in the ENS of rats and mice (Krantis, 2000; Fletcher et al., 2001; Fletcher et al., 2002; Seifi et al., 2014) with activation of receptors modifying a number of gut functions (Auteri et al., 2014, 2015). GABA is also localised to enteroendocrine cells in the mucosa (Davenger et al., 1994; Krantis, 2000) and thus is proposed to be a neurotransmitter and endocrine mediator in the ENS. In addition, GABA is known to modulate the function of immune cells. Immune cells are known to express GABA receptors, transporters and enzymes (Wu et al., 2017) while GABA has been shown to act on antigen presenting cells to inhibit inflammatory responses (Bhat et al., 2010) and to inhibit T-cells responses via the GABAA receptor (Tian et al., 1999).

GABA in the ENS is produced via the same synthesising enzyme found in the CNS; L- glutamate decarboxylase (GAD), which utilises glutamate as a substrate. The highest L- glutamate decarboxylase activity is in the myenteric plexus (Jessen et al., 1979; Erdo and Bowery, 1986; Krantis, 2000); accordingly, GABAergic neurons are predominantly found within the MP where they are thought to be interneurons and motor neurons.

In rats and humans between 5 and 8% of myenteric neurons in the colon are GABAergic, where they colocalise with the inhibitory neurotransmitter SOM and to a lesser extent with NOS and enkephalin (ENK) (Nichols et al., 1995; Krantis, 2000; Hyland and Cryan, 2010). GABAergic neurons are found at a slightly higher proportion in the mouse colon (14%) and are immunoreactive for NOS, calretinin or SP (Sang and Young, 1996). Most GABAergic interneurons and motor neurons in the murine colon project anally, perhaps mediating descending reflexes, and often for greater distances than VIP or NOS circular motor neurons. Some GABA motor neurons to the circular muscle project orally and may possess an excitatory function (Sang et al., 1997). In the small intestine however, GABAergic neurons 44

are sparse; less than 5% of neurons in the MP of the mouse small intestine are GABAergic, and hence they are often excluded from colocalisation studies and have not been studied in detail in this region (Sang and Young, 1996; Sang et al., 1997; Li et al., 2011).

Across most species including humans, GABA-immunoreactive nerve fibres can be found within the circular muscle (Jessen et al., 1986; Sang and Young, 1996; Sang et al., 1997; Krantis, 2000). In the mouse colon, GABA terminals that are also immunoreactive for either NOS, SP or calretinin are present in the circular muscle, and hence GABA is likely to be present in both excitatory and inhibitory motor neurons (Sang and Young, 1996). In the human colon, GABA colocalises with NOS in nerve fibres innervating the muscle (Nichols et al., 1995), but the exact role of GABA in neuro-muscular transmission in the gut remains largely unknown. It is recognized though, that GABA receptor-mediated motility effects require intact enteric circuitry, as rat colonic smooth muscle cells in isolation do not respond to GABA (Grider and Makhlouf, 1992). In the rat gut, GABA-immunoreactive fibres project to the submucosa and mucosa and are abundant in the colon particularly within myenteric ganglia and interconnecting trunks, often surrounding cell bodies within a ganglion, some of which are GABA-immunoreactive themselves (Jessen et al., 1986; Krantis, 2000). In the mouse colon GABA-immunoreactive terminals are present in the submucosal and myenteric plexuses, where myenteric terminals are also immunoreactive for calretinin, and thus it is highly probable that they are cholinergic since in the mouse calretinin neurons are cholinergic as they are other species (Qu et al., 2008; Furness, 2006). Descriptions of GABAergic innervation in small intestine are comparably less informative (Sang and Young, 1996; Sang et al., 1997).

8.1 GABA receptors and their function

GABA signalling includes transport of GABA across the intestinal epithelium (Mazzoli and Pessione, 2016). In humans, a H+/GABA symporter mediates the uptake of luminal GABA, and has been shown to be found on the apical membrane of intestinal epithelial cells (Thwaites et al., 2000; Chen et al., 2003). In the rodent gut, the intestinal absorption pathway of GABA can occur via amino acid carrier proteins that typically function in nutrient absorption; in the rat small intestine GABA has been shown to share a transporter with the amino acid β-alanine (Nacher et al., 1994). Alternatively, GABA can cross the epithelium to

interact with enteric neural elements, following rupture of the epithelial barrier due to inflammation or invasion of bacteria (Pothoulakis et al., 1998; Sansonetti et al., 1999).

Within the ENS, GABA can act through excitatory ionotropic GABAA and GABAC receptor subtypes which are both pentameric chloride channels (Fletcher et al., 1998; Auteri et al., 2015). Unlike its inhibitory role in the central nervous system, in the ENS GABA has an excitatory effect on enteric neurons via GABAA and GABAC receptors due to their high intracellular Cl- concentration, which is established and maintained by a Na+-K+-2Cl- symporter (Liu et al., 2013). Additionally, GABA can act through inhibitory metabotropic 2+ GABAB receptors, which operate either presynaptically to depress Ca influx via voltage- activated Ca2+ channels to decrease ACh release from cholinergic neurons, or postsynaptically where they are coupled to inwardly-rectifying K+ channels (Hyland and Cryan, 2010; Auteri et al., 2015). Therefore GABA can have both an excitatory and inhibitory influence on neurons via the 3 receptor subtypes.

Most immunohistochemical studies demonstrating neuronal GABA receptor expression have been performed on the rat gastrointestinal tract and in the colon. Immunoreactivity for

GABAA receptors has been localised to myenteric neurons, a third of which are also immunoreactive for NOS (Krantis et al., 1995, 2000). GABAB receptors are localised to muscle fibres, the epithelium and neurons in both plexuses where 50% of GABAB- immunoreactive neurons in the MP also display NADPH-diaphorase activity (Nakajima et al., 1996) thus, the pattern of localisation of GABAA and GABAB receptors in the rat suggests that they may modulate inhibitory NO transmission (Hyland and Cryan, 2010). Very little data exist on GABAC receptor function and expression, but it has been demonstrated that GABAC receptors are expressed on most calretinin immunoreactive neurons and on over half of NOS neurons in the rat colon (Fletcher et al., 2001). In the MP of the murine colon, a number of GABAA receptor subunits are expressed on the somata of 5HT-, SOM-, ChAT- and NOS- immunoreactive neurons and on axonal compartments of NOS neurons in SMP

(Seifi et al., 2014). This localisation pattern in the murine MP suggests that GABAA receptors regulate neuronal activity post-synaptic to the GABA release sites, but are expressed presynaptically on SMP neurons, which could result in an autoregulatory function- further controlling the release of co-expressed neurotransmitters (Seifi et al., 2014). Additionally,

GABAB receptors are known to be found on enteric neurons in the mouse colon (Casanova et al., 2009), and only limited studies reporting GABAC receptor expression in mouse model 46

exist. One study has demonstrated that contraction of longitudinal muscle of the mouse small intestine is modulated by GABAC receptors (Zizzo et al., 2007). Surprisingly little additional literature exists on GABA receptor expression in the small intestine, thus GABA receptor subunit expression on neurochemically defined cell types has been limited to few animal models and intestinal regions and it remains to be fully investigated.

Most of what is known about GABA receptor expression in the GIT comes from pharmacological studies in which the contribution of GABA receptors to gut function is often studied in the MP, where it appears that these receptors exist in both excitatory and inhibitory reflex pathways. Ascending excitatory pathways result in muscular contractions mediated by ACh and tachykinins, while descending inhibitory reflexes lead to muscle relaxation via release of transmitters such as NO, VIP and ATP (Grider, 2003; Auteri et al., 2015).

In rodent models, activation of the GABAA and GABAC receptor systems typically leads to the activation of cholinergic and noncholinergic enteric neurons and subsequent release of ACh or NO resulting in either contraction or relaxation in gut muscle responses (Tonini et al.,

1987; Krantis and Harding, 1987; Zizzo et al., 2007; Auteri et al., 2014). GABAB receptors appear to exist in pathways resulting in reduction in presynaptic ACh release from neurons and a reduction in contractile activity (Sanger et al., 2002; Auteri et al., 2014), or in reduction of inhibitory NANC transmitter release (Tonini et al., 1989a). Inhibition by GABAB receptors has also been reported in human gut (Gentilini et al., 1992). Figure 1.8 illustrates the contribution of GABA receptors on motility output in the rodent small intestine.

It is however, difficult to predict the overall functional output of GABA receptor activation as this varies according to the species studied. In the mouse colon, activation of GABAA receptors increases propulsive activity (Auteri et al., 2014) however in guinea pig colon, antagonism of GABAA receptors produces the same effect (Frigo et al., 1987).

The net functional effect of GABA also appears to depend on the region specific localisation of receptors along the GIT; GABAA receptors modulate VIP and NO –induced relaxation in the rat stomach (Krantis et al., 1998), but in the jejunum and ileum they also mediate cholinergic contractions (Krantis and Harding, 1987).

Figure 1. 8 GABA receptor activation and signalling underlying motor output in the rodent small intestine. A schematic representing ENS signalling underlying GABA receptor-mediated motility changes. GABAA and GABAC receptor systems usually exist in pathways resulting in the activation of cholinergic and noncholinergic enteric neurons with release of ACh or NO respectively, resulting in either contraction or relaxation in gut muscle responses. GABAB receptors appear to exist in pathways resulting in reduction in presynaptic ACh release from neurons and a reduction in contractile activity. The overall functional output of GABA receptor activation however is difficult to predict as it is species- and region-dependent. Figure taken from Auteri et al., 2015.

8.2 GABA in secretion and as an endocrine mediator

Aside from motor output mainly under control of the MP, the presence of GABA in submucosal neurons and in mucosal fibres (Jessen et al., 1979) suggests that GABA influences other intestinal functions such as secretion. In fact, GABA has been shown to evoke gastrin and mucous secretion from the mucosal epithelium of the rat stomach (Guo et al., 1989; Erdö et al., 1989), while GABAA antagonism decreases short-circuit current, an indicator of fluid and electrolyte transport across the epithelium, in the guinea pig and rat intestine (Hardcastle et al., 1991; MacNaughton et al., 1996). The GABA-mediated response

in the guinea pig is partially reduced by atropine, implicating the activation of cholinergic secretomotor neurons.

It is probable that GABA also functions as an endocrine mediator in addition to its role as an enteric neurotransmitter since GABA is expressed in mucosal enteroendocrine cells in the rat stomach and intestine (Davenger et al., 1994; Krantis et al., 1994). GABA has been shown to modulate 5-HT release from EC cells in the guinea pig via GABAA and GABAB receptors

(Schwörer et al., 1989) while GABAB receptors colocalise with 5-HT in mucosal cells of the rat stomach and intestine (Nakajima et al., 1996). It is therefore likely that the various GABA receptor subtypes modulate the release of 5-HT from the mucosa. In this manner, since 5-HT is known to trigger peristaltic activity (Grider et al., 1996) and activate the afferents of secretomotor pathways (Kirchgessner et al., 1992; Bertrand et al., 2000), GABA may indirectly activate reflex pathways via epithelial intermediates.

9. AIMS OF THESIS

There are three key aims in this thesis:

The first is to determine how the enterotoxins CT and TcdA and the associated microbial- derived metabolite GABA acting at the level of the gut mucosa, activate the underlying enteric circuitry in vitro. ISNs in the MP project the mucosa, to other myenteric neurons including the formation of networks of ISNs and to the SMP. They also receive input from neurons in submucosal ganglia. Their extensive connectivity implies they are likely involved in both motility and secretion. Since there is a lack ISNs in the mouse SMP, the myenteric plexus has formed an integral focus of this thesis and is extensively examined in the majority of my studies. Accordingly, Chapter 3 and Chapter 4 addressed whether myenteric ISNs are key elements in the pathways activated by enterotoxins and GABA respectively using in vitro incubation methods and intracellular recording in the guinea pig.

The second aim is to further define the role of endogenous GABA in the ENS as a means to elucidate the mechanisms through which microbial metabolites act and contribute to disease. In Chapter 5, Wnt1-Cre;R26R-GCaMP3 mice in which enteric neurons and glia express a fluorescent calcium indicator, were used to examine the actions of GABA and its receptor subtypes on the circuitry of the MP.

The third aim is to further examine the effects of enterotoxins on the enteric circuitry. Chapter 6 and Chapter 7 extended this investigation with in vivo ileal loop incubations of CT and TcdA respectively in Wnt1-Cre;R26R-GCaMP3 mice.

CHAPTER 2: MATERIALS AND METHODS

In Chapters 3, 4 and 6, intracellular recording was the primary method used, while in Chapters 5 and 7 Ca2+ imaging was used extensively. To avoid repetition, the fundamental aspects of these methods are outlined in this chapter, together with the other techniques that require a detailed explanation. Experimental details unique to each chapter will be explained in their respective chapters.

1. ELECTROPHYSIOLOGY

1.1 Tissue preparation for intracellular recording

In Chapters 3 and 4 experiments were performed using guinea pigs (weighing 170-350 g) of either sex. In Chapter 6 male mice on a C57BL/6 background, including Wnt1-Cre;R26R- GCaMP3 mice that express the fluorescent calcium indicator GCaMP3 in all enteric neurons and glia, aged 8-12 weeks were used for experimentation. Guinea pigs were killed by being stunned by a blow to the head and then having their carotid arteries and spinal cords severed, in accordance with the guidelines of the University of Melbourne Animal Experimentation Ethics Committee. The abdominal cavity was cut open and 5-10 cm segments of jejunum were removed immediately distal to the duodenal-jejunal junction. The tissue was flushed clean and placed in physiological saline (composition in mM per litre: NaCl, 118; KCl, 4.6; CaCl₂, 2.5; MgSO₄, 1.2; NaH₂PO₄, 1; NaHCO₃ 25; d-glucose, 11), bubbled with carbogen gas (95% O₂, 5% CO₂). In Chapters 3 and 4, segments of guinea pig jejunum were tied off at both ends and either physiological saline (0.3-0.5 mL), CT (with or without an antagonist) in saline (12.5 μg/mL), TcdA in saline (12.5 μg/mL), or GABA (with or without an antagonist) (100 µM) in saline, was injected into the lumen of the segments which were then incubated at 35°C for 90 minutes in a heated water bath. Mice that had undergone ileal loop surgery in Chapter 6 (see section 2 below) were killed by cervical dislocation, as approved by the University of Melbourne University Animal Experimentation Ethics Committee and distal ileum was dissected out of the abdominal cavity and placed in physiological saline.

Following the incubation period, the lumen of the intestinal segment (either jejunum or ileum) was flushed clear of its contents with physiological saline and placed in fresh 51

physiological saline also containing nicardipine (1.25 μM) and hyoscine (1 μM) to relax the muscularis externa and minimise the contractions of any circular muscle that remained after the dissection, which could disturb the recording process. The tissue segment was cut open along the mesenteric border and pinned flat with the mucosal side up in a dissecting dish lined with silicone elastomer (Sylgard 184, Dow Corning, North Ryde, NSW, Australia). Tissues were dissected with fine forceps under a dissecting microscope to produce three different electrophysiological preparations (Figure 2.1). (1) The mucosa and its underlying connections were left intact in one half of the preparation while the myenteric plexus and underlying longitudinal muscle were exposed in the other half (mucosa-LMMP) (Chapter 3) (Kunze et al., 1995; Bertrand et al., 1998; Bertrand et al., 2000; Gwynne and Bornstein, 2009). (2) In Mucosa-SMP preparations, the mucosa and submucosa were left intact in one half, while the muscle layers and the MP were dissected away. In the other half, the mucosa was removed, revealing the SMP (Chapter 3) (Gwynne et al., 2009). (3) In LMMP preparations, the mucosa and submucosa were removed (Chapters 3, 4 and 6). The preparation was transferred and pinned flat with the mucosal side up into a recording bath (a plastic mould with a volume of approximately 1-2 mL, lined with Sylgard) using 50 μm tungsten pins. The preparation was continuously superfused with warmed (35-36°C) physiological saline bubbled with 95% O₂, 5% CO₂, at a flow rate of 4-6 mL/min. The preparation was then allowed to equilibrate for an hour under these conditions before commencing electrophysiological experiments.

Figure 2. 1 Tissue preparation for intracellular recording. A diagram representing the three experimental preparations: Mucosa-LMMP, Mucosa-SMP and LMMP. A neuron in either the SMP or MP is impaled with the recording electrode, while the stimulating electrode is placed on the mucosa opposite its fibre tract or on an interganglionic fibre tract leading into the ganglion as to provide focal stimulation to the neuron in that ganglion (Gwynne et al., 2009). Neurons were recorded close to the mucosal wall (first row), further away from the wall or from preparations cleared of mucosa completely. Note the orientation of the tissue in the recording bath, in order to avoid the diffusion of substances being released off the mucosal half and into the recording area.

1.2 Impaling and recording from neurons

Ganglia were visualised at 200x magnification under an upright compound microscope (BX51W1, Olympus, Australia (NSW)), using a long distance objective (LMPlanF1 20x/0.40 ∞/o, Olympus, Australia (NSW)).

Neurons were impaled with intracellular glass microelectrodes (95-200 MΩ tip resistance) made by pulling glass capillaries (Harvard GC100F-15, ID 0.58 mm x OD 1.0 mm) on a microelectrode puller (P-87, Sutter Instrument CO, U.S.A.). They were filled with an electrolyte solution consisting of 1 mol/L KCl with or without 2% biocytin (Sigma Aldrich, Castle Hill, NSW, Australia) to allow neurons to be identified morphologically after experiments. A 250μm silver wire was inserted into the glass electrode just before the tip and submerged in the KCl solution in order to record voltage changes in the cell and apply current to the cell. A neuron contained within a chosen ganglion was impaled with the recording electrode held in place by a micro-manipulator (Leitz, Germany). The recording electrode was lowered into the cell using the manipulator whilst simultaneously vibrating the tip of the electrode or “buzzing”. The recording electrode was connected to an amplifier (Axoprobe- 1A, Axon Instruments, U.S.A) whose voltage output was digitized at 10 kHz by a data acquisition system (Digidata 1440A, Molecular Devices, California) and displayed as two trace records (voltage and current traces) on a personal computer monitor using the software Axotape 2.0.2 (Axon Instruments U.S.A.). The recording circuit was grounded by 2 ground electrodes ( 250 μm silver wire) positioned in the organ bath with the preparation, while the upright compound microscope and both of the manipulators were situated on an anti- vibration table (Technical Manufacturing Cooperation, U.S.A.) enclosed in a grounded metal Faraday cage.

Electrical stimulation

A stimulating electrode (made from 50 μm tungsten wire) placed in a micromanipulator (Model: MP-1, Narishige, Japan) and two ground electrodes (250 μm silver wire), were connected to a Master-9 stimulator (A.M.P.I, Israel) via a stimulation isolation unit (Iso-flex, A.M.P.I, Israel). The stimulating electrode was positioned on top of a fibre tract of a chosen ganglion, or on an interganglionic fibre tract leading into it, to stimulate the axons of, and evoke post-synaptic potentials in, the impaled cells.

Electrophysiology protocol

The electrophysiological properties of the two major electrophysiological classes of neurons in the ENS (AH and S neurons) (Hirst et al., 1974) were determined in a number of ways. The excitability of neurons was established by injecting depolarising current pulse steps, 500 ms in duration and in 50 pA increments over the range of 50-350 pA, through the recording electrode. The membrane potential of the neurons was held at -50mV and was allowed to return to this value in between each change in current pulse amplitude. The number of action potentials (APs) triggered by each pulse was counted and the duration of AP firing during each pulse was measured. Duration of firing was measured from the start of the first AP to the end of the last AP triggered by each pulse. When a single AP was observed, the duration was taken to be the time between the start and end of the AP. Neurons that failed to fire were treated as having a duration of 0 ms. Instantaneous firing frequency was measured as the interspike interval (ISI) between the first two spikes at the 300 pA current step. Hyperpolarising current pulses (500 ms duration, 50 pA increments over 50-350 pA) were also injected through the electrode to determine any changes to the input resistance (IR, MΩ) of the neurons. The input resistance was determined by measuring the change in membrane potential at each pulse amplitude and was then calculated using Ohm’s law (V = IR; voltage

= current x resistance). Ih currents can be detected as a characteristic ‘sag’ present during hyperpolarising steps. The magnitude of the Ih-induced rectification was measured as the difference between the maximum hyperpolarisation and the hyperpolarisation at the end of each current step. The gradient of this voltage current relationship was then determined to calculate the change in IR due to the conductance change underlying the Ih current, in accordance with Ohm’s law.

A focal stimulating electrode was positioned on top to the mucosa or on interganglionic fibre tracts leading into the ganglion of choice. Single stimuli were employed to evoke antidromic APs and fast excitatory post synaptic potentials (fEPSPs). The amplitude (change in mV, from baseline to peak amplitude) and duration (D50) (time taken to return to half amplitude from peak amplitude) of the APs were measured as well as the amplitude and full duration (time taken to return to baseline from peak amplitude) of fEPSPs. In AH neurons, the amplitude and duration (D50 and full, time taken to return to half amplitude from peak amplitude, and the time from the beginning of the AP to when baseline is returned, respectively) of AHPs were measured. In addition, the amplitudes of after-depolarising

potentials (ADP) were measured from the base of the ADP to its peak amplitude. The amplitude and full duration of fEPSPs was measured from single pulses evoked at -90mV, where the threshold stimulus pulse was increased to supramaximal stimulus to ensure changes in fEPSP amplitude were not due to variations on the number of presynaptic fibres activated. Trains of stimuli (10 pulses, 20 Hz) were applied at resting membrane potential

(RMP) to evoke slow excitatory postsynaptic potentials (sEPSPs). The amplitude and D50 of sEPSPs were measured. The RMP of the neurons at the end of electrophysiological experiments was examined and any spontaneous activity throughout the experiment was noted.

1.3 Morphology of biocytin-filled neurons

After experiments, the preparations were fixed overnight in Zamboni’s fixative (2% formaldehyde/0.2% picric acid in 0.1 M phosphate buffered saline, PBS, pH 7.0) (Chapter 3) or in 4% formaldehyde (in 0.1 M phosphate buffer, pH7.2) at 4°C (Chapter 6). Zamboni’s fixative was cleared with 3 x 10 min washes in dimethylsulfoxide (DMSO), followed by 3 x 10 min washes in phosphate buffered saline (PBS) and the mucosa was removed if present. Formaldehyde was cleared with 3 x 10 min PBS washes.

The preparations were placed in a humidified dark container to incubate in 1% Triton X-100 (ProSciTech, Thuringowa, QLD, Australia) for 30 mins. After another 3 x 10 min washes with PBS, the preparations were incubated in Steptavidin Alexa Fluor 594 (1:200; Molecular Probes, Mulgrave, Vic, Australia) for 2.5 hours at room temperature. The label was then washed off from the preparations with 3 x 10 minute PBS washes. The preparations were mounted onto glass microscope slides in DAKO mounting medium (Dako, Carpinteria, CA). The impaled neurons were visualised under fluorescence using 20x, and 40x objectives on a Zeiss Pascal Confocal microscope where images were taken as a z-series at a resolution of 512 x 512 pixels or on a Zeiss Axio Imager M2 microscope. The contrast and brightness of the images were adjusted using LSM image browser software (version 4.2.0.121) or FIJI software (ImageJ ® 1.51a).

1.4 Analysis and statistics

The recorded data were analysed using AxoScope computer software (version 10.2.0.14, Axon Instruments U.S.A.). Injection of depolarising and hyperpolarising current pulses into the neurons was performed 2 times per cell while the single and train stimulus regime was repeated 3 times to obtain averages for these parameters. The results are presented as mean ± SEM. Statistical comparisons included analysis of the firing over the whole range of current pulses using two way ANOVA, where statistical values are given for the entire curve and 2- sample t tests for independent groups with P < 0.05 considered statistically significant.

2. ILEAL LOOP SURGERY

2.1 Surgical preparation

In Chapters 6 and 7 experiments were performed using male C57BL/6 or Wnt1-Cre;R26R- GCaMP3 mice aged 8-12 weeks. Mice undergoing the ileal loop procedure were fasted overnight (up to 16 hours) but allowed ad libitum access to water to allow the emptying of intestinal contents. On the day of the procedure, the animal was collected from the in-house animal facility.

Within 10-15 mins of collection, the animal was placed in a closed, recirculating perspex container and anaesthesia was induced by inhalation of isoflurane (Ceva, Glenorie NSW, Australia) at a concentration of 4-5% via a Fluovac system (Harvard Apparatus Holliston MA, US). The animal was briefly removed from the container, still under anaesthesia, to be shaved along the abdomen from the ribcage to the top of the thigh. The animal was then transferred to a nose cone on a heated mat (T6-100 temperature controller, CWE Inc, US) where isoflurane was administered at 2-3% via the Fluovac system, as required to ensure stable breathing and loss of the pedal and tail withdrawal reflexes. Excess isoflurane was scavenged by the same system. Reflexes were monitored every 5-10 mins throughout the procedure.

The shaved area was cleaned with a sterile gauze swab (7.5 cm 8ply BSN Medical, Mt Waverley VIC, Australia) sprayed with 80% ethanol first along the midline and then either

lateral side, ensuring the midline was kept cleanest. The swabbing process was repeated with a betadine antiseptic liquid (Virginia QLD, Australia) swab.

2.2 Surgical procedure

Surgical instruments were placed in a dry heat bead steriliser (GerminatorTM 500, Glass bead Dry Sterilizer, US) for sterilisation. Approximately 100ul of the anaesthetic bupivacaine hydrochloride 0.5% (Astra Zeneca, North Ryde NSW, Australia) in a 1 mL syringe (Terumo, Tokyo Japan) was injected with a 25G (0.50 x 19 mm) needle (Terumo) at several (usually five) points along the midline where the incision would be made (20ul injected at each point). A period of 10 mins was allowed to elapse for the bupivacaine to take effect. A cutaneous abdominal incision ~3 cm in length was made along the linea alba, to minimize bleeding, starting from ~1-1.5 cm above the rectum up to the ribcage. The same incision was then followed along the underlying muscle layer. The animal was draped in a sterile piece of gauze with a small window cut over the area for incision. The gauze was moistened with sterile physiological saline heated to 37°C.

The caecum was located in the abdominal cavity as a landmark and was gently displaced along with the distal ileum without affecting the blood and extrinsic nerve supply. The abdominal contents were regularly kept moist throughout the procedure with the heated saline. The first few, most distal centimetres of ileum, were disregarded for the loop. A 24G (0.8 x 38 mm) needle was pierced through the mesentery at the distal end of the intended loop. Silk suture (DYSILK, U.S.P: 4/0, Met: 1.5, Dynek Pty Ltd South Australia) was fed through and the ileal segment was tied off with a double knot.

4-5cm oral to this point, silk suture was tied around the gut segment as before, but loosely enough to allow a syringe to pass through. Either the physiological saline or the toxin in saline was injected into the ileum filling the region using a 25G needle (approximately 100- 200 µl of incubating solution). The ileum was slightly elevated at the injection point to reduce backflow out of the loop. The silk suture was quickly tightened below the injection site with a double knot. The ileal loop and abdominal organs were gently returned into the abdomen. The incision in the muscle was sutured first, down the midline using a 24G needle. The process was repeated for the cutaneous incision. Surgical glue (Vetbond tissue adhesive,

3M St Paul MN, US) was applied over the suture material to seal and secure the site, so as to not lose any peritoneal fluid.

The animal was placed in a recovery cage on a heated mat with access to bedding and water and monitored every 30 mins during the duration of the incubation. Animals recovered consciousness ~ 15 mins following the procedure, after which they moved around freely in their cages. After 3.5 hours (Chapter 6), or 2.5-3.5 hrs (Chapter 7) animals were killed by cervical dislocation as approved by the Melbourne Animal Experimentation Ethics Committee.

3. CRYOSECTIONING

Segments of ileal loop tissue and off target tissue (jejunum) were fixed in 4% formaldehyde overnight (Chapter 7). The preparations were cleared with 3 x 10 min PBS washes. The preparations were cryo-protected in a 20% sucrose/PBS solution overnight and then dissected into 2-3 mm rings of intestine. They were placed in OCT medium (Tissue Tek, Elkhart IN, USA) and frozen in isopentane cooled by liquid nitrogen then stored at -80°C until they were sectioned. Sections of tissue 20 μm in thickness were cut on a cryostat (Microm HM 525, Fronine Laboratory Supplies, Riverstone, NSW, Australia) and mounted onto positively charged slides (SuperFrostPlus, Menzel-Glaser, Braunschweig, Germany) where they were left to dry for an hour at room temp. Sections were processed with Hematoxylin and Eosin (H&E) staining (Table 2.1).

Table 2. 1 Protocol for H&E staining.

Steps/Method Duration

Wash in running tap water 10 sec

Immerse in Mayer’s Haemotoxylin 60 sec

(Choral Hydrate 100g; Potassium Aluminium Sulphate 100g; Haemotoxylin 4g; Citric Acid 2g; Sodium Iodate 0.4g; Distilled water 2L; Glacial Acetic Acid 20ml)

Wash in running tap water 10 sec

Scott’s tap water 30 sec

(Magnesium Sulphate 40g; Sodium Bicarbonate 7g; Distilled water 2L)

Wash in running tap water 10 sec

Immerse in 0.1% Eosin (with 2g/100ml CaCl2) 30 sec

Wash in running tap water 10 sec

Immerse in 70% Ethanol 2 dips

Immerse in absolute Ethanol 2 dips

Immerse in absolute Ethanol 4 dips

Immerse in absolute Ethanol x2 60 sec each

Histolene 120 sec

Histolene 120sec

Mount in DPX and place a coverslip on Leave to dry overnight

4. CELL CULTURE ASSAY

Vero cells (Sigma-Aldrich, Castle Hill, NSW, AUS), derived from the kidney of an African green monkey (Cercopithecus aethiops) (Ammerman et al., 2008) were cultured to test the cytotoxicity of TcdA prior to use in experimentation.

Vero cell lines are often used in cytotoxicity assays for Clostridial enterotoxins (Mahony et al., 1989; Eastwood et al., 2009) and for other enterotoxins such as Escherichia coli (Konowalchuk et al., 1977). TcdA acts on epithelial cells to induce biochemical changes, such as alterations to the cytoskeleton and tight junction permeability (Hecht et al., 1988); in cell culture this is associated with changes in cell morphology such as rounding and detachment (Mahida et al., 1996; Eastwood et al., 2009).

Vero cells were cultured in Dulbecco’s modification of Eagle medium (DMEM), supplemented with 10% heat inactivated fetal bovine serum (FBS) (both from Life Technologies, Mulgrave, Victoria, AUS) using 100 x 15 mm Falcon ® Petri dishes (In Vitro Technologies, Noble Park, VIC, AUS) at 37 °C in a humidified atmosphere. Confluent monolayers were counted with a Biorad TC20 TM automated cell counter and removed using TrypLE Express (Life Technologies, Mulgrave, Victoria, Australia). Cells were resuspended in fresh DMEM at approximately 105 cells/ml. 1 ml of cell suspension was pipetted into each well of a 12-well plate (In Vitro Technologies, Noble Park, VIC, AUS) and left to cultivate overnight at 37 °C. The following day, toxin activity was assayed by adding 1 µg of TcdA into 1ml of DMEM in each well. As a negative control, only DMEM was added into some wells. Plates were incubated at 37 °C and images of the cells were taken after 2, 4 and 24 hours of incubation using a Zeiss Axio Imager M2 microscope and with an Axiocam 506 mono camera (from Zeiss, Australia) using AxioVisionLE64 (v 4.9.10) software and a 20x objective.

5. IMMUNOHISTOCHEMISTRY

5.1 Tissue preparation

Segments of either ileum (Chapters 5 and 7), jejunum or proximal colon (Chapter 7) were dissected from mice of a C57Bl6 background including Wnt1-Cre;R26R-GCaMP3 mice. Tissue was opened along the mesenteric border and pinned flat on an elastomer-lined dish and fixed for 1 hour 15 min with 4% formaldehyde at 4°C. Preparations were rinsed with 3 x 10 min PBS washes and microdissected to reveal either the SMP (Chapter 7) or the MP (Chapters 5 and 7). The wholemount preparations were first permeabilized with 1% triton X- 100 in PBS (ProSciTech, Thuringowa, QLD, Australia) for 30 min at room temperature and washed in PBS (3 x 10 min). They were then double or triple labelled with a combination of primary antisera for 24-72hrs at 4°C and rinsed with PBS (3 x 10 min). Incubations in secondary antisera were at room temperature for 2 hrs and 15 mins. All preparations were rinsed of excess secondary antisera 3 times, for 10 mins before they were mounted on slides with Dako fluorescent mounting medium (Carpinteria, California, USA). Specific details regarding the immunostaining protocol including antiserum combinations, are described in the individual chapters outlined above (Chapters 5 and 7).

5.2 Imaging and analysis

Wholemounts of myenteric and submucosal preparations were viewed under 20x and 40x objectives respectively on a Zeiss Axio Imager M2 microscope. Images were acquired with an Axiocam 506 mono camera using Zen 2.3 (blue edition) software (all from Zeiss, Australia).

For cell counts and colocalisation studies, proportions of each neuronal subtype were determined by examining co-expression with the pan neuronal marker Hu.

At least 150 Hu+ cell bodies in each myenteric, and 30 Hu+ cell bodies in each submucosal preparation were examined. The mean proportion of each neuronal subtype, was determined by calculating averages from 3 animals.

The data are expressed as mean ± SEM and n = the number of cells examined. Statistical analyses were performed using unpaired t-tests with P < 0.05 considered statistically

significant. Comparisons were performed using using GraphPad Prism 5.0 (GraphPad Softwares, San Diego California).

6. CALCIUM IMAGING

6.1 Tissue preparation for Ca2+ imaging

Wnt1-Cre;R26R-GCaMP3 mice of either sex (Chapter 5), or male Wnt1-Cre;R26R-GCaMP3 mice (chapter 7) aged 8-12 weeks were killed by cervical dislocation, as approved by the University of Melbourne University Animal Experimentation Ethics Committee. A segment of ileum (Chapter 5) or ileal loop (Chapter 7) was dissected out and placed in physiological saline bubbled with 95% O₂, 5% CO₂ in an elastomer-lined dish. The tissue was cut along the mesenteric border and pinned flat. The mucosa, submucosa were removed by microdissection to obtain preparations of SMP (Chapter 7). Then circular muscle was removed to obtain preparations of the underlying myenteric plexus and attached longitudinal muscle (LMMP) (Chapters 5 and 7). The preparations of SMP and MP were immobilized by stretching the tissue (plexus uppermost) over an inox ring which was then clamped by a matched rubber O- ring (Vanden Berghe et al., 2002). A maximum of 5 rings were prepared from each segment of ileum. The tissue was transferred to an organ bath for imaging, where it was constantly superfused (1 ml/min) with 95% O2: 5% CO2 -bubbled physiological saline at room temperature throughout the experiment

6.2 Experimental protocol

The preparations were imaged using a 20× (NA 0.5) water dipping objective on an upright Zeiss Axioscope microscope with a Zeiss AxioCam MRm camera, images (278 × 278) were acquired at 1 Hz. Neurons within ganglia were stimulated either chemically or electrically. Chemical stimulation (Chapter 5) included the exogenous application (spritz) of GABA (1mM) via pressure injection (2 s duration, 9 psi; Intracel Picospritzer III, Parker Hannifin, Hollis, NH, USA) using a micropipette (tip diameter ~20 μm) situated right at the border of the ganglion. Spritz pipettes were made using a micropipette puller (Model 2-9, Narishige Scientific Instrument, Japan, or P-87, Sutter Instrument Co., Novato, CA, USA) using

borosilicate glass capillaries (Harvard GC1500-15, inner diameter (ID) 0.86 mm x outer diameter (OD) 1.5 mm or Harvard GC100F-10 OD 1.0 mm x ID 0.5 mm)

To electrically stimulate the neurons, a focal stimulating electrode (tungsten wire; 50 μm diameter) was placed on an interganglionic fibre tract leading into the ganglion of choice where a single pulse and a train of 300 μs pulses (1 s, 20 Hz, ; Grass SD9 or S88 stimulator, Grass Instruments, Warwick, RI, USA) were elicited. For time-controlled experiments (Chapter 5), ganglia were stimulated first with a single pulse and then a train of 20 pulses 5 minutes later. This stimulation regime was repeated 2 times 10 minutes apart. For non time- controlled experiments (Chapter 7) this stimulation regime was performed once on 3 different ganglia for each plexus (MP and SMP).

After some experiments, tissues were fixed overnight with 4% formaldehyde at 4°C and examined post-hoc with immunohistochemistry to further identify neuron subtypes.

6.3 Analysis

Analyses were performed using custom-written directives in IGOR Pro (WaveMetrics, Lake Oswego, Oregon, USA) (Boesmans et al., 2013). Regions of interest were drawn over a selected area of the cytoplasm for each neuron. The intensity of the intracellular calcium 2+ ([Ca ]i) transient signal for each response was calculated and expressed as the maximum 2+ 2+ increase in [Ca ]i from the baseline signal (ΔFi/F0). [Ca ]i transients were only considered if the intensity of the transient signal was more than 5 times the intrinsic noise.

7. MEASUREMENT OF SHORT-CIRCUIT CURRENT IN VITRO USING USSING CHAMBERS

7.1 Tissue preparation

In Chapter 7 a segment of ileal loop was dissected out of the animal and placed in an elastomer-lined dish, in saline bubbled with 95% O₂, 5% CO₂. The ileal segment was cut along the mesenteric border and pinned flat, so that all intestinal layers remained intact.

Tissue was gently stretched and mounted over an opening (5.5 mm pin circle diameter, 4 mm reservoir opening, CHM8; World Precision Instruments, Inc. (WPI), Sarasota, FL, USA) 64

between two halves of an Ussing chamber. The two chamber halves were held together by custom-made support apparatus (Department of Physiology, University of Melbourne, Victoria, Australia) whereby the mucosal side of the preparation was facing one side of the chamber and the serosal side, the other. The chamber was connected to a superfusate- containing reservoir. An isolated volume (~10 ml) of physiological saline from separate reservoirs was continuously superfusing the mucosal and serosal sides of the preparation. The physiological saline in the reservoirs was supplied with 95% O₂, 5% CO₂ via injection ports, which also provide ‘gas lift’ allowing constant circulation of saline and the reservoirs were water-jacketed so that the circulating saline was maintained at 37°C.

7.2 Electrical measurements

Transepithelial voltage potential (Vt) was measured by a potentiometer (VCC600 Single Channel Voltage-Current Clamp; Physiological Instruments, Inc., World Trade Drive, San Diego, CA, USA) and two voltage-sensing electrodes (SR4 Reference Electrode Calomel Separable; ThermoFisher Scientific, Station Road, Auchtermuchty Fife, Scotland, UK) each bathed in separate 3 M KCl solutions, and connected to each chamber half via salt bridges (polyethylene tubing OD 1.52 mm x ID 0.86 mm; Tyco Electronics, Huntingdale, VIC, Australia, containing 3% agar melted in 3 M KCl solution) at each side of intestinal preparation. Separate electrodes (Ag-AgCl electrodes; WPI) allowing current to pass across the intestinal preparation were connected to each half chamber by larger 3M KCl salt bridges (ID 5 mm; WPI). These were necessary to clamp spontaneous Vt to zero; the current required to do so was the short-circuit current (ISC). The chambers were set up with saline alone in the absence of a preparation and ‘zero-ed’ by offsetting any voltage differences between the two voltage-sensing electrodes and compensating for fluid resistance before the preparation was mounted. Following a 30 min equilibration period, short-circuit current (ISC) was measured throughout the experiment.

7.3 Data analysis

Both data collection and analysis were performed using AcqKnowledge 3.9.0 software (BIOPAC Systems, Inc., SDR Clinical Technology, Middle Cove, NSW, Australia). The maximum change in ISC from baseline (ΔISC) was measured and compared between control and treatment groups. 65

8. DRUGS AND TOXINS USED

Drugs were diluted in distilled water to make stock solutions and then again in saline to working concentration on the day of experimentation. Drugs used were hyoscine (1μM), nicardipine (1.25μM), (from Sigma Aldrich, Castle Hill NSW, Australia), CT (12.5 µg/ml, List Biologicals, Campbell, CA), ), tetrodotoxin (TTX, 1 μM, Alomone Labs, Jerusalem, Israel), granisetron (1 μM) (Smith Kline Beecham, Harlow Essex, UK), TcdA (12.5 µg/ml, a kind gift from Dr Tor Savidge Baylor Medical College, Houston and Charalabos (Harry) Pothoulakis, David Geffen School of Medicine UCLA, CA) neurokinin 1 (NK1) antagonist SR140333 (100nM) and NK3 antagonist SR142801 (100nM) (both kind gifts from Dr Emonds-Alt, Sanofi Recherche, Montpellier, France), GABA (100 µM, 1mM), hexamethonium bromide (both from Sigma Aldrich, Castle Hill, NSW, Australia), bicuculline (10 µM), CGP54626 HCl (1 µM) and TPMPA (100 µM) (all from Tocris Bioscience, Avonmouth Bristol UK), Bupivacaine hydrochloride 0.5% (Astra Zeneca, North Ryde NSW, Australia) and isoflurane (Ceva, Glenorie NSW, Australia).

CHAPTER 3: CHOLERA TOXIN INDUCES SUSTAINED HYPEREXCITABILITY IN MYENTERIC, BUT NOT SUBMUCOSAL, AH NEURONS IN GUINEA PIG JEJUNUM

This Chapter has been published in Frontiers in Physiology (2017); 8: 254. It is reproduced here with minimal alterations.

ABSTRACT

AH neurons, which have been identified as a population of intrinsic sensory neurons, are a source of excitatory input to the enteric secretomotor pathways activated during Cholera toxin (CT)-induced hypersecretion. This chapter therefore examines effects of CT in the intestinal lumen in vitro on myenteric and submucosal AH neurons using electrophysiology. CT pre-treatment increased the excitability of myenteric, but not submucosal AH neurons. CT induced excitability in these neurons regardless of the presence or absence of intact mucosa. Coincubation with tetrodotoxin or SR142801 (NK3 receptor antagonist), but not

SR140333 (NK1 antagonist) or granisetron (5-HT3 receptor antagonist) prevented the increased excitability induced by CT. Increased excitability was associated with a reduction in the characteristic AHP and an increase in the ADP of these neurons. We concluded that CT increases excitability of myenteric, but not submucosal, AH neurons. This is neurally mediated and depends on NK3, but not 5-HT3 receptors. Therefore, CT may act to amplify the secretomotor response to CT via an increase in the activity of the afferent limb of the enteric reflex circuitry.

1. INTRODUCTION

When the bacterium Vibrio cholera invades the gut, its exotoxin, cholera toxin (CT), causes hypersecretion in the small intestine, which can produce severe diarrhoea that quickly leads to dehydration and death if left untreated. The enteric nervous system (ENS) has been implicated in the harmful effects of CT in the small intestine since the 1980s (Cassuto et al., 1981a; Cassuto et al., 1982a). The ENS is a complex nerve circuitry embedded within the

walls of the gastrointestinal tract which regulates vital gut functions, including motility and secretion. It incorporates two distinct ganglionated networks- the myenteric (MP) and submucosal (SMP) plexuses.

In one long-standing model of toxin-induced hypersecretion, CT is postulated to activate persistent release of 5-HT from the mucosa, which then activates secretomotor reflex pathways in the ENS. The components of the neuronal reflex pathways include 5-HT3 receptors and probably several neuronal subtypes from each plexus communicating via nicotinic synapses. The secretomotor efferents release vasoactive intestinal peptide (VIP) which binds to specific receptors on enterocytes, activating an adenylyl cyclase-cAMP pathway to drive water and electrolyte secretion (Lundgren, 2002; Farthing, 2000). However effects of CT are more complex than this. Recent work has focused on the effects of CT exposure on the properties of the enteric neurons in the secretomotor and motility reflex pathways (Kordasti et al., 2006; Gwynne et al., 2009; Fung et al., 2010). We have previously shown that luminal incubation of CT in isolated guinea pig jejunum induces a sustained increase in excitability of submucosal secretomotor neurons, the final neurons of the secretomotor pathways (Gwynne et al., 2009). Thus, one action of CT is to enhance the response of secretomotor neurons to activity in secretomotor pathways, but the question remains as to whether the properties of other elements of these pathways are also altered.

Intrinsic sensory neurons (ISNs) are central to the enteric circuitry and are the initial neurons in the secretomotor pathways. They are also referred to as intrinsic primary afferent neurons (IPANS) (Kirchgessner and Gershon, 1988a) or as AH-type neurons from their distinctive electrophysiological feature of a prolonged after-hyperpolarising potential (AHP) following an action potential. Here, we refer to them as AH neurons. The prolonged AHP is critical in determining the excitability of AH neurons, since it has the capacity to limit firing rate and slow excitatory transmission (Bertrand and Thomas, 2004). Other currents exhibited by AH neurons are also important for the regulation of cell excitability (Galligan et al., 1990; Rugiero et al., 2002a; Nguyen et al., 2005; Chambers et al., 2014). There are also mechanosensitive myenteric neurons in guinea pig small intestine that lack the characteristics of AH neurons (for review see Schemann and Mazzuoli (2010)); in guinea pig colon, similar neurons appear to be S-type interneurons (Spencer and Smith, 2004), Whether they have sensory functions in small intestinal reflex pathways remains to be established.

AH neurons are present in the SMP and MP (Furness, 2000). In the MP, AH neurons project circumferentially and synapse with other AH neurons, as well as virtually all other types of neurons (Pompolo and Furness, 1998; Bornstein et al., 1991b; Kunze et al., 1993) and this may also be the case in the SMP (Song et al., 1992; Evans et al., 1994; Pan and Gershon, 2000; Reed and Vanner, 2001). Thus, AH neurons form interconnected networks (Thomas et al., 2000; Bornstein et al., 2012). Given this ability to form networks, AH neurons are not always strictly ‘afferent’, as they can act as interneurons under some conditions (Wood, 1994; Bertrand et al., 1997; Kunze and Furness, 1999; Thomas and Bornstein, 2003; Chambers et al., 2005).

In this study we used intracellular recording to analyse the effects of pre-treatment with CT in the lumen on the firing of AH submucosal AH neurons and myenteric AH and S (the other major electrophysiological class of neuron in the myenteric plexus) in the guinea pig jejunum. We found that prior acute exposure to CT enhanced the firing of myenteric AH neurons, but not that of myenteric S neurons or submucosal AH neurons. This effect was neurally mediated and depended on activation of NK3 tachykinin receptors, but not 5-HT3 receptors.

2. METHODS

2.1 Tissue preparation and electrophysiology

Following the incubation, mucosa-LMMP, mucosa-SMP and LMMP preparations were dissected and recordings were taken from neurons in ganglia in the ‘first row’ close to the mucosal wall in both mucosa-LMMP and mucosa-SMP preparations. Recordings were also taken up to 1 cm away from the edge of the mucosa in mucosa-LMMP preparations and from any ganglion in LMMP preparations lacking all mucosa. Dissections and the 69

electrophysiological protocol are both described in full detail in Chapter 2. The measurement of the Ih-induced rectification and the amplitudes of after-depolarizing potentials (ADP) are shown in Figure 3.1.

2.2 Immunohistochemistry for biocytin-filled neurons

Some impaled neurons were filled with 2% biocytin (Sigma Aldrich, Castle Hill NSW, Australia) during intracellular recordings and were processed after experimentation to reveal their morphology (Chapter 2 section 1.3). In order to identify neurons as either a subclass of motor or interneuron, after each experiment, when possible, immunoreactivity for nNOS was examined. A primary antibody against nNOS (sheep anti-nNOS, Emson) was applied and incubated for 2-3 nights. Excess primary antibody was washed with PBS. A secondary antibody to visualise nNOS expression (Donkey anti-sheep 488, Molecular Probes) was added together with streptavidin anti-biotin alexa 594 (Molecular Probes) to allow identification of impaled neurons and incubated for 3 hours in dark humid containers. Secondary antibodiess were removed by washing in PBS and the preparations were mounted on glass slides in DAKO mounting medium (Carpinteria, CA, USA). Preparations were visualized using 40x and 63x oil objectives on a BIO-Rad MRC 1024 confocal microscope mounted on a Zeiss fluorescence microscope (Gladesville, NSW, Australia) and images of impaled neurons were taken as a z series using separate filters for green and red at a resolution of 512 x 512 pixels. Images were later merged (Image J ® software) to identify impaled neurons as nNOS positive (+) or negative (-).

2.3 Drugs and toxins

Drugs were diluted in distilled water to make stock solutions and then again in saline to the working concentration on the day of experimentation. Drugs used were hyoscine (1 μM), nicardipine (1.25 μM) (both from Sigma Aldrich), CT (12.5 µg/ml, List Biologicals, Campbell, CA), tetrodotoxin (TTX, 1 μM, Alomone Labs, Jerusalem, Israel), granisetron (1 μM) (Smith Kline Beecham, Harlow Essex, UK), neurokinin 1 (NK1) antagonist SR140333 (100 nM) and NK3 antagonist SR142801 (100 nM) (both kind gifts from Dr Emonds-Alt, Sanofi Recherche, Montpellier, France).

2.4 Analysis and statistics

The electrophysiological data were analysed using AxoScope computer software (version 10.2.0.14, Axon Instruments U.S.A.) as described in Chapter 2, section 1.4.

In the text below, quantitative comparisons are provided for numbers of APs and duration of firing for 300 pA depolarizing pulses. In addition, the number of APs fired and the firing duration have been plotted against the entire current pulse amplitude range to produce stimulus-response curves for each condition. Here statistical comparisons include analysis of the firing over the whole range of current pulses using two way ANOVA, where statistical values are given for the entire curve. Tukey’s post-hoc tests were used to identify differences for individual current pulse amplitudes. Any such differences are specified in the results. Where no specification exists, differences in firing are to be taken as occurring over the entire curve, rather than at individual data points. Statistical analyses were made using 2-sample t tests for independent groups or two way ANOVA for the stimulus-response curves, with P < 0.05 considered statistically significant.

Figure 3. 1 Measurements of Ih-induced rectification and ADP amplitude.

The magnitude of the Ih-induced rectification, was measured as the difference between the maximum hyperpolarisation and the hyperpolarisation at the end of each current step (A). The gradient of this voltage current relationship was measured in the same way as the input resistance of the cell (see Methods) and represents the change in input resistance due to the conductance change underlying the Ih current. The amplitude of an after-depolarising potential (ADP) was measured from the base of the ADP to its peak amplitude (B).

3. RESULTS

In this study, CT incubation produced enhanced secretion in all preparations as previously described (Gwynne et al., 2009). Increased fluid accumulation was observed in jejuna that were injected with CT compared to saline controls after 90 mins.

A total of 60 submucosal AH neurons and 111 myenteric AH neurons was examined electrophysiologically. All neurons displayed phasic firing properties (firing rapidly at the onset of the depolarisation, then accommodating to the stimulus), regardless of treatment. The morphologies of 12 myenteric and 46 submucosal AH neurons were examined after electrophysiological experiments and processed to confirm and correlate their morphology with their electrophysiological properties. As expected all AH neurons had large cell bodies and displayed multiple axonal processes, S neurons typically had smaller cell bodies, short dendrites and a single axon (Bornstein et al., 1984; Bornstein et al., 1991a; Bornstein et al., 1994) (Figures 3.2 D, 3.3 D and 3.6 D).

3.1 CT increases excitability of myenteric AH neurons and the effect is neurally mediated

We have previously shown that most submucosal S neurons exhibit an increase in excitability following CT treatment, but only when recorded close to intact mucosa (Gwynne et al., 2009). Thus, the firing properties of control and CT (12.5 μg/mL)-treated AH neurons in the first row of ganglia, close to the mucosa in mucosa-LMMP preparations were examined.

Myenteric AH neurons in CT-treated preparations (n = 29) were significantly more excitable than control neurons (n = 24). CT-treated neurons fired more action potentials (P < 0.001, at 200-350 pA, Tukey’s post-hoc analysis) and for a longer duration (P < 0.001, at 50-350 pA, Tukey’s post-hoc analysis) than controls (firing at 300 pA control 2.9 ± 0.3 APs over 69 ± 6 ms, CT 4.8 ± 0.5 APs over 120 ± 9 ms (Figure 3.2)), with firing more likely to occur at 50 pA following CT incubation (Control: 0.42 ± 0.2 APs; CT 1.2 ± 0.2 APs, P = < 0.05, 2- sample t test) implying a lowered threshold for firing following CT pre-treatment. No change in the ISI was observed between the two groups (Control: 23.9 ± 1.3 ms; CT: 23.8 ± 1.5 ms, at 300 pA) indicating that their maximum firing frequency was unchanged at about 40 Hz, despite being more excitable.

The CT-induced increase in excitability was neurally mediated. Neurons incubated with CT together with TTX (1µM) (n = 6) fired significantly fewer action potentials (P < 0.001) and had a shorter duration of firing (P < 0.001) than neurons incubated with CT alone (CT + TTX: 2.7 ± 1.1 APs at 300 pA, CT + TTX: 56 ± 20 ms at 300 pA (Figure 3.2)). Firing was returned to a control level following co-incubation of CT with TTX (P > 0.05 between control and CT and TTX-treated preparations) (Figure 3.2).

Figure 3. 2 CT increases the excitability of myenteric AH neurons close to the mucosa. Example electrophysiological recordings of action potential firing evoked by current pulses of 50 pA (top panels) and at 300 pA (bottom panels) from a control (A) and CT (12.5 µg/ml) pre-treated (B) neuron. Graphical representations of the mean number (C) and total duration (C’) of APs fired at each depolarizing current amplitude (50 - 350 pA). Following CT treatment, AH neurons fired significantly more APs (^P < 0.001) and for longer (^ P < 0.001) compared to neurons in control preparations. Firing was reduced following co-incubation of CT with TTX (1 μM), shown in greyscale (^P < 0.001, firing and duration between CT and CT and TTX-treated preparations), returning to a control level (P > 0.05 control vs CT and TTX preparations). A confocal micrograph of a myenteric AH neuron injected with biocytin during recording (D) arrows denote axons.

3.2 Effect of CT on the membrane properties and synaptic potentials of myenteric AH neurons

To further examine this CT-induced excitability increase, a number of membrane properties and synaptic potentials of AH neurons were analysed. The late AHP is critical in determining the excitability of AH neurons, since it has the capacity to limit both firing and slow excitatory transmission (Chambers et al., 2014). The amplitude of the AHP (P < 0.01, n = 17) and its duration (P < 0.05, n = 15) were significantly reduced following CT treatment (amplitude: control 6.5 ± 0.8 mV, CT 3.9 ± 0.5 mV; duration: control: 15.5 ± 2.6 s, CT 8.5 ± 2.0 s) (shaded, Table 3.1). While the exact physiological function of the ADP has not been determined, our computational model predicts that an increase in ADP magnitude can increase the excitability of AH neurons in some ways (Chambers et al., 2014). Pursuant to this, we found a significant increase in ADP amplitude observed in CT-treated neurons compared to neurons in the control group (P <0.05, n = 17; control 8.5 ± 1.1 mV, CT 13 ± 1.3 mV) (shaded, Table 3.1).

CT pre-treatment did not alter the resting membrane potential, AP amplitude, AP duration

(D50), input resistance or the Ih rectification in the myenteric AH neurons (Table 3.1). The amplitude and duration of sEPSPs that were evoked by trains of electrical stimuli were also unchanged by CT treatment. Similarly CT pre-treatment did not produce any changes in the incidence of spontaneous APs or anode break APs (Table 3.1).

Table 3. 1 Membrane properties and synaptic potentials of myenteric AH neurons close to mucosa following CT treatment. The membrane and synaptic properties of neurons in the control group and CT- treated cells are compared.

Parameter Control (n) CT-treated (n)

RMP (mV) -55.4 ± 1 (17) -53.5 ±1.1 (24)

Input Resistance (MΩ) 115.0 ± 6.8 (22) 120.9 ± 6.5 (22)

Ih current (MΩ) 25.9 ± 3.6 (11) 27.2 ± 4.9 (11)

AP amplitude (mV) 63.6 ± 3.0 (14) 66.4 ± 2.5 (17)

APD50 (ms) 1.4 ± 0.1 (14) 1.3 ± 0.1 (17)

*AHP amplitude (mV) 6.45 ± 0.8 (14) 3.9 ± 0.5 (17) (P = 0.009)

8.46 ± 2.0 (15) (P = *AHP full duration (s) 15.5 ± 2.6 (13) 0.036)

13.0 ± 1.3 (17) (P = *ADP amplitude (mV) 8.5 ± 1.1 (14) 0.014)

slow EPSP amplitude (mV) 8.5 ± 1.1 (6) 7.4 ± 1.9 (4)

Spontaneous AP firing % 8 % (12) 8% (13)

Incidence Anode Break APs % 42% (12) 46% (13)

*P < 0.05, n = sample size, RMP = resting membrane potential, IR = input resistance, AP = action potential, (Ih) = hyperpolarisation-activated cation current, D₅₀ = half duration, AHP = after-hyperpolarising potential, ADP= after-depolarising potential, slow EPSP = slow excitatory post-synaptic potential. 2-sample t-test.

3.3 Firing of submucosal AH neurons is unaltered by CT

The firing properties of control and CT (12.5 μg/mL)-treated submucosal AH neurons in the first row of ganglia, next to the mucosa, in mucosa-SMP preparations were examined.

There was no significant difference between the stimulus-response curves for submucosal AH neurons in the CT-treated group (n = 33) and the equivalent curves recorded under control conditions (n =27). The number of APs fired and the duration of firing evoked by the 300 pA current pulse stimulus were unchanged (Control 4.4 ± 0.7 APs; CT 5.1 ± 1.0 APs, Control: 156 ± 25 ms; CT 159 ± 30 ms, P > 0.05 (Figure 3.3)). Furthermore, CT treatment did not affect input resistance, resting membrane potential and the amplitudes and durations of electrically evoked slow excitatory synaptic potentials (sEPSPs) of the recorded neurons (data not shown).

Figure 3. 3 CT did not affect the excitability of submucosal AH neurons close to the mucosa. Example electrophysiological recordings of action potential firing at current pulses of 100 pA (top panels) and at 350 pA (bottom panels) from a control (A) and a CT (12.5 µg/ml) pre- treated (B) neuron. Graphical representations of the mean number (C) and total duration (C’) of APs fired at each depolarizing current amplitude (50 - 350 pA). The number of APs fired and their duration were not affected by CT pre-treatment. A confocal micrograph of an AH neuron in the SMP injected with biocytin during electrophysiological recording (D) arrows denote axons.

3.4 Excitability of myenteric AH neurons is affected by proximity to intact mucosa

We tested whether the presence of, or proximity to, the mucosa affects the excitability of AH neurons in control and CT conditions, as we have reported previously for submucosal secretomotor neurons (Gwynne et al., 2009).

In contrast to submucosal secretomotor neurons (Gwynne et al., 2009), the presence of mucosa had an inhibitory effect on the excitability of myenteric AH neurons. In control conditions, AH neurons from stripped LMMP preparations (n =13) fired more action potentials (P < 0.05) and for a longer duration throughout the pulses (P < 0.001) than AH neurons recorded in the first row of ganglia adjacent to the mucosa in mucosa-LMMP preparations (n = 24, Control 1st row 2.9 ± 0.3 APs; Control stripped 3.3 ± 0.4 APs, at 300 pA; Control 1st row 69 ± 6 ms, Control stripped: 91 ± 15 ms at 300 pA) (Figure 3.4). In contrast, the ISI was greater in stripped preparations than adjacent to the mucosa (adjacent to mucosa 24 ± 1 ms, stripped: 30 ± 3 ms, at 300 pA, P < 0.05) indicating that the maximum instantaneous firing frequency was higher when the mucosa was present (42 Hz versus 33 Hz respectively).

Similarly, in CT-treated preparations, myenteric AH neurons from fully stripped LMMP preparations (n = 5), exhibited more action potentials (P < 0.05) than those close to intact mucosa (n = 29, CT 1st row: 4.8 ± 0.5 APs; CT stripped: 5.9 ± 0.7 APs at 300 pA). However, there was no significant change in the duration of firing (CT 1st row: 120 ± 9 ms; CT stripped: 124 ± 21 ms at 300 pA, P > 0.9) (Figure 3.4) or in the ISI (CT 1st row: 24 ± 2 ms; CT stripped: 19 ± 3 ms, at 300 pA, P > 0.05).

Proximity to the mucosa also affected the excitability of myenteric AH neurons. CT-treated myenteric AH neurons impaled further from the mucosa (n = 6) fired more APs (P < 0.001) and for longer (P < 0.05) than CT-treated neurons in the first row of ganglia (at 300 pA CT 1st row: 4.8 ± 0.5 APs; CT away: 6.5 ± 1.4 APs, CT 1st row: 120 ± 9 ms, CT away: 144 ± 31 ms). No difference in the ISI was observed between the two groups (CT 1st row: 24 ± 2 ms; CT away: 20 ± 1 ms, P > 0.1). Furthermore, CT-treated AH neurons from stripped preparations displayed firing (number of action potentials and duration of firing both P > 0.05) comparable to those from CT-treated preparations located away from the mucosa. Thus all subsequent recordings were taken from preparations stripped of mucosa to further examine the components involved in sustained neuronal excitability.

Figure 3. 4 Presence of the mucosa reduces action potential firing in myenteric AH neurons. Graphical representations of the number of APs fired and the duration of firing during prolonged depolarisations in myenteric AH neurons from the two control (preparations with and without mucosa) (A-A’) and two CT (12.5 µg/ml)-treated (preparations with and without mucosa) (B-B’) groups. In control conditions, neurons recorded from preparations without mucosa showed significantly higher number of APs fired (A) (*P < 0.05) and longer duration of firing (A’) (^ P < 0.001), compared to neurons located close to the mucosa. Following CT treatment, neurons from preparations devoid of mucosa displayed significantly higher number of APs fired (B) (*P < 0.05) compared to neurons located close to the mucosa. The duration of firing (B’) however, remained unaltered.

3.5 CT’s effect on myenteric AH neurons is mediated by NK3, but not NK1 or 5-HT3 receptors

Since slow EPSPs in AH neurons are primarily mediated by tachykinins acting via NK1 and NK3 tachykinin receptors (Bertrand and Galligan, 1995; Alex et al., 2001; Johnson and Bornstein, 2004) and slow EPSPs suppress AHPs (Grafe et al., 1980) and therefore modify excitability of AH neurons, the effects of NK1 and NK3 antagonists on the CT-induced changes in firing properties of myenteric AH neurons were examined. The CT-induced increased firing in AH neurons was prevented by co-incubation with the NK3 receptor antagonist (SR142801, 100 nM). Following luminal co-incubation of CT with SR142801, myenteric AH neurons (n = 6) displayed a significant reduction in the number ( P < 0.001) and duration of action potentials fired (P < 0.001) when compared with pre-incubation with CT alone (CT: 5.9 ± 0.7 APs; CT + NK3 antagonist: 3.2 ± 0.7 APs, CT: 124 ± 21 ms, CT + NK3 antagonist: 70 ± 16 ms, at 300 pA (Figure 3.5)). Addition of the NK1 antagonist (SR140333, 100 nM), however, had no effect on the number of APs fired when compared with pre-incubation with CT alone (CT + NK1 antagonist: 6.5 ± 1.8 APs, at 300 pA, n = 6, P > 0.1). However, duration of action potential firing was increased (P < 0.01, at 300 pA, CT + NK1 antagonist: 172 ± 44 ms) (Figure 3.5).

An established mechanism by which CT induces a massive secretion in the small intestine includes 5-HT release by enterochromaffin (EC) cells (Farthing, 2000) and 5-HT is known to excite AH neurons via 5-HT3 receptors on their mucosal terminals (Bertrand et al., 2000).

Luminal co-incubation of CT with the 5-HT3 receptor antagonist granisetron (1 μM, n = 6), however, did not alter the number and duration of APs fired during the protocol, when compared with pre-incubation with CT alone (at 300 pA, CT + granisetron: 5.7 ± 0.9 APs, CT + granisetron: 119 ± 17 ms, P > 0.5 (Figure 3.5)).

Figure 3. 5 Co-incubation of CT with a NK3 but not NK1 or 5-HT3 receptor antagonist prevents CT-induced hyperexcitability in myenteric AH neurons. Electrophysiological recordings are displayed showing a control neuron (A) and a CT (12.5µg/ml) + NK3 (100 nM) antagonist-treated neuron (B) at 50 pA (A, B) and 300 pA (A’, B’) current pulses. These neurons exhibit comparable action potential firing. Pre-treatment with the CT + NK3 antagonist significantly reduced the number of APs fired (C) (^P < 0.001) during depolarisations and their firing duration (C’) (^ P < 0.001), over a range of current pulse amplitudes (effect shown graphically in greyscale). This effect was not observed following the CT + NK1 antagonist (100 nM) or CT + 5-HT3 antagonist (1 μM) co- incubation (D, D’). Instead, duration of action potential firing was increased with the addition of the NK1 receptor antagonist (# P < 0.01) (D’).

3.6 Myenteric S neurons show no change in firing following CT incubation

Since myenteric AH neurons were hyperexcitable after CT incubation and are known to have synapses on almost all other types of myenteric neuron including interneurons and motor neurons which display S-type electrophysiology (Furness et al., 2004a), we examined the excitability of 12 myenteric S neurons from control preparations and 24 S neurons from CT- treated preparations. No significant effects of CT incubation on excitability, resting membrane potential, input resistance, AP properties or spontaneous AP firing were identified (Figure 3.6, Table 3.2). A diverse range of S neurons exists that can be characterised by their morphology, axonal projections and immunoreactivity for nNOS. While we found no difference in excitability between control and CT-treated S neurons in the MP as a whole, it is likely that we have only sampled a small portion of the different subclasses of S neurons that exist and we cannot rule out effects in theses sub-populations. We also cannot exclude that these sites in the network respond to CT exposure with immediate effects during incubation. The morphology and axonal projections were examined in 11 control S neurons; 2 neurons projected orally, 1 projected circumferentially 8 had axons projecting anally. Immunoreactivity for nNOS was detected in 4 of these neurons (all projected anally and 2 were identified as inhibitory motor neurons with another being a descending interneuron and the 4th being unclassifiable). Axonal projections were examined in 21 (of 24) CT neurons; 4 projected orally, 4 projected circumferentially and 13 anally. Nineteen were classified further as interneurons (N = 10) or motor neurons (N = 9) and 5 (of 14) were NOS immunoreactive.

Figure 3. 6 Effects of CT on the action potential firing properties of myenteric S neurons close to the mucosa. Example electrophysiological recordings of action potential firing at current pulses of 100 pA (top panels) and at 300 pA (bottom panels) from a control (A) and CT (12.5 µg/ml) pre- treated (B) neuron. Following CT treatment, firing number (C) of myenteric S neurons and firing duration (C’) was unchanged compared to neurons in control preparations. A confocal micrograph of a myenteric S neuron injected with biocytin during recording (D) arrow denotes axon.

Table 3. 2 Membrane properties and synaptic potentials of myenteric S neurons in control and CT-treated preparations.

Parameter Control (n) CT-treated (n)

RMP (mV) -36.7 ± 4.6 (9) -37.4 ± 1.5 (17)

Input Resistance (MΩ) 165.9 ± 17.8 (8) 158.5 ± 7.0 (15)

AP amplitude (mV) 52.2 ± 2.1 (11) 55.2 ± 1.9 (20)

APD50 (ms) 0.80 ± 0.03 (11) 0.78 ± 0.04 (20)

Fast EPSP Amp (mV) 31.0 ± 2.4 (11) 26.8 ± 1.5 (19)

N = sample size, RMP = resting membrane potential, AP = action potential, ADP = after- depolarising potential, D₅₀ = half duration, Fast EPSP = fast excitatory post-synaptic potential. 2-sample t-test.

4. DISCUSSION

The ENS has been implicated in CT-induced hypersecretion for over 30 years (Cassuto et al., 1981a; Farthing, 2000; Lundgren, 2002; Burleigh and Banks, 2007), but the specific neurons affected and their roles in the enteric neural circuit have not been fully characterised. We have reported that incubation of the guinea pig jejunum with CT in the lumen in vitro induces a sustained increase in the excitability of submucosal secretomotor neurons (Gwynne et al., 2009). We have now examined the effects of CT on excitability of other functionally distinct subclasses of neurons and identified important changes in the firing properties of myenteric AH neurons. CT pre-treatment increased excitability of myenteric AH neurons, but not submucosal AH neurons or myenteric S neurons. This effect was dependent on neural activity and required activation of NK3 tachykinin receptors during the incubation period. Blocking

NK1 or 5-HT3 receptors during the incubation did not alter the effect of CT. This suggests that CT produces hypersecretion via prolonged increases in the responsiveness of both the afferent and efferent limbs of secretomotor circuits (Figure 3.7).

4.1 CT increases the excitability of myenteric AH neurons, but not submucosal AH neurons or myenteric S neurons

Prior incubation of CT in the lumen of isolated segments of guinea pig jejunum caused myenteric AH neurons to respond to smaller depolarizations and for longer, but the maximum firing frequencies remained constant. The increased excitability of myenteric AH neurons was not accompanied by increased spontaneous firing of these neurons or changes in slow EPSPs. Thus, the increased excitability is due to a change in the intrinsic properties of these AH neurons rather than a change in their inputs.

Submucosal AH neurons were unaffected by prior luminal incubation with CT. The large sample of recordings of submucosal neurons were made in ganglia adjacent to intact mucosa, where increased excitability of secretomotor neurons results from identical CT incubations (Gwynne et al., 2009). However, excitability of myenteric AH neurons was increased both in ganglia adjacent to intact mucosa and in fully stripped preparations. Thus, it is unlikely that the failure to detect a change in excitability was due to either the study being underpowered or the experimental conditions.

Myenteric AH neurons respond to mechanical deformation of the muscle (Kunze et al., 2000), muscle contractile activity (Kunze et al., 1999) and chemical stimuli applied to the mucosa (Kunze et al., 1995; Bertrand et al., 1997; Gwynne and Bornstein, 2007a), but not to mucosal deformation (Bertrand et al., 1997). In contrast, the very limited published work on sensory transduction by submucosal AH neurons indicates that they respond to mucosal deformation (Kirchgessner et al., 1992). Thus, our data suggest that CT preferentially enhances responses of enteric pathways to some sensory modalities, but not others.

Luminal CT excites propulsive intestinal motility (Fung et al., 2010) and triggers hypersecretion in the guinea pig small intestine (Carey and Cooke, 1986; Gwynne et al., 2009). The motility effects were observed when CT was present in the lumen (Fung et al., 2010), over an equivalent incubation period to the present study. Contractile responses of rat jejunum in vivo are enhanced after 120 min of CT incubation and persist for several hours after washout of the toxin. (Kordasti et al., 2006). Thus, the excitability of myenteric interneurons and motor neurons might be affected by incubation with CT. These neurons are typically classed electrophysiologically as S neurons. The number of functional subtypes of S neurons, however, is large and some sub-populations form as little as 1% of the total, although their outputs diverge to contact many myenteric and submucosal neurons (Pompolo and Furness, 1995; Bornstein et al., 2004; Moore and Vanner, 2000). Such neurons did not contribute significantly to our sample, as recordings from at least 1,000 neurons in each condition would be required. Thus, we cannot rule out significant effects on some myenteric S neurons, but our data indicate that most interneurons in pathways regulating secretion or motility and the output neurons of motility pathways are unaffected by CT.

4.2 CT increases the excitability of myenteric AH neurons via a mechanism requiring neural activity

To test whether the increased excitability of myenteric AH neurons seen after CT incubation depended on neural activity during the incubation period, we blocked voltage-dependent sodium channels with TTX while CT was in the lumen. This prevented the CT-induced increased excitability of the AH neurons, as seen with submucosal secretomotor neurons (Gwynne et al., 2009). We have shown that CT does not penetrate the mucosa during this type of incubation, with staining for the CT-B subunit confined to the mucosa and not in the underlying SMP, MP or muscle (Gwynne et al., 2009). Thus, CT does not interact directly

with enteric neurons, strongly implicating indirect activation of enteric neural pathways by luminal CT as the mechanism responsible for the enhanced excitability. It is unlikely that the failure of CT to enhance excitability is due to persistence of TTX beyond the incubation, because impalements were often 5-6 hours after the toxins had been washed out and TTX is reversible in guinea pig myenteric plexus (Hirst et al., 1974).

It is also possible that activation of extrinsic primary afferent neurons with axon collaterals in the myenteric plexus contributes to induction of hyperexcitability in myenteric AH neurons. That such collateral innervation of myenteric plexus exists in guinea pig small intestine is indicated by the findings of Takaki and Nakayama (1990), who showed that stimulation of mesenteric nerves as they enter the ENS evokes sEPSPs in myenteric AH neurons. However, extrinsic primary afferents probably do not have a major role, because long-term extrinsic denervation of the rat small intestine does not prevent CT-induced hypersecretion in vivo (Sjöqvist, 1991; Turvill et al., 2000a). Further, submucosal ganglia in guinea pig small intestine receive inputs from SP-containing extrinsic nerve fibres (Costa et al., 1981), but submucosal AH neurons are unaffected by CT.

4.3 CT probably acts to reduce IK channel activity in AH myenteric neurons

Myenteric AH neurons express a wide variety of different ion channels that contribute to their excitability (Chambers et al., 2014). For example, inflammation-induced increased excitability of these neurons is associated with an increase in an hyperpolarization-activated cyclic nucleotide dependent cation conductance (Ih), (Linden et al., 2003) whose activation reduces the amplitude and duration of the slow AHP in these neurons (Chambers et al.,

2014). In contrast, CT does not cause inflammation or alter Ih, but does reduce both the amplitude and duration of the AHP in myenteric AH neurons (Table 3.1). AHPs in these neurons result from the opening of an intermediate conductance calcium-dependent potassium channel (IK) as a consequence of Ca2+ entry during the action potential (Hirst et al., 1974; Hirst et al., 1985b; Vogalis et al., 2002a; Neylon et al., 2004). Realistic computational modelling shows that reducing IK conductance following an action potential increases numbers of action potentials fired and durations of firing evoked by depolarizing current pulses (Chambers et al., 2014). In the model, this occurs with little change in resting membrane potential, input resistance or the maximum frequency of firing, findings similar to our observations of the effects of CT pre-treatment. The ADP triggered by individual action 88

potentials in AH neurons is enhanced by CT. However, our computational simulations indicate that changes in the ADP do not produce changes in action potential firing that match those seen with CT pre-treatment (Chambers et al., 2014) thus the function of this particular membrane event remains unclear.

The reduction in the AHP produced by CT may be due to an increased drive in a protein kinase pathway(s). IK channel phosphorylation by either protein kinase A (PKA) or protein kinase C (PKC) is a major source of potassium channel regulation in AH neurons (Del Carlo et al., 2003; Vogalis et al., 2003). The open probabilities of these channels are reduced by phosphorylation by PKA, which suppresses the AHP (Vogalis et al., 2003). As the CT effect depends on activation of NK3 tachykinin receptors (see below), which are coupled to phospholipase C, the PKC pathway may contribute to suppression of the AHP following CT treatment. PKD, a downstream target of PKC, is of particular interest. In guinea pig ileum, stimulating the NK3 receptor responsible for slow EPSPs in AH neurons, translocates PKD from the cytosol to the plasma membrane where it is subsequently phosphorylated (Poole et al., 2008). PKD activation may account for the membrane changes underlying CT-induced changes in AH neuron excitability.

4.4 Role of mucosal mediators in CT-induced effects on enteric neurons

The current model of the mechanism underlying CT-induced hypersecretion can be summarised as CTB-subunits binding to GM-1 gangliosides in the apical membrane of enterochromaffin cells (Farthing, 2000; Burleigh and Banks, 2007). This is believed to result in depolarization, increased Ca2+ entry via voltage-dependent Ca2+ channels and release of serotonin from the basolateral surface. The serotonin is thought to activate 5-HT3 receptors on mucosal terminals of myenteric AH neurons (Bertrand et al., 2000), thereby activating secretomotor pathways.

We found that only secretomotor neurons close to intact mucosa exhibit increased excitability in CT pre-treated preparations (Gwynne et al., 2009). However, the enhanced excitability of myenteric AH neurons was seen in in preparations lacking any mucosa and thus could not be due to ongoing activity of CT in the mucosa. These findings highlight the conclusion that the prolonged effects of CT are not simply due to an irreversible action on mucosal epithelial cells, whether enterocytes or enterochromaffin cells, as has been widely postulated. Further,

mucosal mediators may have different effects on the afferent and efferent limbs of the neurogenic pathway, or perhaps the reflex limbs are affected by different mucosal mediators.

Our data show that increased excitability of myenteric AH neurons does not depend on activation of 5-HT3 receptors during CT exposure as shown by insensitivity to granisetron during CT incubation. Several other lines of evidence indicate that, although CT-induced hypersecretion depends on 5-HT3 receptor activation, other mediator/receptor combinations are also activated by CT. For example, luminal CT increases propulsive motor patterns in guinea pig jejunum within minutes of exposure, an effect enhanced by blocking 5-HT3 receptors (Fung et al., 2010). In this case, luminal CT increases activity of 5-HT3 receptors, but the pathway excited inhibits propulsive motor activity. Other mediators released from the mucosa may also have a role in CT-induced effects as EC cells contain several other mediators (e.g. cholecystokinin, secretin, ATP) (Cooke et al., 2003; Engelstoft et al., 2013; Gribble and Reimann, 2016) and presumably the same release processes act on these along with serotonin. Further, it seems reasonable to assume mechanisms operating in enterocytes and EC cells also operate in enteroendocrine cells (Barber et al., 1986; Eklund et al., 1989; Lundgren, 1998; Gribble and Reimann, 2016; Fothergill et al., 2017). Future studies aimed at identifying mediators responsible for activation of this pathway will be important for understanding how CT affects the enteric neural circuitry. Further, release of different mediator types from the mucosa may also explain why myenteric AH neurons close to mucosa fired less than those further away/in the absence of mucosa, if different pathways can be activated by different mediators, with there being a possibility of mediator release differentially activating neurons close to the mucosa.

4.5 Role of tachykinin receptors

Myenteric AH neurons express both NK1 and NK3 tachykinin receptors. However, the effect of CT incubation on myenteric AH neurons depends on NK3, but not NK1, tachykinin receptors (Figure 3.5). This is consistent with earlier studies suggesting that these receptors mediate distinct membrane events (Johnson & Bornstein 2003). NK3 receptors activate the Nav1.9 sodium channel via a PKC pathway (Copel et al., 2009). Computational simulations indicate that increasing Nav1.9 conductance increases both the number of action potentials and the duration of firing of AH neurons, but this is associated with an increased amplitude of the resulting AHP (Chambers et al., 2014). Whether activation of NK1 receptors modulates 90

Nav1.9 in these neurons is unknown. Thus, whether the CT-induced increased excitability relates to the Nav1.9 activation by NK3 receptors requires further investigation.

Unlike myenteric AH neurons, submucosal AH neurons do not express NK3 receptors (Jenkinson et al., 1999), although they do express NK1 receptors (Portbury et al., 1996; Lomax et al., 1998; Harrington et al., 2005). It is tempting to suggest that the absence of NK3 receptors directly acting on the submucosal AH neurons accounts for the failure of CT incubation to enhance their excitability. However, VIP secretomotor neurons also rarely express NK3 receptors (Jenkinson et al., 1999), while the cholinergic NPY-containing and calretinin-containing secretomotor neurons do, and both VIP and NPY, but not calretinin, submucosal neurons are made hyperexcitable by CT incubation via an NK3 and NK1 receptor-dependent process. The nature of the second messenger systems activated by these two different tachykinin receptors in distinct classes of enteric neurons has not been identified. While both subtypes are usually thought to activate phospholipase C, NK1 receptors are known to activate two alternative pathways in other systems (Quartara and Maggi, 1997). These issues require further investigation. Further to this, the exact role of NK3 receptors in intestinal hypersecretion remains unclear. NK3 receptor antagonism was demonstrated to reduce excitability of submucosal secretomotor neurons by Gwynne et al., (2009) in an in vitro CT incubation model that was shown to produce luminal hypersecretion, and so it is possible that NK3 block could reduce the hypersecretion. However this was not tested the study of Gwynne et al., (2009). Indeed NK3 receptors have been implicated in the secretory response to distension of rat small intestine (Larsson et al., 2008). However selective agonists of NK3 have had minimal effects on electrogenic ion transport across the guinea pig small intestine (Reddix and Cooke, 1992).

4.6 New proposed circuit

The data presented here suggest a new model for CT-induced neurogenic hypersecretion. This is illustrated in Figure 3.7, which shows the secretomotor circuit (simplified) and how luminal exposure of CT may produce hyperactivity at several sites. These include the intestinal epithelium itself, a major class of intrinsic sensory neurons, myenteric AH neurons, and the output neurons of the secretomotor pathways with both cholinergic and non- cholinergic (VIP) secretomotor neurons being affected.

We postulate that CT acts to release several mediators from EC (and EE) cells in the mucosa including serotonin which activate underlying neural pathways. Initial exposure to CT activates at least three neural pathways, one of which involves serotonin acting on 5-HT3 receptors to inhibit motility, perhaps via the 5-HT3 dependent local inhibitory reflex pathway described by Gwynne and Bornstein (2007b). This immediate effect may be unrelated to the persistent hypersecretion or prolonged motility increases seen after CT exposure in vivo (Mathias et al., 1976).

The second pathway, highlighted in this present study, probably involves a mediator other than serotonin, does not appear to involve 5-HT3 receptors, and produces prolonged activation of NK3 tachykinin receptors on myenteric AH neurons. This may well be due simply to increased activity of these neurons during the incubation period, as they form recurrent excitatory networks with each other and communicate via NK3 receptor mediated slow EPSPs (Bertrand and Galligan, 1995; Alex et al., 2001; Thomas and Bornstein, 2003; Johnson and Bornstein, 2004). Modelling has shown that increasing the excitability of the AH neuron network by reducing IK channel activity increases the gain in the output of the circuit, thereby amplifying sensory input into enteric neural pathways (Thomas et al., 2000; Thomas and Bornstein, 2003) and potentially leading to irreversible increases in network firing (Chambers et al., 2005). These neurons also have excitatory outputs to, and receive similar inputs from, the submucosal intrinsic sensory neurons, (Galligan et al., 1988; Kirchgessner et al., 1992; Vanner and Macnaughton, 2004; Gwynne and Bornstein, 2007b) thereby adding a further level of amplification to the intrinsic sensory network.

The final pathway involves activation of 5-HT3 receptors, presumably on the mucosal terminals of intrinsic sensory neurons, and leads to long lasting increases in excitability of the secretomotor neurons that innervate secretory enterocytes in the crypt regions. These increases in excitability involve both cholinergic and non-cholinergic secretomotor neurons and depend on proximity to the mucosa (Gwynne et al., 2009). The effect would be to amplify any activity in the secretomotor circuits arising from ongoing sensory stimulation or CT-stimulated release of mucosal mediators like serotonin.

Thus, we postulate that CT induces hypersecretion in the small intestine by indirectly increasing the excitability of both the input and the output arm of the secretomotor reflex pathway. This in turn is amplified by increased excitability of secretomotor neurons leading

to increased secretion and potentially distension of the intestinal segment, thereby further exciting intrinsic sensory neurons. Thus, we suggest that CT produces hypersecretion via a form of neural plasticity operating at several levels within the enteric nervous system.

Figure 3. 7 Schematic of secretomotor pathways of the guinea pig ENS activated by CT. CT = Cholera toxin, SMP = submucosal plexus, CM = circular muscle, MP = myenteric plexus, LM = longitudinal muscle, VIP = vasoactive intestinal peptide, NPY= neuropeptide Y, Ach = acetylcholine, ChAT = choline acetyltransferase, nAChR = nicotinic acetylcholine receptor, SP = substance P, Calb = calbindin, INT= interneuron, MOT = motor neuron, 93

NK1R = neurokinin 1 receptor, NK3R = neurokinin 3 receptor, 5HT = serotonin, 5HTR = serotonin receptor.

This simplified schematic of the secretomotor circuit in the guinea pig small intestine integrates our current data with previous work and presents an updated model for CT-induced hypersecretion. It highlights several sites of sustained excitability following CT exposure. Firstly at the mucosal epithelium where CT produces a large release of 5-HT. In myenteric ISNs (in pink) and in secretomotor neurons in the SMP (in yellow). In this study, other myenteric and submucosal neuron subtypes did not show sustained changes in excitability; we cannot exclude however, acute effects occurring during CT incubation. Taken together the revised model proposes that CT acts at the mucosal epithelium to release a range of mediators from EC and EE cells. This CT exposure activates 5-HT-dependent and independent neural pathways: 1) a path where 5-HT acts on 5-HT3 receptors to inhibit motility 2) a 5-HT - independent prolonged activation of myenteric ISNs (via NK3 receptors), possibly due to another mucosally-derived mediator 3) and a 5-HT dependent pathway, presumably via 5-

HT3 receptors on mucosal terminals of myenteric ISNs which leads to prolonged excitation of secretomotor neurons in the SMP.

4.7 Conclusions and future directions

In light of the present study and previously published data (Gwynne et al., 2009), we propose a schematic diagram of the enteric circuitry involved in hypersecretory and propulsive effects of acute exposure to CT in the guinea pig small intestine (Figure 3.7). Prior exposure to CT leads to sustained excitability of specific groups of neurons in both enteric plexuses with the afferent arm residing in the myenteric plexus and efferent arm in the submucosal plexus. However, we cannot exclude the possibility that some enteric neurons are only activated during the incubation phase. The mechanisms involved in the increased excitability of these neurons and the involvement of substances other than 5-HT from the mucosa needs to be further investigated.

CHAPTER 4: CLOSTRIDIUM DIFFICILE TOXIN A AND GABA SIGNALS CONVERGE TO MAKE MYENTERIC INTRINSIC SENSORY NEURONS HYPEREXCITABLE

ABSTRACT

C. difficile toxins, TcdA and TcdB, induce life threatening diarrhoea in susceptible patients. Antibiotic-associated shifts in stool metabolome profiles linked with increased GABA neurotransmitter production represent a significant new risk factor in such patients. GABA is found in enteric neurons and enteroendocrine cells, and is a major metabolic product of antibiotic-associated microbial dysbiosis. Because little is known about how luminal GABA signals may affect enteric nervous system responses to C. difficile toxins, I investigated this question by first examining the effects of GABA and TcdA in the intestinal lumen separately. In Chapter 3 I showed that myenteric AH neurons are made hyperexcitable by the enterotoxin CT, and so the firing properties of myenteric AH neurons were examined via electrophysiology flowing in vitro incubations of GABA and TcdA in guinea pig jejunum. TcdA and GABA both induced hyperexcitability in myenteric AH neurons. Co-incubation of

GABA with bicuculline (GABAA antagonist) further enhanced the increased excitability produced by luminal GABA. In fact, bicuculline alone produced hyperexcitability in these neurons. Addition of either CGP54626 (GABAB antagonist) or TPMPA (GABAC antagonist) to the luminal incubation solution inhibited the GABA induced excitability. Thus, luminal

TcdA, GABA and the GABAA antagonist bicuculline at the level of the mucosa make myenteric intrinsic sensory neurons hyperexcitable suggesting convergence of neural pathways activated by C. difficile toxins and bacterially-produced GABA.

1. INTRODUCTION

As set out in the introduction (Section 6), Clostridium difficile (C. diff) infection (CDI) is increasingly becoming a healthcare concern with TcdA being the major pathogenic agent in rodents. In light of the results of Chapter 3 (Koussoulas et al., 2017) which highlighted a key role of the myenteric plexus and sensory neurons in CT-induced hypersecretion, I examined the effects of luminal TcdA on the same population of myenteric intrinsic sensory neurons 95

affected by CT, to investigate any generality in the action of enterotoxins. Little is known about the participation of enteric neurons in TcdA-mediated hypersecretion, but an electrophysiological study of submucosal enteric neurons of the guinea-pig small intestine (Xia et al., 2000) demonstrated that TcdA excites secretomotor neurons when directly applied to the cell body.

Further, data from both human and animal models suggest that specific bacterial metabolites, notably γ-aminobutyric acid (GABA) originating from the unique ecosystem of flora and fauna in the gut- the microbiome, may be associated with susceptibility to and recurrence of CDI (Barrett et al., 2012; Dann et al., 2015). An emphasis on the role of luminal GABA as a mechanism underlying disease recurrence is a newly emerging research focal point and was investigated in this study by examining the effects of luminal GABA on the same sensory system.

Via intracellular recording, it was found that myenteric AH neurons are made hyperexcitable by both luminal TcdA and GABA and hence are a common point through which enterotoxins and GABA operate. The GABA-induced effects were inhibited by GABAB and GABAC receptor antagonists, but enhanced by a GABAA antagonist, indicating involvement of at least two distinct GABA- activated pathways.

2. METHODS

2.1 Tissue preparation and electrophysiology

Experiments were performed using guinea pigs (weighing 170-350 g) of either sex. The animals were killed as outlined in Chapter 2, section 1.1. The abdominal cavity was cut open and 5-10 cm segments of jejunum were removed immediately distal to the duodenal-jejunal junction. The tissue was flushed clean and placed in physiological saline. The lumens of jejunal segments were incubated with as outlined in Chapter 2, section 1 with physiological saline (control), TcdA (12.5 μg/mL) in saline, GABA (100 µM) in saline or GABA with an antagonist. The concentration of 12.5 µg/mL for TcdA to be incubated was selected based on effective doses which result in demonstrable enterotoxicity in mouse and rabbit in vivo ileal loop assays (Kelly et al., 1994; Castagliuolo et al., 1994). Similarly concentrations of GABA

and GABA receptor antagonists used were based on previous publications (Tonini et al., 1989a; Sanger et al., 2002; Zizzo et al., 2007).

Following the incubation LMMP preparations were dissected and intracellular recording techniques were implemented. Dissections and the electrophysiological protocol are both described in full detail in Chapter 2 sections 1.1 and 1.2.

2.2 Drugs and Toxins

Drugs kept at -20 °C prior to use, were diluted in distilled water to make stock solutions and then again in saline to the working concentration on the day of experimentation. Drugs used were the muscarinic antagonist hyoscine hydrobromide (1 μmol/L) (Keast et al., 1985b) the calcium channel blocker nicardipine (1.25 μmol/L) (Hao et al., 2011) (both from Sigma Aldrich), TcdA (12.5 µg/ml, a kind gift from Dr Tor Savidge, Baylor Medical College, Houston and Prof Charalabos Pothoulakis, David Geffen School of Medicine UCLA, CA),

GABA (100 µM) (Sigma Aldrich), the GABAA receptor antagonist bicuculline methiodide

(10 µM) (Frigo et al., 1987), the GABAB antagonist CGP54626 HCl (1 µM) (Brugger et al.,

1993) and the GABAC antagonist (1,2,5,6-Tetrahydropyridin-4-yl) methylphosphinic acid (TPMPA) (100 µM) (Ragozzino et al., 1996) (all from Tocris Bioscience, Avonmouth Bristol UK). TcdA was purified to homogeneity on SDS/PAGE from broth culture supernatants of C. difficile strain 10,463 as described by (Pothoulakis et al., 1988). 2.2 Analysis and statistics The electrophysiological data were analysed using AxoScope computer software (version 10.2.0.14, Axon Instruments U.S.A.). The injection of depolarising and hyperpolarising current pulses into the neurons was performed twice per cell, while the single and train stimulus regime was repeated 3 times to obtain averages for these parameters. Data are presented as mean ± SEM where n = number of neurons.

In the text below, quantitative comparisons are provided for numbers of APs and duration of firing for 300 pA depolarizing pulses. Statistical comparisons included an analysis of the firing over the whole range of current pulses using two way ANOVA, where statistical values are given for the entire curve. The Tukey method of post-hoc analysis was used to identify differences for individual current pulse amplitudes and are specified in the results. Where no specification exists, differences in firing are to be taken as occurring over the entire curve,

rather than at individual data points. 2-sample t tests were used for independent groups with P < 0.05 considered statistically significant.

3. RESULTS

3.1 TcdA increases excitability of myenteric AH neurons

Following a 90 min luminal incubation with TcdA (12.5 μg/mL), jejuna showed signs of mucosal damage such villus breakdown compared to saline controls.

The firing properties of TcdA -treated AH neurons were examined in LMMP preparations by injection of depolarising current pulses (50 pA increments, over a 50-350 pA range). Myenteric AH-neurons from TcdA-treated preparations (n= 7) were substantially more excitable than control neurons (n = 10). There was a significant difference in the stimulus response curves; TcdA-treated neurons fired more action potentials (P < 0.001 at 250–350 pA, Tukey’s post-hoc analysis) and for a longer duration (P < 0.001, at 300 pA, Tukey’s post-hoc analysis), than controls (firing at 300 pA control: 3.7 ± 0.7 APs; TcdA: 8.1 ±1.0 APs, at 300 pA, control: 76 ± 12 ms; TcdA: 170 ± 17 ms, Figure 4.1 ) with firing more likely to occur at 50 pA following TcdA incubation (Control: 0.9 ± 0.4 APs; TcdA: 2.6 ± 0.7 APs, P < 0.05, 2-sample t-test). While the neurons were more excitable, their maximum firing frequency was unchanged indicated by no change in the interspike interval (ISI) between the two groups (Control: 22 ± 1 ms; TcdA: 18 ± 3 ms, at 300 pA, P > 0.05).

TcdA-treated neurons were found to have a more depolarised resting membrane potential (RMP) compared to controls (Control: -55.5 ± 2.2 mV; TcdA: -47.1 ± 2.3 mV, P < 0.05,

Table 4.1). TcdA pre- treatment did not alter input resistance (IR), Ih current, the incidence of anode break action potentials (APs) or spontaneous APs in myenteric AH neurons (Table 4.1).

It was rarely possible to record from one neuron long enough to obtain all of the parameters required for full electrophysiological classification; antidromic action potentials often could not be triggered without moving the stimulating electrode, which frequently dislodged the recording electrode. Thus a full characterisation of membrane properties and synaptic input was limited to a subset of neurons under each condition (Tables 4.2 and 4.3). 6 of the 10 control neurons and 1 out of 7 TcdA-treated neurons were characterised (Table 4.2). Further

recordings from TcdA-treated neurons would be required to derive meaningful comparisons with controls; since TcdA supplies were extremely limited, this could not be pursued.

Figure 4. 1 TcdA makes myenteric AH neurons hyperexcitable. Example electrophysiological recordings of action potential firing at current pulses of 50 pA (top panels) and at 300 pA (bottom panels) from a control (A-A’) and TcdA (12.5 µg/ml) pre- treated (B-B’) neuron. Graphical representations of the mean number (C) and total duration (C’) of APs fired at each depolarizing current amplitude (50 - 350 pA). Following TcdA treatment, AH neurons fired significantly more APs (^P < 0.001) and for longer (^ P < 0.001) compared to neurons in control preparations.

3.2 GABA increases excitability of myenteric AH neurons

Following luminal incubation with GABA (100 µM), myenteric AH neurons (n = 12) displayed an increase in the number of APs fired (P<0.05) and the duration of firing throughout the pulse range (P < 0.01) when compared with controls (n=10) (firing at 300 pA control: 3.7 ± 0.6 APs; GABA: 4.5 ± 0.4 APs, control: 76 ± 12 ms; GABA: 103 ± 11 ms, Figure 4.2). GABA-treated neurons did not show any differences in firing at the lowest current pulse amplitude i.e. firing threshold (Control: 0.9 ± 0.4 APs; GABA: 1.0 ± 0.3 APs, P > 0.05, 2-sample t-test) or in RMP (Control: -55.5 ± 2.2 mV; GABA: -53.9 ± 1.5 mV, P >

0.05), and ISI (Control: 22 ± 1 ms; GABA: 22 ± 1 ms, at 300 pA, P > 0.05). IR, Ih current, the incidence of anode break APs or spontaneous APs were also unchanged (Table 4.1). Of the 12 GABA-treated neurons, 2 had their electrophysiology characterised in full. These neurons did not display any obvious differences in AP amplitude, AP duration D50, after- hyperpolarising potential (AHP) amplitude, after-depolarising (ADP) amplitude, slow excitatory post synaptic potential (sEPSP) amplitude, or sEPSP duration D50 compared to their controls (Table 4.2).

3.3 Co-incubating GABA with bicuculline enhances the excitability, while bicuculline alone produces hyperexcitability in myenteric AH neurons

Co-incubation with bicuculline (10 µM) (n = 5) further enhanced the increased excitability produced by luminal GABA compared to incubation with GABA alone, in both the number of APs fired (P < 0.001) and firing duration (P < 0.001) (firing at 300 pA GABA + bicuculline: 6.1 ± 0.7 APs; 148 ± 19 ms), with firing more likely to occur at 50 pA compared to GABA-treated neurons (GABA: 1.0 ± 0.3 APs; GABA + bicuculline: 2.4 ± 0.5 APs, P < 0.05, 2-sample t-test) (Figure 4.2). In fact, bicuculline pre-incubation in the lumen alone (n = 8) produced hyperexcitability in the firing number (P < 0.05) and duration (P < 0.01) in these neurons compared to GABA alone, but not compared to GABA+ Bicuculline (firing at 300 pA bicuculline: 5.4 ± 0.7 APs; 142 ± 17 ms) and did not alter the threshold of firing (GABA: 1.0 ± 0.3 APs; bicuculline: 1.3 ± 0.4 APs, P > 0.05, 2-sample t-test). For both GABA + bicuculline-treated and bicuculline-treated neurons, RMP (GABA+ bicuculline: -54.0 ± 2.0 mV; bicuculline: -50.0 ± 2.4 mV ) and ISI (GABA+ bicuculline: 21 ± 1 ms; bicuculline: 29 ± 5 ms, at 300 pA) were unchanged in addition to other neuronal properties including IR, the incidence of anode break APs and spontaneous APs compared to both controls and GABA-

100

treated neurons (Table 4.3). Membrane properties and synaptic input was not characterised in full or both GABA + bicuculline-treated and bicuculline-treated neurons (Table 4.3) due to recording instability.

Figure 4. 2 The effects of GABA and the GABAA antagonist Bicuculline, on the action potential firing properties myenteric AH neurons. Electrophysiological recordings are displayed showing a control neuron (A) a GABA (100 µM) -treated neuron (B) and a GABA & Bicuculline (Bic) (10 µM)- treated neuron (C) at 50 pA (A- C) and 300 pA (A’- C’) current pulses. Pre-treatment with GABA significantly increased the number of APs (D) fired during depolarisations and their firing duration (D’) over a range of current pulse amplitudes (*P < 0.05), (effect shown graphically). Firing was enhanced further following co-incubation of GABA with Bicuculline, compared to GABA alone, shown graphically in greyscale (D and D’) (^P < 0.001).

101

3.4 The GABAB antagonist CGP54626 and GABAC antagonist TPMPA inhibit the GABA induced excitability

The GABAB antagonist CGP54626 (1 µM) and the GABAC antagonist TPMPA (100 µM) co- incubated in the lumen with GABA, blocked the GABA-induced increased firing in AH neurons. Following co-incubation of GABA with CGP54626, myenteric AH neurons ( n= 6) displayed a significant reduction in the number (P < 0.01) and duration (P < 0.01) of action potentials fired when compared with pre-incubation with GABA alone (firing at 300 pA GABA: 4.5 ± 0.4 APs; GABA + CGP54626: 3.9 ± 0.7 APs, GABA: 103 ± 11 ms; GABA +

CGP54626: 91 ± 18 ms, Figure 4.3). Following co-incubation of GABA with the GABAC antagonist TPMPA, myenteric AH-neurons (n = 8) similarly showed a reduction in the number (P < 0.01) and duration (P < 0.001) of APs fired when compared with pre-incubation with GABA alone (firing at 300 pA, GABA + TPMPA: 3.6 ± 0.8 APs, GABA + TPMPA 65 ± 13 ms, Figure 4.3). Firing was returned to control level after the co-incubation with either the GABAB or GABAC antagonist. The threshold for firing was comparable to controls (Control: 0.9 ± 0.4 APs; GABA + CGP54626: 0.3 ± 0.3 APs; GABA + TPMPA: 0.3 ± 0.2 APs, both P>0.05). For both antagonist groups RMP (GABA + CGP54626: -57.7 ± 3.3 mV; GABA + TPMPA: -57.3 ± 1.9 mV ) and ISI (GABA + CGP54626: 24 ± 3 ms; GABA + TPMPA: 20 ± 3 ms, at 300 pA) in addition to other neuronal properties including IR, the incidence of anode break APs and spontaneous APs were no different from controls and GABA-treated neurons as shown in Table 4.3. Of the 6 GABA + CGP54626-treated neurons, 2 had their electrophysiology characterised in full. These neurons did not display any obvious differences in AP amplitude, AP duration D50, AHP amplitude, ADP amplitude, sEPSP amplitude, or sEPSP duration D50 compared to their controls and GABA-treated neurons (Table 4.3). Similarly 4 of 8 GABA + TPMPA-treated neurons had their electrophysiology characterised in full but did not show differences in these membrane and synaptic properties compared to controls and GABA-treated neurons (Table 4.3).

102

Figure 4. 3 Co-incubation with either a GABAB or a GABAC antagonist, prevents the GABA-induced hyperexcitability of myenteric AH neurons. Graphical representations of the number of APs fired and the duration of firing during prolonged depolarisations in myenteric AH neurons following pretreatment with GABA (100

µM) and a GABAB antagonist (1µM) (A-A’) and with GABA and a GABAC (100 µM) antagonist (B-B’) Co-incubating GABA with a GABAB antagonist significantly decreased the number of APs (A) fired during depolarisations and their firing duration (A’) over a range of current pulse amplitudes (#P < 0.01), effect shown in grey scale. A similar reduction in AP firing (B) and duration (B’) was observed following the co-incubation of GABA with a

GABAC antagonist (#P < 0.01, number) (^P < 0.001, duration), shown in grey scale.

103

Table 4. 1 Membrane properties of myenteric AH neurons following TcdA and GABA pre-treatment.

Parameter Control (n) TcdA-treated (n) GABA-treated (n)

*RMP (mV) -55.5 ± 2.2 (10) -47.1 ± 2.3 (6) 53.9 ± 1.5 (12)

Input Resistance (MΩ) 155.4 ± 8.3 (10) 144.5 ± 22.3 (7) 145.5 ± 12.4 (12)

Ih current (MΩ) 2.3 ± 1.9 (10) 1.4 ± 1.8 (7) 6.6 ± 2.2 (12)

Interspike interval (ISI) (ms) 22 ± 1 (9) 18 ± 3 (7) 22 ± 1 (12)

Spontaneous AP firing % 20% (10) 14% (7) 0% (12)

Incidence Anode Break APs % 30% (10) 43% (7) 25% (12)

*P < 0.05 Control vs Tcda and TcdA vs GABA. N = sample size, RMP = resting membrane potential, IR = input resistance, AP = action potential, (Ih) = hyperpolarisation-activated cation current. 2-sample t-test, Fisher’s exact test.

104

Table 4. 2 Membrane properties and synaptic potentials of myenteric AH neurons following TcdA and GABA pre-treatment.

GABA-treated Parameter Control (n) TcdA-treated (n) (n)

AP amplitude (mV) 58.7 ± 1.3 (6) 73.5 (1) 56.4 ± 0.9 (2)

APD50 (ms) 1.2 ± 0.04 (6) 1.1 (1) 1.5 ± 0.1 (2)

AHP amplitude (mV) 5.9 ± 0.5 (6) 6.5 (1) 3.8 ± 0.3 (2)

ADP amplitude (mV) 14.3 ± 1.5 (6) 17.7 (1) 13.5 ± 0.5 (2)

slow EPSP amplitude (mV) 14.5 ± 3 (5) 4.2 (1) 9.8 ± 0.7 (2)

AP = action potential, D₅₀ = half duration, AH P= after-hyperpolarising potential, ADP = after-depolarising potential, slow EPSP = slow excitatory post-synaptic potential. 2-sample t- test.

105

Table 4. 3 Membrane properties and synaptic potentials of myenteric AH neurons following various GABA receptor antagonist pre-treatments.

Parameter Control (n) GABA+Bic Bic (n) GABA+CGP GABA+TPMPA (n) (n) (n) RMP (mV) -55.5 ± 2.2 -54.0 ± 2.0 -50.0 ± 2.4 -57.7 ± 3.3 -57.3 ± 1.9 (10) (5) (4) (6) (6) Input 155.4 ± 8.3 146 ± 13 151 ± 7 201 ± 45 162 ± 11 Resistance (10) (5) (4) (6) (8) (MΩ) Interspike 22 ± 1 (9) 21 ± 1 (5) 29 ± 5 (5) 24 ± 3 (6) 20 ± 3 (8) interval (ISI) (ms) Spontaneous 20% (10) 0% (5) 0% (8) 33% (6) 12.5% (8) AP % Incidence 30% (10) 40% (5) 50% (4) 0% (6) 0% (8) Anode Break APs % AP amplitude 58.7 ± 1.3 (6) 58.3 ± 3.7 (2) 64.1 ± 3.3 (4) (mV)

APD50 (ms) 1.2 ± 0.04 (6) 1.1 ± 0.2 (2) 1.0 ± 0.1 (4) AHP amplitude 5.9 ± 0.5 (6) 9.0 ± 2.3 (2) 6.8 ± 1.0 (4) (mV) ADP amplitude 14.3 ± 1.5 (6) 12.0 ± 3.7 (2) 11.3 ± 0.9 (4) (mV) slow EPSP 14.5 ± 3 (5) 19.5 ± 0.1 (2) 4.5 ± 1.5 (2) amplitude (mV) Bic = Bicuculline, CGP = CGP54626, n = sample size, RMP = resting membrane potential, IR=input resistance, AP= action potential, D₅₀ =half duration, AHP = after-hyperpolarising potential, ADP = after-depolarising potential, slow EPSP = slow excitatory post-synaptic potential. 2-sample t-test, Fisher’s exact test.

106

4. DISCUSSION

The ENS has been implicated in the pathophysiological effects of TcdA (Xia et al., 2000), but its contribution remains largely undefined. Here I examined the effect of luminal in vitro incubations of TcdA on the ENS in guinea pig jejunum and found that TcdA exposure leads to hyperexcitability of myenteric AH neurons. Further, GABA, a metabolic by-product of the microbiome that is implicated in increasing susceptibility to CDI (Dann et al., 2015), was also found to make myenteric AH neurons hyperexcitable presumably acting at the level of the mucosa. The results of this study suggest that myenteric AH neurons are a common pathway through which enterotoxins and some bacterially-derived metabolites act.

4.1 TcdA makes myenteric AH neurons hyperexcitable

In vitro incubation of TcdA in guinea pig jejunum increased the excitability of myenteric AH neurons, causing them to fire at smaller depolarisations and for longer but the maximum firing frequencies remained unchanged. In addition, TcdA depolarised the RMP of myenteric AH neurons. These changes in excitability were not accompanied by a change in spontaneous firing, so it is likely that the increased excitability is due to a change in the properties of these AH neurons rather than a change in their inputs. However, due to limited availability of toxin, the full electrophysiological characterisation of neurons following toxin treatment was restricted so that other parameters such as changes in sEPSPs, which would confirm this, could not be definitively determined. An examination of other membrane property changes should be a point of future study. This could well include toxin-induced alterations of the AHP - the main modulatory mechanism and source of stabilising inhibition for these neurons (Wood and Mayer, 1979a; Thomas et al., 2000; Bertrand and Thomas, 2004) – which occur following CT exposure (Chapter 3). In models of inflammation in the guinea pig colon, the AHP is reduced in amplitude (Linden et al., 2003; Lomax et al., 2005) and since TcdA induces a powerful inflammatory response (Castagliuolo et al., 1994) this could be a possible mechanism. However inflammation has been shown to augment the hyperpolarisation- activated action current (Ih ) (Linden et al., 2003), but this was found to be unchanged after TcdA incubation. Thus, a detailed examination of such membrane properties is required.

107

4.2 Pathways activated by TcdA

While CT does not penetrate the mucosa during this type of incubation (Gwynne et al., 2009), TcdA is known to cause gross mucosal damage and a breakdown of epithelial barrier function (Hecht et al., 1988; Kelly et al., 1994). Since signs of mucosal damage were observed in TcdA-treated jejuna, the possibility that the toxin is directly interacting with enteric neurons cannot be excluded. In the SMP of the guinea pig small intestine, TcdA excites AH and S neurons when applied directly to the cell body by inducing membrane depolarisation and spontaneous discharge (Xia et al., 2000). By direct action, TcdA could therefore be contributing to hyperstimulation of mucosal secretion via excitation of the secretomotor neurons responsible for driving secretion. Whether or not TcdA incubated luminally directly interacts with myenteric or submucosal neurons in this way to alter their excitability, is unknown. If TcdA directly interacts with neurons in the SMP that project to the MP (Kirchgessner and Gershon, 1988b; Song et al., 1998), changes in their excitability could affect other elements of the circuit and thus be responsible for the enhanced excitability of myenteric AH neurons observed.

Aside from the possibility of direct excitation, luminal TcdA might also alter the firing properties of enteric neurons via indirect neural pathways. Capsaicin treatment, which only acts on extrinsic primary afferents, not enteric neurons (Bartho et al., 1982; Sharkey et al., 1984) prevents the TcdA-induced response in vivo (Castagliuolo et al., 1994) implicating the involvement of extrinsic neural pathways. Indeed extrinsic axon collateral innervation of the MP exists in the guinea pig small intestine since stimulation of mesenteric nerves that supply the gut evokes sEPSPs in myenteric AH neurons (Takaki and Nakayama, 1990). Thus, it is possible that that the toxin activates extrinsic neurons with axon collaterals in the MP to induce hypexcitability in the MP.

The extrinsic pathways activated by TcdA involve substance P (SP) and calcitonin gene- related peptide (CGRP)- mediated inflammation, including mast cell degranulation (Castagliuolo et al., 1994; Pothoulakis et al., 1994; Keates et al., 1998), the products of which are thought to act at the mucosa and to contribute the hypersecretion (Pothoulakis et al., 1993). Mast cell-deficient and SP receptor (NK1-R)-deficient mice show attenuated responses to TcdA (Castagliuolo et al., 1998a; Wershil et al., 1998) and since SP is a putative neurotransmitter for slow excitatory transmission in the ENS (Alex et al. 2001; Gwynne and 108

Bornstein 2007b) enteric nervous input may reinforce the secretory and inflammatory responses of TcdA (Xia et al., 2000). In agreement with this, SP Receptor (SPR) activation in enteric neurons in the SMP and MP in response to intraluminal TcdA has been observed in the rat ileum (Mantyh et al., 1996), thus the changes in excitability I observed might well be the result of tachykinin-mediated excitation. In fact neurokinin (NK) receptors are expressed on enteric neurons in the guinea pig small intestine (Portbury et al., 1996; Johnson and Bornstein, 2004) while CT acts to produce sustained excitation in neurons via tachykinin- mediated pathways in both plexuses of the guinea pig small intestine (Gwynne et al., 2009; Chapter 3).

4.3 GABA makes myenteric AH neurons hyperexcitable

GABA incubated luminally was found to increase the excitability of the same population of neurons made hyperexcitable by TcdA and CT. This change was not accompanied by a change in other membrane properties or synaptic input, however the sample size examining this was limited. Since the increase in excitability produced by GABA was small, the mechanism underlying this may be difficult to determine. Effects in excitability were prolonged, which may indicate changes in membrane properties rather direct action of GABA and so a more extensive examination of these properties is needed to confirm this.

In the guinea pig ileum, GABA applied directly onto myenteric AH neurons causes depolarisation via activation of chloride conductance, mediated by GABAA receptors and is rapidly desensitising, while S neurons are insensitive to GABA (Cherubini and North, 1984a). No change in RMP accompanied the increased excitability observed in this study, thus it is likely that the pathway through which GABA acts luminally differs from the direct cellular action of GABA on myenteric AH neurons.

It is probable that GABA also functions as an endocrine mediator and the excitation of AH neurons may be the result of a release of neural intermediates. GABA is expressed in mucosal enteroendocrine cells in the rat stomach and intestine (Davenger et al., 1994; Krantis et al., 1994), it has been shown to modulate release of histamine from mast cells in the guinea pig ileum (Luzzi et al., 1987) and prostaglandins from interstitial cells in the rabbit small intestine (Girdhar et al., 1981). Exposure to histamine and prostaglandin E2 (PGE2) in vitro increases the excitability of myenteric AH neurons in the guinea pig small intestine (Dekkers

109

et al., 1997a; Tamura and Wood, 1992). Further, GABA has been shown to inhibit serotonin

(5-HT) release from EC cells in the guinea pig small intestine via GABAA and GABAB receptors (Schwörer et al., 1989), while GABAB receptors colocalise with 5-HT in mucosal cells of the rat stomach and intestine (Nakajima et al., 1996). Consequently it is likely that GABA regulates the release of 5-HT from the mucosa via multiple GABA receptor subtypes. 5-HT is of particular interest as it is a key mediator in regulating motor (Grider et al., 1996; Ellis et al., 2013; Gwynne et al., 2014) and secretory reflexes (Cooke et al., 1997a, 1997b) and importantly excites intrinsic sensory neurons in the SMP and MP that are the afferents of secretomotor pathways in the guinea pig (Kirchgessner et al., 1992; Bertrand et al., 2000). GABA may well indirectly activate reflex pathways running through though the MP via 5- HT, but it is unknown if GABA-induced modulation of epithelial 5-HT is the cause of excitation in this study.

Further, TcdA also directly acts at the mucosa to release a range of inflammatory mediators including PGE2 (Pothoulakis and Lamont 2001) as well as triggering extensive mucosal mast cell degranulation and histamine release via extrinsic pathways (Castagliuolo et al., 1994), while CT has been long known to induce a release of 5-HT from EC cells (Beubler et al., 1989b; Bearcroft et al., 1996). Thus, it might be that enterotoxin-excited pathways converge early on at the mucosal epithelium to excite sensory neurons downstream. An interaction of GABA at the mucosa may hypersensitize the underling neural networks to exacerbate the effects of TcdA and contribute to disease susceptibility in CDI observed in human and animal models (Dann et al., 2015). Whether GABA potentiates the effects of TcdA is unknown. Due to limited toxin supplies, this question was not addressed but remains a vital point to examine, ideally using an in vivo model such as the ileal loop which is well suited for prolonged incubations that more closely resemble the pathophysiology of CDI.

4.4 GABA acts via GABA receptor subtypes in separate pathways

Co-incubation with GABA receptor antagonists revealed distinct pathways through which luminal GABA may act. As mentioned previously depolarisation of AH neurons by GABA (Cherubini and North, 1984a) is direct, while the luminal exposure is likely to recruit a number of pathways. The data in this study implicate a dual effect of GABA: a pathway excited by GABA acting through GABAC and GABAB receptors that is responsible for the hyperexcitability of AH neurons and another where GABA suppresses the excitability of AH 110

neurons via GABAA receptors. When GABAA receptors are antagonised and therefore the inhibitory influence of GABA is removed, this results in the excitatory GABAC pathway to be larger and accounts for the enhanced excitability produced by the incubation of GABA and bicuculline together. Since bicuculline per se induced prolonged excitation of these neurons it is likely that endogenous GABA is activating a mucosal pathway via GABAA receptors that produces on going inhibition of AH neurons, most likely via an indirect neural mechanism. This might well be due to GABA inhibiting the release of sensory intermediates such as 5-HT from the mucosa, where bicuculline has been shown to enhance 5-HT release (Schwörer et al., 1989). Indeed, myenteric and submucosal AH neurons are excited by via 5-HT receptors at their nerve terminals (Kirchgessner et al., 1992; Bertrand et al., 2000; Pan and Gershon, 2000).

The pathways excited by GABAB and GABAC receptors probably operate separately to the inhibitory GABAA-mediated pathway since, GABAB receptors have also been shown to inhibit 5HT release from the mucosa in the guinea pig small intestine (Schwörer et al., 1989), but produce increased excitability of myenteric neurons in this study. There is also the possibility of a third pathway in which GABAB activation inhibits the GABAA pathway, promoting that of GABAC which would also account for the often inhibitory role of GABAB receptors in functional studies (Giotti et al., 1983; Ong and Kerr, 1987; Marcoli et al., 2000).

Where GABAB and GABAC receptors are expressed in rodent models remains relatively unclear. While GABAB receptors colocalise with 5-HT in the mucosa of the rat stomach and intestine (Nakajima et al., 1996) they are also localised to muscle fibres and neurons in both plexuses. Very little data about GABAC receptor function and expression exist, but it has been reported that GABAC receptors are expressed on some inhibitory and excitatory neuronal subtypes in the rat colon (Fletcher et al., 2001). Thus it is challenging to ascertain where GABA is acting within enteric pathways to produce prolonged excitability increases of AH neurons and this should be a future point of investigation. Since antisera raised against GABA receptors subtypes have been shown to be mostly effective in the rat, in addition there is a lack of studies showing GABAC expression in other species, immunohistochemical localisation of receptors was not pursued.

111

4.5 AH neurons, enterotoxins and GABA in the context of the circuit

The neural pathways through which CT and TcdA operate are typically thought of as separate. The hypersecretion induced by CT occurs via nicotinic synapses and the actions of VIP within the ENS and mucosa (Eklund et al., 1979; Cassuto et al., 1981b, 1982a). TcdA activates extrinsic sensory neurons containing SP and CGRP and an inflammatory response involving degranulation of mast cells (Pothoulakis and Lamont 2001), while participation of the ENS in the effects of TcdA has only been implicated by very few studies (Mantyh et al., 1996; Xia et al., 2000). Analogous to the effects on the ENS following luminal exposure to CT (Chapter 3), TcdA induced hyperexcitability in myenteric AH neurons. My study is the first to demonstrate a generality in the neural actions of luminal enterotoxins in the gut as the mechanisms they activate converge on the same neuronal population. It is also worth mentioning that an earlier study has shown that in the guinea pig, a parasitic gut infection with the nematode T. spiralis produces hyperexcitability in myenteric AH neurons in the afferent limb of gut reflexes (Chen et al., 2007). Moreover, luminal GABAergic signals appear to present a risk factor in disease pathogenesis of CDI (Dann et al., 2015) and my study shows they act on the same enteric sensory system. This suggests that AH neurons play a key role in the mechanisms underlying microbial interaction with the ENS. Indeed myenteric AH neurons display supressed excitability in germfree mice lacking microbiota, which is restored following colonisation with bacterial communities (McVey Neufeld et al., 2013).

Thus, it is clear that AH neurons perform an integral role in the plasticity of enteric circuits in healthy and perturbed states, acting as a ‘relay centre’ for incoming luminal signals. This is likely due to unique characteristics such as prolonged AHPs that function in gating their outputs (Wood and Mayer, 1979a; Bertrand and Thomas, 2004) and their ability to form self- reinforcing interconnected networks that have the capacity to encode graded ongoing stimuli (Thomas et al., 2000; Thomas and Bornstein, 2003; Chambers et al., 2005).

Activation of the afferent limb of secretomotor pathways by TcdA may contribute to the hypersecretion produced by the toxin, since myenteric AH neurons in the guinea pig are known to project to and synapse with most other myenteric neurons and provide excitatory input to neurons in the SMP responsible for driving secretion (Bornstein et al., 1987; Galligan et al., 1988; Furness et al., 1990b; Furness et al., 2004a; Gwynne and Bornstein, 2007b). They may also contribute to changes in the gut motor patterns that TcdA has been shown to 112

produce in rabbit in vivo (Burakoff et al., 1995). Whether such functional changes are produced by luminal TcdA in other species including the guinea pig remains to be investigated.

The role of GABA in gut function is unclear, and so it is often studied in the MP due to readily measurable contractility output. The consequences of the prolonged increased excitability of myenteric AH neurons produced by luminal GABA and the extent of its participation in GABA-induced motility is challenging to predict. This is simply due to the difficulty in predicting the overall functional output of GABA exposure due to various receptor subtype expression and activation (Tonini et al., 1987; Krantis and Harding, 1987; Zizzo et al., 2007; Auteri et al., 2014) and the species studied (Tonini et al., 1989b; Auteri et al., 2014). Nonetheless the results of this study suggest that GABA induces sustained changes in enteric circuity that are likely to participate in motor reflexes.

The changes in excitability of AH neurons examined in this study were prolonged, persisting for hours after exposure. This indicates potential alterations in key membrane events or similar mechanisms and it is therefore likely that excitability changes are plastic and not due to direct activation of transiently excited receptors. Thus a number of indirect pathways are probably involved in neuronal activation, which are best studied using in vivo incubation models that better preserve extrinsic and mucosal pathways. Consequently the effects of enterotoxins on the enteric circuitry have been examined the in vivo ileal loop for Chapters 6 and 7.

4.6 Conclusions and future directions

This chapter examined the involvement of the ENS following luminal incubations of TcdA and GABA- a bacterially-derived metabolite associated with CDI, in vitro. I demonstrated that TcdA produces prolonged hyperexcitability in myenteric AH neurons, whether this occurs directly or via an indirect neural pathway will be an important question to address in future studies. I established that GABA also induces increased excitability in these same neurons, possibly via at least 2 distinct enteric pathways. It is unknown whether activation of these pathways occurs via neural intermediates; this will be an interesting point to investigate further. Additionally, to provide insight into the role of GABA in CDI susceptibility, the

113

interaction between both luminal TcdA and GABA together should be an essential focus of future work.

114

CHAPTER 5: GABAERGIC TRANSMISSION IN THE MP OF THE MURINE ILEUM

ABSTRACT

γ-Aminobutyric Acid (GABA) and its receptors, GABAA,B,C, are expressed in several locations along the gastrointestinal tract. Nevertheless, a role for GABA in enteric synaptic transmission remains elusive. In Chapter 4 I demonstrated that both luminal GABA and TcdA activate neural pathways that excite the same population of intrinsic sensory neurons. This chapter aims to examine the nature GABAergic transmission in the gut given the emerging role of luminal GABA in CDI pathogenesis. In the mouse small intestine, about 8% of all myenteric neurons were found to be GABA-immunoreactive (GABA-IR) including some Calretinin-IR and some neuronal nitric oxide synthase (nNOS-IR) neurons. Further Ca2+ imaging was used to examine the activity of myenteric neurons where many that did not express GABA or nNOS (the majority), some GABA- IR, Calretinin-IR or Neurofilament-M (NFM)-IR but rarely nNOS-IR neurons responded to exogenous GABA application with 2+ [Ca ]i transients. GABAA,B,C and nicotinic receptor blockade reduced the number of neurons responding to GABA application. To examine the possible effects of endogenous GABA, interganglionic fibre tracts were electrically stimulated. Single pulse-evoked calcium responses were unaffected by blockade of each of the 3 GABA receptor subtypes. Calcium responses evoked by 20 pulse trains were potentiated by GABAC receptor blockade. These data suggest that GABAA and GABAB receptors are not involved in synaptic transmission, but for the first time here, suggest a novel role for GABAC receptors in modulating slow synaptic transmission.

1. INTRODUCTION

Several authors have proposed γ-Aminobutyric Acid (GABA) to be a putative neurotransmitter in the ENS. GABA is known to be expressed in the gut mucosa, musculature, and enteric neurons and is a prominent by-product of the microbiome (Jessen et al., 1986; Krantis, 2000; Barrett et al., 2012; Savidge, 2016). Of particular interest is the role

115

of luminal GABA signals in gastrointestinal disease states where data from both human and rodent models show microbial-derived GABAergic signals confer disease susceptibility to Clostridium difficile(C. diff) infection (CDI) (Dann et al., 2015). In Chapter 4 I demonstrated that both C. diff toxin A (TcdA) and luminal GABA act through the same sensory system of the ENS in the small intestine. However, most work examining GABA in the gut has been performed in the colon where neural distribution of GABA appears to greater and more readily studied (Krantis, 2000). Since the pathophysiology of TcdA will be examined in vivo using the well-established mouse ileal loop model in Chapter 7, and in light of our new exciting data implicating a role for GABA in CDI pathogenesis, the focus of this chapter is GABAergic transmission the murine small intestine .

Within enteric plexuses, GABA-immunoreactive (IR) neurons are predominantly found within the MP and are thought to be interneurons and motor neurons (Jessen et al., 1986; Hills et al., 1987; Sang and Young, 1996). Across a number of rodent models between 5-14% of myenteric neurons in the colon are GABA-IR (Jessen et al., 1979; Sang and Young, 1996; Krantis, 2000; Li et al., 2011), however GABA-IR neurons are sparse in the small intestine. While studies have shown that a small percentage of neurons in the MP of the mouse small intestine are GABA-IR, GABA is otherwise often excluded from colocalisation studies and is generally not studied in detail in this region (Sang and Young, 1996; Li et al., 2011). Yet GABA-IR terminals are present within the MP of the murine small intestine (Sang and Young, 1996), which certainly suggests a role for GABAergic transmission in this region of the gut.

Within the ENS, GABA can have both an excitatory and inhibitory influence on neurons depending on the receptor it activates. GABA can act through excitatory ionotropic GABAA and GABAC receptor subtypes which are both pentameric chloride channels (Fletcher et al., 1998; Auteri et al., 2015). Unlike its inhibitory role in the central nervous system, in the ENS

GABA excites enteric neurons via GABAA and GABAC receptors due to their high intracellular Cl- concentration (Liu et al., 2013). Additionally, GABA can act through inhibitory metabotropic GABAB receptors, which operate either presynaptically to depress Ca2+ influx via voltage-activated Ca2+ channels to decrease ACh release from cholinergic neurons, or postsynaptically where they are coupled to inwardly-rectifying K+ channels (Hyland and Cryan, 2010; Auteri et al., 2015).

116

Studies have demonstrated neuronal GABAA, GABAB and GABAC receptor expression in both the small and large intestine of the rat (Krantis et al., 1995; Nakajima et al., 1996; Poulter et al., 1999; Fletcher et al., 2001). In the large intestine of the mouse, immunohistochemical studies have demonstrated that GABAA and GABAB receptors are expressed on enteric neurons (Casanova et al., 2009; Seifi et al., 2014), while functional contractility studies in the small intestine of the mouse indicate that all 3 receptor subtypes are in fact present in the ENS and contribute to gut function in this region (Sanger et al., 2002; Zizzo et al., 2007). Indeed, most of what is actually known about GABA receptor expression comes from pharmacological functional studies where use of GABA receptor agonists and antagonists reveal that these receptors exist in both excitatory and inhibitory reflex pathways.

Activation of the GABAA and GABAC receptor system in the mouse typically either leads to the activation of cholinergic enteric neurons resulting in contraction in gut muscle response (Zizzo et al., 2007; Auteri et al., 2014) or activation of nNOS inhibitory motor neurons that relax gut muscle (Zizzo et al., 2007). GABAB receptors appear to exist in pathways resulting in reduction in ACh release from neurons and a consequential reduction in contractile activity (Sanger et al., 2002; Auteri et al., 2014). The net functional output of GABA acting at these receptors however, is difficult to predict and varies depending on the species studied (Frigo et al., 1987; Auteri et al., 2014) and the specific localization of the different GABA receptor subtypes in gut regions (Krantis and Harding, 1987; Krantis et al., 1998). Thus, while GABA and its receptors have been implicated in the functional physiology of the gut, the nature of GABAergic transmission remains elusive; indeed a synaptic function of GABA and its receptors in the ENS has not in fact been identified.

In this multifaceted study, an array of techniques including Ca2+ imaging making use of Wnt1-Cre;R26R-GCaMP3 mice that express a genetically encoded fluorescent calcium indicator in the ENS allowing a broad survey of neural transmission, was used to elucidate the evasive role of GABA in the gut. This study is the first to identify myenteric neuronal subtypes that express functional GABA receptors via the use of specific GABA receptor antagonists and post-hoc immunohistochemistry. My data suggest that GABAA and GABAB receptors do not have a synaptic function within the ENS but have I demonstrated a novel role for GABAC receptors in modulating slow synaptic transmission.

117

2. METHODS

2.1 Mice

Mice with a C57Bl6 background of either sex aged 8-12 weeks including Wnt1-Cre;R26R- GCaMP3 mice, that express the fluorescent Ca2+ indicator in all neural crest-derived cells (including enteric neurons and glia) (Zariwala et al., 2012; Boesmans et al., 2013), were killed by cervical dislocation; a procedure approved by the University of Melbourne Animal Experimentation Ethics Committee (see Chapter 2).

2.2 Immunohistochemical analysis of the expression of GABA in neurons and its co- expression with other neuronal subtype markers

Wholemounts of ileum from mice were immunostained. Segments of distal ileum were microdissected to remove the mucosa, submucosa and circular muscle, exposing the myenteric plexus and then permeabilized (Chapter 2, section 5.1). Preparations were incubated in primary antisera overnight at 4°C (Table 5.1) to identify GABA, neuronal nitric oxide synthase (nNOS) and calretinin (calr) -immunoreactive (IR) populations. Incubations in secondary antisera (Table 5.2) were at room temperature for 2 hrs and 15 mins. All preparations were rinsed of excess secondary antisera 3 times with PBS, for 10 mins before they were mounted on slides with Dako fluorescent mounting medium (Carpinteria, California, USA).

2.3 Imaging, analysis and statistics

For colocalisation studies, wholemount preparations were viewed using a Zeiss Axio Imager M2 microscope and images were acquired with an Axiocam 506 mono camera using Zen 2.3 (blue edition) software (all from Zeiss, Australia). Images were taken using 20x or 40x objectives.

The proportion of each neuronal subtype (calr-IR, NOS-IR and GABA-IR), was obtained by examining co-expression with the pan neuronal marker Hu. At least 200 Hu+ cell bodies were examined in each preparation. The mean proportion of each neuronal subtype was determined by calculating averages from 3 animals. The data are expressed as mean ± SEM and n = the number of cells examined. Statistical analyses were performed using unpaired t-

118

tests with P < 0.05 considered statistically significant. Comparisons were performed using using GraphPad Prism 5.0 (GraphPad Softwares, San Diego California).

For analysis of GABAergic varicosities apposing calr-IR and NOS-IR neuronal subtypes, up to 5 micrographs of ileal myenteric preparations were obtained using a Zeiss LSM800 confocal microscope (Zeiss, Australia). Approximately 30 neurons from 3 animals were examined in total for each subtype of enteric neuron. High-resolution confocal z-stacks (.czi files) with a sampled resolution of 1024 x 1024 pixels using a 63x/1.40 Oil DIC M27 objective, with a 1.8 x software zoom and z steps of 0.43µm were obtained. Close contacts were investigated qualitatively by individually stepping through each z-stack using the image analysis software Image J and identified as an overlap between two fluorophores on a merged image. 3D rendered images of neurons receiving GABA varicosities were generated using 3D image analysis software Imaris (Bitplane, version 8.4).

2.4 Tissue preparation for Ca2+ imaging

A segment of the distal ileum was removed from the animal, placed in physiological saline and dissected to reveal the myenteric plexus and attached longitudinal muscle (LMMP) as described in Chapter 2, section 6.1. A maximum of 5 rings could be prepared from each segment of ileum. Preparations were superfused (1 ml/min) with 95% O2: 5% CO2 bubbled physiological saline at room temperature throughout the experiment via a gravity-fed inflow system that included a manual valve to switch between saline and drug-containing saline solutions. Neurons within the myenteric ganglia were stimulated either chemically or electrically as described in Chapter 2, section 6.2. 2+ Following the first spritz application of GABA, the amplitude of GABA-evoked [Ca ]i responses was significantly reduced at the second time point suggesting desensitization of receptors, in accordance with Cherubini and North (1984a). For this reason the spritz pipette was moved away from the ganglion between each application to prevent desensitisation to GABA.

To test the effects of antagonists, drug-containing saline was superfused for 10 mins into the organ bath after either the first spritz or electrical stimulation regime, so that the first spritz or electrical time point was taken as the control response. Each ringed preparation was only

119

exposed to an antagonist or a combination of antagonists, once. After some experiments, tissues were fixed overnight with 4% formaldehyde at 4°C and examined post hoc using immunohistochemistry to identify neuron subtypes that responded.

2.5 Post-hoc immunohistochemistry for Ca2+ imaging experiments

Myenteric preparations from Ca2+ imaging experiments were incubated in primary antisera for 48-72 hours (Table 5.1) to identify GABA, nNOS, neurofilament-M (NFM) and calr-IR neurons. Incubations in secondary antisera (Table 5.2) were room temperature for 2 hrs and 15 mins. All preparations were rinsed of excess secondary antisera 3 times, for 10 mins before they were mounted on slides with Dako fluorescent mounting medium (Carpinteria, California, USA).

2.6 Drugs

Drugs used included GABA, hexamethonium bromide (both from Sigma Aldrich, Castle Hill, New South Wales, Australia), Bicuculline, CGP54626 and TPMPA (all from Tocris Bioscience, Avonmouth Bristol UK). All drugs were diluted in distilled water to make stock solutions and then again in physiological saline to their working concentration on the day of experimentation.

2.7 Analysis and statistics

Analyses were performed using custom-written directives in IGOR Pro (WaveMetrics, Lake Oswego, Oregon, USA) (Chapter 2, section 6.3). Regions of interest were drawn over a selected area of the cytoplasm for each neuron, excluding the nucleus because GCaMP3 is absent from the nuclei (Tian et al., 2009; Yamada and Mikoshiba, 2012). The intensity of the 2+ intracellular calcium ([Ca ]i) transient signal for each response was calculated and expressed 2+ 2+ as the maximum increase in [Ca ]i from the baseline signal (ΔFi/F0). [Ca ]i transients were only considered if the intensity of the transient signal was more than 5 times the intrinsic noise. For both time control and antagonist experiments the ΔFi/F0 of the second GABA spritz or second electrical stimulation response, is normalised and expressed as a fraction of

120

the first (%ΔFi/F0). A minimum of 3 animals were examined for each experimental set; unless stated otherwise.

Ganglia of interest that were examined with Ca2+-imaging were processed by post-hoc immunohistochemistry and relocated using a fluorescence microscope and images were taken using 20x or 40x objectives. Responding GCaMP+ neurons were identified immunohistochemically by matching the micrographs with the Ca2+ imaging videos.

Data are presented as the mean % ΔFi/F0 of the control ± SEM where n = number of neurons examined or as mean ± SEM and n = the number of cells examined. Statistical analyses were performed using unpaired t-tests with P < 0.05 considered statistically significant. Comparisons were performed using using GraphPad Prism 5.0 (GraphPad Softwares, San Diego California).

Table 5. 1 Primary antisera used for immunostaining.

Primary Antisera Raised in Dilution factor Source GABA Rabbit 1:1000 Sigma nNOS Sheep 1:1000 Gift from P. Emson Hu Human 1:5000 Gift from Dr V. Lennon Calretinin Goat 1:1000 SWANT NFM Rabbit 1:500 Merck Millipore

Table 5. 2 Secondary antisera used for immunostaining.

Secondary Antisera Raised in Dilution factor Source Anti-rabbit AF 647 Donkey 1:400 Molecular Probes Anti-sheep AF 594 Donkey 1:100 Molecular Probes Anti-sheep AF 488 Donkey 1:400 Molecular Probes Anti-human 594 Donkey 1:750 Jackson Immuno Labs

121

3. RESULTS

3.1 Expression of GABA in the distal ileal myenteric plexus

In this study, I found that GABA immunoreactivity was expressed in the somata of 8 ± 0.7% of Hu-IR cells in the myenteric plexus of the mouse ileum (n = 3 animals, at least 200 Hu-IR neurons per preparation; 2087 Hu-IR neurons examined in total, Figure 5.1). Previous studies have also reported GABA-IR neurons in the mouse small intestine constitute less than 10% of the myenteric population (Sang and Young 1996; Li et al., 2011).

In mouse ileum, most (90%) nNOS-IR myenteric neurons are inhibitory motor neurons to muscle while calr–IR neurons have been identified as interneurons and excitatory neurons to longitudinal and circular muscle and also intrinsic sensory neurons (ISNs) (Qu et al., 2008). nNOS immunoreactivity was observed in 24 ± 2% of neurons (n = 3 animals; 1023 Hu+ neurons, Figure 5.1) while 33 ± 1 % of neurons were immunoreactive for calretinin (n = 3 animals; 1064 Hu+ neurons, Figure 5.1). Comparable proportions have been observed in the murine ileum previously (Sang and Young, 1996; Qu et al., 2008; Li et al., 2011). Further to this, of the total number of Hu-IR neurons examined, GABA colocalised with some nNOS-IR (2 ± 0.1 %) and calr-IR neurons (4 ± 0.6%) Figure 5.1.

Next, GABA-IR varicosities which indicate presynaptic sites containing GABA (Jessen et al., 1986; Mann et al., 1997) were examined. The markers of potential neuronal targets (nNOS and calretinin) were selected because they reveal the whole shape of the neuron (Neal and Bornstein 2007a, 2008) allowing for analysis of GABA close contacts.

GABA-IR terminals were identified surrounding GABA-IR and nNOS-IR neurons within myenteric ganglia and in the muscle. Appositions were qualitatively defined as apparent contacts between the target neuron and GABA varicosity with no intervening black pixel (Neal and Bornstein 2007a, 2008). GABA-IR varicosities were found to closely appose the cell bodies of 61 ± 13 % calr-IR neurons (total of n = 33 calr-IR neurons examined) and 73 ± 18% of nNOS-IR neurons (total of n = 30 nNOS-IR neurons examined) Figure 5.2.

122

Figure 5. 1 Immunoreactivities of Hu, GABA, nNOS and calretinin in the distal ileum of the mouse. Images of representative myenteric ganglia, illustrating neurons stained for calr, (A), Hu (B) GABA (C), merged (D) (top row) and nNOS (E), Hu (F) and GABA (G) merged (H) (bottom row). Merged image of Hu, calr and GABA (D) demonstrates colocalisation of GABA with calr in some Hu-IR neurons. Merged image of Hu, nNOS, and GABA (H) show no apparent colocalisation of GABA with nNOS in Hu-IR neurons. Scale bars = 20μm. Filled arrows indicate colocalisation of GABA with the neuronal subtype marker, whereas open arrows indicate a lack of colocalisation between the markers.

123

Figure 5. 2 GABA varicosities form close contacts with neurons immunoreactive for calretinin and nNOS. 3D -rendered images of confocal micrographs showing a calr-IR (A) and a nNOS-IR neuron (B, asterisk) receiving close contacts from GABAergic appositions (A’, B’).

124

3.2 Some myenteric neurons including GABA-IR neurons respond to exogenous GABA 2+ with [Ca ]i transients

Overall, 10 ganglia (from 6 animals) were exposed to GABA (1 mM) spritz and some (63/290; 22 ± 3%) GCaMP+ cells, identified as neurons based on cell body size (~20 µm 2+ diameter) (Gabella and Trigg, 1984), responded with [Ca ]i transients (∆Fi/F0 = 0.23 ± 0.03, n = 63). The vast majority of all GABA responding GCaMP+ neurons were found to be GABA-/NOS- (72 ± 6%) or GABA-IR (23 ± 7%) (Figure 5.3). They were rarely NOS-IR or GABA-IR/NOS-IR. The overall amplitude of calcium transients of GABA-/nNOS- responders (GABA-/nNOS-: ∆Fi/F0 = 0.24 ± 0.03, n = 47) was no different to that of the other immunoreactive groups. However, nNOS-IR responders exhibited a significantly smaller response amplitude than GABA-IR responders, (GABA-IR: ∆Fi/F0 = 0.21 ± 0.03, n =

11, nNOS-IR: ∆Fi/F0 = 0.09 ± 0.02, n = 4, P < 0.05). Only one responder was GABA-

IR/nNOS-IR (GABA-IR/nNOS-IR: ∆Fi/F0 = 0.6896, n = 1). Interestingly within each ganglion, most GABA-IR neurons responded to GABA spritz (75 ± 12 %), while nNOS-IR neurons rarely did (3 ± 2 %).

Calretinin and neurofilament M (NF-M) are markers of excitatory and sensory neurons in the murine gut (Sang and Young, 1996; Qu et al., 2008). As I found that the vast majority of GABA-responding neurons were not GABA-IR/nNOS-IR and had large somata (indicative of enteric sensory neurons, Qu et al., 2008) and some calretinin neurons receive close contacts with GABAergic varicosities (Figure 5.2), preliminary experiments were conducted by another group member to further examine the identities of GABA-responding myenteric 2+ neurons. Of 16 neurons which displayed GABA-evoked [Ca ]i responses 7 were identified as calr-IR neurons and 9 were NFM-/calretinin- neurons via post hoc immunohistochemistry.

125

Figure 5. 3 Identity of neurons in the mouse ileum responding to pressure injection of GABA (1mM). 2+ Fluorescence images of some neurons that responded to GABA spritz with [Ca ]i transients highlighted by coloured arrows (A-A’). Preparations were processed by post-hoc immunohistochemistry to identify the neuronal subtype of GCaMP+ cells (B-E). Some responders were GABA-IR (indicated by green and red-coloured arrows) (C). Open arrows indicate nNOS-IR neurons that did not respond to GABA (D). Both responding and non- responding neurons are shown in the merged image (E). Scale bar = 20µm. Colour-matched 2+ corresponding traces of GABA-induced [Ca ]i responses of previously highlighted neurons in (A’), black arrow indicates GABA spritz (F). A pie chart displaying the percentage of neuron subtypes that responded to exogenous application of GABA. The majority of responders were not IR for GABA or nNOS, but responders did include some GABA positive neurons (G).

126

2+ 3.3 Half the GABA-induced [Ca ]i transients are due to synaptic activation

Acetylcholine activating nicotinic receptors is the primary means of neurotransmission in the ENS, thus hexamethonium (nicotinic antagonist, 200 µM) is commonly used as an agent to block most synaptic transmission in the gut. I examined the effects of hexamethonium on 2+ GABA-evoked [Ca ]i responses which revealed two groups of neurons. Hexamethonium blocked responses to GABA application in 15/30 neurons (at the time control point: 3/33 stopped responding, P = 0.0006, Fisher’s exact test), (Table 5.3) but increased the amplitude 2+ of hexamethonium- resistant GABA-induced [Ca ]i responses (%∆Fi/F0 control: 48 ± 6 %, n

= 30; %∆Fi/F0 hex: 102 ± 34% n = 15, P = 0.029). Ten initially responsive neurons were identified via immunohistochemistry after recordings were complete: 8/10 were GABA- /nNOS- and 2 were GABA-IR. 3/10 neurons that were GABA-/nNOS- stopped responding after hexamethonium; responses in both GABA-IR neurons were blocked by hexamethonium.

3.4 GABA-evoked effects are mediated through GABA A, B and C receptors

At first, ganglia were stimulated chemically via GABA spritz (1 mM) to produce time control data (Table 5.3) enabling appropriate comparisons for subsequent antagonist experiments. During antagonist experiments, bicuculline, CGP54626 and TPMPA were washed in to examine the contributions of GABA receptor subtypes on exogenous GABA-evoked effects.

2+ The GABAA receptor antagonist, bicuculline (10 µM) abolished [Ca ]i responses in 22/35 neurons (at the time control point : 3/33 stopped responding, P=0.0001, Fisher’s exact test,

Table 5.3, Figure 5.4 A, B). Similarly, the GABAC receptor antagonist, TPMPA (100 µM) blocked a significant number of responses (14/28 neurons, P = 0.0005, Fisher’s exact test, Table 5.3, Figure 5.4 G, H). A combination of bicuculline (10 µM) and TPMPA (100 µM) (examined in 2 ganglia, from 2 animals) abolished responses in a significant number of neurons (16/22 neurons, P = 0.0001, Fisher’s exact test), but this proportion did not differ significantly from the proportion blocked by either antagonist alone (combination vs bicuculline alone, P > 0.8; combination vs TPMPA alone, P > 0.5, Fisher’s exact test, Table 5.3).

127

2+ The GABAB antagonist, CGP54626 (1 µM) abolished a significant number of [Ca ]i responses to GABA relative to time controls (7/20 neurons abolished, P = 0.03, Fisher’s exact test) (Table 5.3, Figure 5.4 E, F).

Neurons that still responded in the presence of bicuculline displayed an increase in the 2 amplitude of GABA-induced [Ca +]i responses (%∆Fi/F0 control: 48 ± 6 %, n = 30; %∆Fi/F0 bic: 95 ± 21% n = 13, P = 0.0037) (Figure 5.4 C, D, I). Similarly, TPMPA application increased the amplitude of residual responses (%∆Fi/F0 control: 48 ± 6 %, n = 30; % ΔFi/F0 TPMPA: 109 ± 39% n = 14, P = 0.0255) (Figure 5.4 G, H, I). However in contrast to the 2 effects of bicuculline and TPMPA individually, the amplitudes of residual [Ca +]i responses did not increase significantly in the presence of the combined antagonists (%∆Fi/F0 control:

48 ± 6 %, n = 30; % ΔFi/F0 bic & TPMPA: 75 ± 38% n = 6, P > 0.1). Like the other receptor 2 antagonists, the amplitudes of [Ca +]i responses in the responding neurons with GABAB receptors blocked by CGP54626 were significantly larger than control responses (%∆Fi/F0 control: 48 ± 6 %, n = 30; % ΔFi/F0 CGP54626: 133 ± 25% n = 13, P < 0.0001) (Figure 5.4 E, F, I).

128

Table 5. 3 Number of neurons responding to GABA in control conditions and in the presence of drug antagonists.

Time Point 1 2

Time controls 33 30

Antagonists Control In the presence of drug/s

Hexamethonium (nAChR antagonist) 30 15^

Bicuculline (GABAA antagonist) 35 13^

TPMPA (GABAC antagonist) 28 14^

CGP54626 (GABAB antagonist) 20 13*

Bicuculline + TPMPA 22 6^

*P < 0.05, ^P < 0.001 Fishers exact test

129

Figure legend opposite 130

2+ Figure 5. 4 GABA (1mM) - evoked [Ca ]i transient responses in the presence of antagonists. 2+ Fluorescence images of neurons responding to GABA (1 mM) spritz with [Ca ]i transients with their corresponding traces (shown in ms, scale x103) under control conditions (A-G) and after antagonists are washed in (A’-G’). Black arrows outlined with white on fluorescence images highlight neurons that stopped responding following receptor blockade, coloured arrows highlight neurons that displayed a potentiation in their response; each are colour- matched to corresponding traces. Following bicuculline (10 µm) wash in, some neuronal responses were abolished (black trace) (A’-B’). The small black arrow in (B-B’) indicates the point of GABA application on the trace that corresponds to the highlighted neuron in (A-A’). The amplitudes of neuronal responses that were not abolished by bicuculline, were potentiated as (blue trace) (C’-D’) (P < 0.01). CGP54626 (1 µm) abolished some neuronal 2+ [Ca ]i responses (black trace) and amplitudes of neuronal responses that were not abolished by CGP54626 were increased (green trace), (E’-F’) (P < 0.0001). TPMPA (100 µM) wash in blocked some neuronal responses (black trace) and increased the amplitude of residual 2+ responses (red trace) (G’-H’) (P < 0.05). Scale bars = 20 µm. The amplitude of [Ca ]i responses following the various GABA receptor antagonist wash in, excluding abolished neurons, is depicted graphically (I). Number of neurons examined are displayed within each histogram.

131

3.5 Involvement of endogenous GABA in synaptic transmission

Initially, ganglia were stimulated electrically to collect time control data for subsequent drug antagonist experiments. Ganglia were stimulated using both single pulse and train stimulation which mimic fast and slow synaptic transmission in the ENS respectively (Foong et al., 2012). Trains of 20 pulses (20 Hz) were used as this stimulus regime has been shown to produce sEPSPs in myenteric neurons in the murine small intestine and colon (Mao et al., 2006; Nurgali et al., 2004). The effects of bicuculline, TPMPA and CGP54626 were then 2+ tested on electrically-induced [Ca ]i transient responses. This would be expected to reveal any involvement of endogenous GABA released from myenteric varicosities.

Bicuculline did not have any significant effect on the number of neurons responding to electrical stimulation relative to time controls (time control single pulse: 10/42 did not respond; bicuculline single pulse: 12/49 did not respond, P > 0.9; time control 20 pulse: 2/58 did not respond; bicuculline 20 pulse: 7/79 did not respond P > 0.3, Fisher’s exact test) (Table 5.4). Similarly CGP54626 had no effect on the number of responders (CGP54626 single pulse: 6/38 did not respond P > 0.4; CGP54626 20 pulse: 4/79 did not respond P > 0.9, Fisher’s exact test) nor did TPMPA (TPMPA single pulse: 8/41 did not respond, P > 0.8; TPMPA 20 pulse: 8/60 did not respond P > 0.1, Fisher’s exact test, Table 5.4).

After bicuculline application, neurons did not display a change in the amplitude of 2 electrically-induced [Ca +]i responses following either single pulse stimulation (single pulse

%ΔFi/F0 control: 106 ± 14% n = 32; % ΔFi/F0 bic: 153 ± 29% n = 37, P > 0.1) or train stimulation (20 pulse % ΔFi/F0 control: 116 ± 10% n = 56; % ΔFi/F0 bic: 122 ± 14% n = 72, P > 0.7, Figure 5.5 A, B, G). Similarly CGP54626 did not alter the amplitude of electrically- 2 induced [Ca +]i responses (single pulse % ΔFi/F0 CGP: 110 ± 20% n = 32 P > 0.8; 20 pulse %

ΔFi/F0 CGP: 106 ± 6% n = 75, P > 0.3, Figure 5.5 C, D, G). The amplitude of single pulse- 2+ evoked [Ca ]i responses was unchanged in the presence of TPMPA (%ΔFi/F0 TPMPA: 139 2+ ± 29% n = 33, P>0.4). However, train-evoked [Ca ]i transient responses were potentiated in amplitude by TPMPA (%ΔFi/F0 TPMPA: 237 ± 53% n = 52, P = 0.02, Figure 5.5 E, F,G).

132

Table 5. 4 Number of neurons responding to electrical stimulation in control conditions and in the presence of antagonists.

Time Point 1 2

1 pulse stimulation time controls 42 32

20 pulse train stimulation time controls 58 56

Antagonists Control In the presence of drug/s

Bicuculline (GABAA, 10 µM ) 1 pulse 49 37 stimulation

Bicuculline (GABAA, 10 µM ) 20 pulse 79 72 stimulation

TPMPA (GABAC, 100 µM ) 1 pulse stimulation 41 33

TPMPA (GABAC, 100 µM ) 20 pulse stimulation 60 52

CGP54626 (GABAB, 1 µM) 1 pulse stimulation 38 32

CGP54626 (GABAB, 1 µM) 20 pulse stimulation 79 75

133

Figure legend opposite

134

2+ Figure 5. 5 Electrically - evoked [Ca ]i transient responses in the presence of antagonists. 2+ Fluorescence images of neurons responding to 20 pulse electrical stimulation with [Ca ]i transients with their corresponding traces (shown in ms, scale x103) under control conditions (A-F) and after antagonists are washed in (A’-F’). Arrows highlighting neurons on fluorescence images are colour-matched to corresponding traces.

2+ Both single pulse and 20 pulse (shown) stimulation produced [Ca ]i responses that were no different in amplitude following bath application of bicuculline (10 µm) (blue trace) (A-A’, B-B’), the black arrow indicates point of electrical stimulation (B-B’). Similarly, CGP54626 2+ (1 µm) did not alter the amplitude of train-evoked [Ca ]i responses (green trace) (C-D, C’- 2+ D’). Trains of electrical stimulation produced [Ca ]i responses that were potentiated in amplitude by bath application of TPMPA (100 µM) (red trace) (E-E’, F-F’) (P < 0.05) . Scale 2+ bars 20 = µm. The amplitude of single pulse and 20 pulse-evoked [Ca ]i responses following GABA antagonist wash in is depicted graphically (G). Number of neurons examined are displayed within each histogram.

135

4. DISCUSSION

Recent studies reveal important roles of GABA particularly as metabolic products of microbiota that are intensified in pathogenic conditions (Barrett et al., 2012; Dann et al., 2015). However the physiological roles of GABA and its receptors remain elusive as their expression appears to be extremely specific in the gut and its ENS. In this study, I used several converging methods to examine the role of GABA and its receptors in the murine myenteric plexus. I demonstrated the expression of GABA in neurons and varicosities surrounding specific enteric neurons and identified enteric neuronal subtypes that express GABA receptors using specific antagonists and Ca2+ imaging. Furthermore, I found that GABA is unlikely to mediate fast or intermediate EPSPs, via any of the GABA receptor subtypes. However I revealed that endogenous GABA is released and may modulate slow excitatory neurotransmission in the ENS via GABAC receptors.

4.1 GABA colocalises with some calretinin and nNOS neurons in the MP of the mouse ileum

It has previously been reported that in the murine small intestine GABA-IR neurons are rare and thus a profile of GABA colocalisation with other neuronal markers has not previously been investigated in mouse ileum (Sang and Young, 1996; Li et al., 2011). Further, prior to this study, though GABA-IR varicosities in the mouse small intestine have been identified, their innervation patterns within the myenteric plexus have not been examined in depth (Sang and Young, 1996).

In accord with previous reports (Sang and Young, 1996; Li et al., 2011), I found that GABA neurons constitute 8% of neurons in the myenteric plexus of the murine ileum. Further, I found that GABA colocalised with nNOS-IR neurons which are often inhibitory motor neurons, and calr-IR neurons, which include intrinsic sensory, excitatory interneurons and motor neurons in the murine small intestine, therefore GABA-IR neurons may constitute subpopulations of such neurons. Previous reports in the mouse colon and guinea pig small intestine are consistent with this suggestion (Williamson et al., 1996; Sang and Young, 1996). In my study, GABA-IR varicosities were found in the circular muscle and within the ganglia, closely apposing approximately 60% of calr-IR neurons and 70 % of nNOS-IR neurons.

136

Thus, some GABA-IR neurons are likely to be interneurons contacting excitatory and inhibitory motor neurons as reported previously for the rat colon (Krantis, 2000). This is also consistent with findings from the murine colon, where widespread GABAergic innervation of neurons throughout the MP has been identified by immunolabelling of the vesicular GABA transporter (VGAT) (Seifi et al., 2014).

4.2 GABA activates myenteric neurons in the mouse ileum

By exogenous application of GABA, I demonstrated that myenteric neurons in the mouse 2+ ileum respond to GABA with [Ca ]i transients, the majority of these being GABA-/NOS- neurons. These also include some calr-IR and smaller numbers of GABA-IR neurons (although these represent most GABA-IR neurons within each ganglion), and to a much lesser extent, NOS-IR neurons.

Previous studies using intracellular recording have shown that GABA excites some myenteric neurons in guinea pig ileum (Cherubini and North, 1984a). This depolarisation was mediated by GABAA receptors and only observed in AH neurons, while S neurons were insensitive to GABA. Since calr-IR neurons responded to GABA in my study and calr is a neuronal marker of AH neurons/intrinsic sensory neurons (ISNs) in the murine small intestine, my results are congruent with the previous findings of Cherubini and North (1984a). Yet calretinin is also present in excitatory interneurons and motorneurons, while some nNOS-IR neurons and GABA-IR neurons responded. These neuronal subtypes are likely to have S-type electrophysiology, thus the inclusion of neurons other than ISNs differs from the early findings in the guinea-pig ileum (Cherubini & North, 1984a). This might well be due to differential GABA receptor expression, which has not been detailed across species.

GABA-IR innervation of neurons within the MP of the mouse and rat colon (Sang and Young, 1996; Krantis, 2000; Seifi et al., 2014) and to a much lesser extent guinea pig ileum (Jessen et al., 1986; Furness et al., 1989) suggests a role for GABAergic transmission in the MP and that GABA-IR neurons may function as interneurons. Indeed GABA-evoked Ca2+ responses in half of the GABA- responding neurons were indirect, and found to be mediated by nicotinic receptors, indicating that these neurons respond via excitatory synaptic inputs from other neurons that express GABA receptors within the circuitry.

137

Further, I found that GABA varicosities make close contacts with many calr-IR and nNOS-IR neurons and hence are likely to be from interneurons innervating these motor neurons. Unsurprisingly calr-IR neurons responded to GABA. Interestingly, nNOS neurons rarely responded despite their presumptive GABAergic innervation. It cannot therefore be concluded that GABAergic varicosities demonstrated in this study form functional contacts or that nNOS neurons express GABA receptors as they have been shown to in the mouse colon (Seifi et al., 2014). This requires further detailed analysis.

4.3 GABAA GABAB and GABAC receptors are expressed in the myenteric circuitry

After demonstrating that GABA activates neurons in the MP of the mouse small intestine, I extended this by identifying GABA receptor subtypes that mediated these excitatory responses. While the expression of GABA receptors has been established in the rodent GIT via immunohistochemical and functional analyses, most information details receptor expression in the rat where GABAA (Krantis and Harding, 1987; Krantis et al., 1995; Poulter et al., 1999), GABAB (Krantis and Harding, 1987; Nakajima et al., 1996), and GABAC (Fletcher et al., 2001) receptors are known to exist.

GABAA and GABAB receptors have been identified in the mouse colon with GABAA receptors expressed on neurochemically diverse cell types including nNOS and ChAT, 5-HT and SOM neurons (Casanova et al., 2009; Seifi et al., 2014; Auteri et al., 2015). Only very limited functional work confirming the expression of GABAC receptors in the mouse model exists. This found that contractility in the longitudinal muscle of the small intestine is modulated by GABAC receptors (Zizzo et al., 2007). In order to deduce GABA receptor expression on enteric neurons in the small intestine, I attempted to identify locations of different receptor subtypes immunohistochemically, but was unable to obtain satisfactory labelling with the available antisera. Most localisation of the various GABA receptor types has been shown in the rat. GABAC receptor expression has only been demonstrated in this species and availability of antisera for GABAC receptors is limited. For these reasons this component of the study was not further pursued. Instead I used specific receptor antagonists to examine the effects of all 3 GABA receptor subtypes.

The GABAA antagonist bicuculline, the GABAB antagonist CGP54626 and the GABAC receptor antagonist TPMPA each prevented calcium responses in many neurons that

138

responded to GABA. These data suggest that a subset of myenteric neurons including nearly all GABA-IR neurons possess GABAA GABAB and GABAC receptors with the possibility of some possessing more than one subtype. A response to GABA that was not blocked by either

GABAA or GABAC antagonists or both was identified, potentially mediated via GABAB receptors.

Interestingly, responses that were not abolished by GABA receptor antagonists and the nicotinic receptor antagonist hexamethonium were significantly larger than control responses in the same neurons, which reveals a potential inhibitory action of GABA.

GABA is generally characterised as an inhibitory neurotransmitter in the CNS that acts - through ionotropic pentameric Cl channels (GABAA and GABAC) and G-protein-coupled receptors (GABAB) (Enz, 2001; Sigel and Steinmann, 2012). However in the ENS and in sub-compartments of the CNS (Chavas and Marty, 2003; Gulledge and Stuart, 2003), including the retina (Fletcher et al., 1998; Varela et al., 2005), GABA-mediated chloride conductance increases via GABAA and GABAC receptors is known to produce depolarisations due to elevated intracellular chloride concentrations (Russell, 2000; Liu et al., 2013).

It is unclear whether the inhibitory action of GABA observed in this study is due to conductance changes or dis-inhibition of unknown inputs within the circuitry. For instance if

GABAA and GABAC receptors are expressed on neurons in a negative feedback arrangement, when either GABA receptor is antagonised the synaptically connected neuron would be disinhibited and exhibit a potentiated response (Figure 5.6 A). Similarly, if inhibitory inputs to GABA-responding neurons from neurons expressing nicotinic receptors are blocked, a disinhibition and amplification of the response in GABA responding neurons could occur (Figure 5.6 B). Possible sources of inhibition within the circuitry could include somatostatin

(SOM) – expressing neurons, which have been presumed to signal via GABAA receptors in the mouse colon while GABAA receptors are also evident on nNOS and 5-HT neurons in the colon (Seifi et al., 2014). To date however, a conclusive inhibitory transmitter in the MP of the murine ileum has not been identified and this should be a point of further examination.

In addition, since the exact immunocolocalisation pattern for GABA receptors in the mouse ileum is unknown, it is possible that the ionotropic GABA receptors are expressed presynaptically mediating the inhibitory effects observed. Indeed, while GABAA receptors in

139

the murine gut are commonly expressed postsynaptically in the MP, in the SMP of the colon

GABAA receptors are expressed presynaptically on axonal compartments of nNOS neurons

(Seifi et al., 2014). Further, GABAB receptors in the mouse act at the presynapse to prevent ACh release from excitatory cholinergic neurons (Sanger et al., 2002; Hyland and Cryan,

2010) and may account for the increase in neuronal activity observed following GABAB receptor blockade in this study. Indeed, while GABAB receptors are generally associated with presynaptic inhibition of ACh release, inhibition of inhibitory neuromuscular transmitter release such as NO, by GABAB activation has been reported in the rabbit, rat and guinea pig (Tonini et al., 1989a; Kilbinger et al., 1999; Bayer et al., 2003). It may be that in this system

GABAB receptors are located on inhibitory inputs to unidentified neurons, so that when they are blocked, inhibition is restored and a number of GABA responses are abolished.

Figure 5. 6 Schematic of a postulated circuit showing dis-inhibition of unknown inputs.

This schematic shows that some myenteric neurons express GABAA and GABAC receptors since the antagonists bicuculline (10 µM) and TPMPA (100 µM) prevented calcium responses in a significant proportion of neurons that initially responded to GABA. Neuronal responses that were not abolished by either antagonist nor by hexamethonium were amplified, suggesting that GABA has both an excitatory and inhibitory influence over enteric neurons, but the exact source of inhibition is unknown. This may occur if both receptor subtypes are expressed on neurons arranged in a negative feedback loop where blocking one receptor dis-inhibits the synaptically-connected neuron thus amplifying its response (A). Additionally GABA-responding neurons may receive inhibitory inputs from neurons expressing nicotinic receptors, which if blocked by hexamethonium (200 µM), potentiate the activity of the GABA responder (B). 140

4.4 The GABAC receptor has a synaptic function and an inhibitory role in neural circuits

Electrical stimulation which evokes endogenous neurotransmitter release was used in this study to ascertain whether any GABA receptors and therefore endogenous GABA contribute to synaptic transmission. A single pulse electrical stimulus applied to interganglionic connectives is typically used to evoke fast to intermediate synaptic responses (Hirst et al., 1974; Bornstein et al., 1986; Monro et al., 2004; Gwynne and Bornstein, 2007b; Foong et al., 2012), while trains of electrical stimuli evoke slow EPSPs (Wood and Mayer, 1978; Johnson et al., 1981; Bornstein et al., 1994; Gwynne and Bornstein, 2007b). In this study, single 2+ pulse-evoked [Ca ]i responses were not affected by antagonism of any of the GABA receptors and trains of stimuli produced calcium responses that were potentiated by antagonising GABAC receptors but were unaffected by GABAA or GABAB blockade.

Further, the GABAC receptor antagonist TPMPA shows strong selectivity for this receptor subtype; it has >100-fold selectivity for GABAC receptors compared with GABAA or GABAB receptors (Ragozzino et al., 1996; Johnston, 2002).

This indicates that release of endogenous GABA modulates slow synaptic transmission via

GABAC receptors, where GABA acting at GABAC receptors tonically inhibits the system. It is intriguing that GABAA or GABAB receptors do not contribute to or modulate either fast or slow synaptic transmission, since exogenous application of GABA showed that each receptor subtype is present and functional studies demonstrate that GABAA and GABAB receptors have a significant role in the control of motility responses in the mouse (Sanger et al., 2002;

Zizzo et al., 2007; Auteri et al., 2014). One study has reported the involvement of GABAC receptors in the control of gut smooth muscle contractility (Zizzo et al., 2007); indeed

GABAC receptors were shown to modulate slow transmission in this study, but do not appear to be involved in fast transmission. It might be that endogenous GABA does not reach relevant GABAA and GABAB receptors, or that different stimulus regimes are required and 2+ so GABAA and GABAB-mediated synaptic transmission cannot be excluded. While Ca 2+ imaging is advantageous for a broad assessment of neural activity, [Ca ]i is not the only measure of neuronal action. Detailed membrane properties such as synaptic events are well studied via electrophysiological techniques such as intracellular recording or whole-cell recording and may be more accurately measured using such methods.

141

While GABAC receptors mediate excitation via a depolarising conductance, intrinsic differences in receptor kinetics may account for the antagonist- revealed inhibition of responses to stimulus trains. GABAA receptors mediate rapid transient currents, while

GABAC receptors have been demonstrated to produce more prolonged responses (Qian and Dowling, 1995; Lukasiewicz and Shields, 1998) where the deactivation rate or closing phase of GABAC receptors is much slower (Amin and Weiss, 1994). In addition, unlike GABAA receptors GABAC receptors display little desensitisation (Matthews et al., 1994).

Consequently, GABAC receptors such as those located on the terminals of retinal bipolar cells are known to produce a (tonic) standing current due to these properties (Hull et al., 2006; Palmer, 2006). It might well be that prolonged opening of chloride channels by endogenous GABA prevents firing by a form of shunting inhibition due to extended chloride conductance and interactions of Na+ inactivation and voltage-gated K+ channel (Kv) opening.

Slow excitatory transmission in enteric neurons is associated with a net decrease in membrane conductance due largely to an inhibition of two potassium conductances; resting + + or ‘leak’ K conductance (gK) and calcium-dependent K conductance (gKCa), which is the conductance mediating the after-hyperpolarisation (AHP) in AH neurons (Grafe et al., 1980). While this applies to AH neurons, the situation involving S neurons is more complex as the K+ channels have not been clearly identified and there is potential for contribution by M- current channels like the KCNQ (Kv7) channels (Brown and Passmore, 2009; Hirst et al., 2015). In addition to suppression of K+ conductances, slow transmission has been shown to be due to an increase in chloride conductance (gCl) in myenteric neurons (Bertrand and Galligan, 1994). The reversal potential for Cl- in guinea pig myenteric neurons is estimated to be between -39 mV (Cherubini and North, 1984a; Bertrand and Galligan, 1994) and -17 mV (Bertrand and Galligan, 1994). Therefore inactivation of resting gK and activation of gCl leads to depolarisation of the neuron relative to the resting potential and similar to the reversal potential of Cl.

Consequently activation and prolonged opening of GABAC receptors could produce a prolonged chloride conductance approximating this value, establishing a new membrane potential. The initial depolarisation of the cell at this membrane potential is likely to activate delayed rectifier K+ channels which are active when the membrane is depolarised to values above -30 mV (Hirst et al., 1985a), producing a repolarising inhibition. This stabilization of the membrane would prevent the activation of voltage gated Na+ channels and further 142

depolarisation therefore allowing GABA to shunt firing by a ‘clamping’ of the membrane in this way. Thus following repetitive stimulation of interganglionic connectives, prolonged opening of these channels by endogenous GABA may be responsible for this shunting inhibition. This would be released when these receptors are blocked, producing the potentiated response observed in this study.

The advantage of broadly surveying the ENS via Ca2+ imaging enabled us to observe the widespread endogenous effects of GABA within the myenteric circuitry for the first time.

The specific actions of GABAC receptors likely require a more detailed analysis of the membrane potential, which would be ideally investigated via intracellular recording. However recording via sharp electrode from a sufficient sample size of neurons, within a reasonable time frame, is outside the scope of this study. Instead we have commenced a preliminary investigation using computational modelling. Previously, modelling conducted by colleagues has generated realistic predictions of neuronal behaviour with a capacity to reproduce experimental observations and has been implemented to understand the influence ionic currents have on the excitability of neurons (Chambers et al., 2014). In this study, the current model incorporates predictions from our previous model (Chambers et al., 2014) and was implemented to examine the effect GABA channels have on the excitability of enteric neurons. The model includes voltage-gated channels which are known to be found on myenteric neurons including Nav1.3 and Nav1.7, responsible for rising phase of action potentials (Sage et al., 2007) and K+ channels including A-type K+ and delayed-rectifier channels, which contribute to the falling phase (Furness et al., 1998; Chambers et al., 2014). Alterations to this model were made to include GABAA and GABAC receptors and a stereotyped synaptic current with a reversal potential of +10 mV and described by an alpha function. This last component served as a model for synaptic potentials mediated by nicotinic acetylcholine receptors (nAChRs) as such responses are the major form of synaptic transmission within ENS (Kirchgessner and Liu, 1998; Zhou and Galligan, 2000; Gwynne and Bornstein, 2007b; Foong et al., 2015). In addition Kv7.2 channels were incorporated into the model. Kv7.2 is known to have a suppressive effect on neuronal firing in the CNS and peripheral neurons (Brown and Passmore, 2009; Cooper, 2011) where the K+ current converts tonic firing to phasic firing. Kv7.2 are expressed by murine enteric neurons (Hirst et al., 2015).

143

This modelling is consistent with the experimental results reported in this study. Following repetitive (train) neural stimulation and activation of synaptic inputs including GABAC synapses, an initial AP is observed in the modelled neuron followed by a prolonged clamping of the membrane potential. In other words, GABAC activation ‘shunts’ firing and prevents further firing. When GABAC synapses are blocked, the inhibitory shunting action is released and firing is increased (Figure 5.7). While this work is preliminary and the model does not include slow EPSPs, it is consistent with the novel experimental finding and paves the way for further investigation into the role of GABAC receptors within the enteric circuitry. Studies should extend to further detailed membrane analyses and the functional effects of such synaptic modulation.

144

Figure 5. 7 Preliminary computer modelling to interpret the possible inhibitory role of GABA in modulating synaptic transmission. Computer modelling was conducted to test the hypothesis that prolonged opening of a Cl- channel via GABAC receptors can prevent firing. A) Here the membrane potential of a neuron is shown. The neuron has been stimulated with a train of the GABAC synapses in the presence of alpha (nicotinic) and GABAA synapses, where the reversal potential of the model of the GABAC synapses is -35mv. The neuron fires initially where GABAC channels remain open due to slow-gating kinetics to produce a prolonged Cl- conductance. The activation of potassium channels including Kv7.2 channels repolarise the membrane potential and shunt further firing. B) When GABAC receptors are inhibited but other synaptic inputs including 20

GABAA synapses are active, firing is increased since the GABAC -mediated inhibition is removed.

145

4.5 Future directions and conclusions

In all, this study extended the profile of GABA within the murine MP in the small intestine where GABA-IR neurons are likely to function as interneurons as they do in other species. It was demonstrated that GABAA, GABAB and GABAC receptors are expressed by enteric neurons where GABA was shown to either excite or inhibit neurons via these receptor subtypes. Release of endogenous GABA was found to modulate synaptic responses to trains of stimuli via GABAC, receptors; this finding was consistent with preliminary computer modelling and is an exciting research point that should be extended further by detailed membrane property analysis and examining the potential functional effects GABAC-mediated transmission may have. If or how this contributes to pathogenic GIT states such CDI, might well pave the way for the identification of exciting therapeutic targets.

146

CHAPTER 6: A MORPHOLOGICAL AND ELECTROPHYSIOLOGICAL STUDY OF MYENTERIC NEURONS IN THE MOUSE ILEUM – EFFECTS OF CT ON THEIR EXCITABILITY

ABSTRACT

In Chapter 3, in vitro incubations of CT were implemented in the guinea pig to examine effects on neural circuitry. In this present study mice were treated with CT in vivo to make use of an incubation method that more closely resembles the physiological internal environment. The properties of myenteric neurons were examined in detail via intracellular recording. This was conducted alongside a parallel study that used Ca2+ imaging to examine the broader effects of CT on the entire circuitry. CT induced spontaneous synaptic activity in some myenteric neurons, suggesting heightened input elsewhere in the circuitry. This source was not identified due to low proportions of sensory neurons in the mouse compared to the commonly studied guinea pig model, and a prominent sampling bias. The data suggest a minor role for myenteric neurons in the contribution to CT-induced hypersecretion.

1. INTRODUCTION

CT produces a massive efflux of water and electrolytes across the intestinal mucosa when incubated in vivo in the small intestine of rodents via a ligated ileal loop (Burrows and Musteikis, 1966; Castagliuolo et al., 1994; Guichard et al., 2013). Chapter 3 explored the effects of in vitro incubations with CT on enteric neuronal firing properties. While in vitro incubations have been used to model CT-induced hypersecretion (Carey and Cooke, 1986; Gwynne et al., 2009), in vivo exposure to enterotoxins prevents damage to the mucosal barrier due to lack of blood supply and thus preserves normal mucosal function, more closely replicating the internal environment. For these reasons in vivo incubations of CT have been used in this study where the ileal loop mouse model has been implemented to take advantage of Wnt1-Cre;R26R-GCaMP3 mice for use in Ca2+ imaging in a joint study.

147

I collaborated with a colleague in a study where she applied Ca2+ imaging to show that in vivo incubation of CT in the mouse ileum increased the number of spontaneously active neurons in the submucosal plexus (SMP), and enhanced their responses to electrical stimulation. Such increases in spontaneous or electrically-evoked activity were not observed myenteric neurons, indeed their spontaneous activity was depressed (Fung, Koussoulas et al., 2017, unpublished). This is surprising, as it is proposed in rats that the MP is required for CT to exert full hypersecretory effects (Jodal et al., 1993) and in Chapters 3 and 4, I demonstrated that in vitro incubations of bacterial toxins including CT in the guinea pig jejunum induced sustained hyperexcitability in myenteric sensory neurons. My colleague also examined the effects of CT on the enteric circuitry by immunolabelling for activity- dependent markers pCREB and c-Fos (Sheng and Greenberg, 1990). In the SMP, neuronal expression of pCREB was significantly higher in CT-treated tissues than in controls, while in the MP CT increased neuronal c-Fos expression compared to controls (Fung, Koussoulas et al., 2017, unpublished). However, despite the high-throughput advantage of Ca2+-imaging, measurement of changes in intracellular calcium is only one indication of neuronal excitability. Of note, the firing and membrane properties of individual neurons cannot be examined with Ca2+ imaging techniques, and these properties have not previously been examined following in vivo incubations of CT in the murine small intestine. In fact, very few studies have examined electrophysiological properties of neurons in the mouse small intestine (Bian et al., 2003; Ren et al., 2003; Mao et al., 2006; Foong et al., 2012). Further to this, an analysis of correlations between electrophysiological and morphological properties of enteric neurons in the mouse ileum has not been performed previously.

In this study, I used intracellular recording to examine the membrane and action potential firing properties of myenteric neurons following CT- treatment in the in vivo mouse ileal loop model to extend Ca2+ imaging observations (Fung, Koussoulas et al., 2017, unpublished), and to determine whether CT induced sustained excitability effects on myenteric neurons. Using this method, I revealed heightened synaptic activity in the MP in the form of an increase in spontaneous synaptic potentials in some neurons. This is also the first study correlating electrophysiological and morphological properties of myenteric neurons in the mouse ileum.

148

2. METHODS

2.1 Mice

Experiments were performed using male mice on a C57BL/6 background, including Wnt1- Cre;R26R-GCaMP3 mice that express the fluorescent calcium indicator GCaMP3 in all enteric neurons and glia, aged 8-12 weeks (see Chapter 2).

2.2 Ileal loop surgery

Ileal loop surgery was performed on mice as described in Chapter 2 (section 2). Briefly, either physiological saline or toxin in saline (CT: 12.5 µg/ml, List Biologicals, Campbell, CA) was injected into the ileum filling the region (100-150 µl of incubating solution). After a 3.5 hour incubation period, the animal was killed by cervical dislocation as approved by the University of Melbourne Animal Experimentation Ethics Committee. The ileal loop segment was dissected from the abdomen for in vitro experimentation.

2.3 Tissue preparation for electrophysiology

Following surgery, ileal loop tissue was dissected out of the animal and placed in physiological saline containing nicardipine (1.25 μM) and hyoscine (1 μM) to prevent smooth muscle contractions. A preparation of MP with attached longitudinal muscle (LMMP) was produced via microdissection (Chapter 2, section 1.1).

The LMMP preparation was transferred to a recording bath continually superfused with 35-

36°C physiological saline bubbled with 95%O2/5%CO2. The preparation was allowed to equilibrate for an hour before commencing electrophysiological experiments.

2.4 Intracellular recording

Neurons were impaled (as outlined in Chapter 2, section 1.2) with glass intracellular recording microelectrodes (95-200 MΩ tip resistance) containing 1 mol/L KCl, with or without 2% biocytin (Sigma Aldrich, Castle Hill, NSW, Australia) to allow them to be identified morphologically after electrophysiological recordings (see Chapter 2 section 1.3).

149

Excitability and synaptic inputs of neurons were determined as previously described in Chapter 2, section 1.2.

2.5 Analysis and statistics

The electrophysiological data were analysed using AxoScope computer software (version 10.2.0.14, Axon Instruments U.S.A.). Injection of depolarising and hyperpolarising current pulses into neurons was performed twice per cell, while the single pulse stimulus regime was repeated 3 times to obtain averages for these parameters. Data are presented as mean ± SEM and n = the number of cells examined.

Statistical analyses were performed using unpaired t-tests with P < 0.05 considered statistically significant. Comparisons were performed using GraphPad Prism 5.0 (GraphPad Softwares, San Diego California).

2.6 Drugs

Drugs used were hyoscine and nicardipine (both from Sigma Aldrich, Castle Hill NSW, AUS). CT was dissolved in distilled water to make stock solutions stored at 4°C and then again in physiological saline to their working concentration on the day of experimentation.

3. RESULTS

3.1 CT in the ileum in vivo produces enhanced secretion

I performed in vivo ileal loop incubations with CT in mice and found enhanced secretion in all preparations as described in our parallel study (Fung, Koussoulas et al., unpublished). Increased fluid accumulation was observed in segments of ileum that were injected with CT (12.5µg/mL) compared to saline controls after 3.5 hours (Figure 6.1).

150

Figure 6.1 Control versus CT -treated ileal loops following 3.5 h incubation in vivo.

Images of ileal loops (marked with yellow asterisks) following 3.5 hour incubation period in vivo with either saline (A) or CT (12.5 µg/mL) (B) injected into the lumen.

3.2 Electrophysiology and morphology of myenteric neurons in the mouse ileum

Experiments were technically difficult and impalements from murine tissue presented a great challenge when compared to guinea pig equivalents. This is likely why very few studies have ever examined the electrophysiological properties of neurons in the mouse small intestine and is the primary reason that neuronal sample sizes were small in this study.

A total of 45 myenteric neurons were examined electrophysiologically, 38 had S-type electrophysiology (20 control, 18 CT) and 7 had AH-type electrophysiology (3 control, 4 CT). Myenteric AH neurons in the murine ileum were far under-represented compared to the guinea pig ileum where AH neurons are much more common and are recorded from in similar proportions to S neurons (Iyer et al., 1988). This is consistent with the work of Bian et al. (2003) and Ren et al. (2003) who found that AH neurons in the mouse ileum are impaled infrequently and represent a small percentage (10%) of total neurons recorded.

151

All AH neurons whose firing properties were examined, displayed phasic firing (firing rapidly at the onset of the depolarisation, then accommodating to the stimulus), regardless of treatment (n = 5), a common feature of AH neuronal firing (Bornstein et al., 1994) (Figure 6.2). Instantaneous firing frequency was measured as the interspike interval (ISI) between the first two spikes at the 300 pA current step. In control neurons, this was found to be 35 ± 10 ms, hence control neurons displayed a maximum firing frequency of 29 Hz (n = 3).

There is a large variety of S-type neurons and they can fire phasically or throughout a depolarising pulse over a range of frequencies (Bornstein et al., 1994). The firing of S-type neurons was characterised for 1 CT-treated neuron and 3 controls. One control S neuron fired tonically (firing throughout the depolarisation) at maximum pulse amplitude (350 pA), while 2 fired a single AP with maximum pulse amplitude (Figure 6.2). The tonically-firing neuron did so at a frequency of 21 Hz (ISI 47 ms).

All AH neurons displayed a characteristic prolonged AHP following AP firing during depolarising current pulses and exhibited an Ih-induced ‘sag’ in the membrane potential during hyperpolarising current pulses (Figure 6.2) (Galligan et al., 1990; Bornstein et al., 1994). Under control conditions, AH neurons had resting membrane potentials (RMPs) of - 59.7 ± 2 mV, n = 3 (Table 6.1). S neurons did not display a prolonged AHP and received substantial fast excitatory input (Figure 6.2), shown by fEPSPs produced by electrical stimulation of their presynaptic inputs, as seen in the guinea pig small intestine (Hirst et al., 1974; Bornstein et al., 1991a; Furness, 2006). S neurons had RMPs of -46.3 ± 4.1 mV, n = 6 (Table 6.1).

The morphologies of 20 myenteric neurons (9 control, 11 CT) were examined after electrophysiological experiments. As observed in the guinea pig small intestine (Bornstein et al., 1984; Bornstein et al., 1991a; Bornstein et al., 1994), all AH neurons (n = 7) possessed Dogiel type II morphology with large smooth cell bodies and multiple axonal processes, (with an average of 3 axons/axonal branches projecting in a circumferential direction) (Figure 6.2 and 6.3).

All S-neurons examined had a single axon, 12 of these neurons displayed lamellar dendrites typical of Dogiel type I morphology (Bornstein et al., 1991a) (Figure 6.2 and 6.3), and 1 152

neuron had filamentous dendrites. Most S neurons (n = 8) projected their axons orally (Figure 6.3), 4 had circumferential projections, including the filamentous neuron, 1 was undetermined. Thus it appears that anally projecting interneurons and motor neurons have not been sampled in this data set. This is a striking point to note since nNOS neurons which would be expected to project anally represent approximately a third of the total neuronal population in the mouse small intestine (Qu et al., 2008). The axon projections of 5 S-type neurons were traced and measured: 2 orally-projecting neurons branched out into the circular muscle, while 1 branched into the longitudinal muscle. Neurons with axons branching and entering the muscle had projection lengths ranging from 250 µm-650 µm. Two S-type neurons had orally-projecting axons that ran for greater distances (1000-1100 µm), but could not be traced to the muscle. These projections were interrupted by the dissection of the tissue and thus were unable to be fully characterized (Figure 6.3).

3.3 CT induces an increase in spontaneous activity in S neurons

As only 1 CT-treated S neuron had its firing properties characterised meaningful conclusions regarding the effects of the toxin on the firing of S neurons cannot be drawn. Further, due to the small proportion of myenteric AH neurons in this study and the random nature of the impalement method, recordings from AH neurons were restricted even though these neurons produced more robust impalements. Therefore with small neuronal sample sizes, statistical comparisons for many of the membrane and firing properties between treatment groups cannot reliably be conducted. Instead for some parameters, values for controls and CT-treated neurons are described here without statistical comparison.

CT treatment did not appear to have a significant effect on the RMP of myenteric neurons for either S-type (control: -46.3 ± 4.1 mV, n = 6, CT: -51.7 ± 4.4 mV, n = 3) or AH neurons (control: -59.7 ± 2 mV, n = 3 CT: -53.3 ± 4.8 mV n = 3) (Table 6.1).

Most of the firing properties of S neurons at RMP were recorded from control neurons including the threshold of firing (S-type; control: 160 ± 10 pA, n = 5), and the maximum number of APs fired during depolarising current pulse steps (S-type; control: 3.5 ± 2.5 APs, n = 3). Following CT-treatment, 1 S neuron fired at a threshold of 150 pA with a maximum of 3 APs, at maximum pulse amplitude (Table 6.1).

153

The threshold of firing for AH neurons for both treatment groups (AH-type control: 183 ± 42 pA, n=3 CT: 100 ± 25 pA, n = 2) and the maximum number of APs fired during depolarising current pulse steps (AH-type; control: 2.5 ± 0, n = 3 CT: 5.3 ± 1.3 APs, n = 2 at 350 pA) were analysed where CT doubled the number of APs fired at maximum pulse amplitude (Table 6.1). ISI in CT-treated AH neurons was 22 ± 0.2 ms (n = 2), indicating that maximum instantaneous firing frequency in CT-treated neurons vs controls was 44 Hz and 29 Hz respectively. All CT-treated AH neurons displayed anode break action potentials (control: 0/2, CT: 3/3). Additionally, while control neurons were quiescent without stimulation, spontaneous firing (proximal process potentials) were seen in 2/4 CT-treated AH neurons; spontaneous firing in AH neurons in the mouse small intestine under control conditions has not previously been reported (Ren et al., 2003; Mao et al., 2006; Foong et al., 2012) (Table 6.1). Thus, a greater sample size may reveal excitatory effects of CT on AH neurons but this was the technically challenging to execute due to time constraints, difficulties in impaling from neurons from the mouse small intestine and the low incidence of AH neurons as previously outlined.

To study synaptic inputs to S neurons, hyperpolarizing current was passed through the recording electrode to bring the membrane potential to -90 mV and hence increase the driving potential for the fast EPSPs (Monro et al., 2004). At this hyperpolarized membrane potential, the amplitudes of electrically-evoked fast EPSPs did not differ significantly between CT (28.6 ± 2.6 mV, n = 12) and controls (22.1 ± 2.9 mV; n = 13, P > 0.2) nor did their duration (Con: 77 ± 10 ms; CT: 104 ± 12 ms, P = 0.09) (Table 6.1, Figure 6.4). However, the proportion of myenteric neurons displaying spontaneous activity (single and bursts of fast EPSPs) increased following CT exposure (CT: 61%, (11) n = 18; control: 15%, (3) n = 20; Fisher’s exact test, P < 0.01; Table 6.1, Figure 6.4), where the durations of recordings that were used to detect spontaneous EPSPs in both control and CT-treated preparations were comparable (Control: 3.7 ± 0.7 mins; CT 3.1 ± 1 mins P > 0.5 ).

154

Figure 6.2 Correlations between electrophysiological and morphological features in myenteric neurons of the mouse ileum. AH neurons fired action potentials phasically during depolarising current pulses followed by a prolonged AHP (as indicated by the arrow head) and exhibited an Ih-induced ‘sag’ in the membrane potential (arrow head) during a 500 ms hyperpolarizing pulse injected at RMP (A). Neurons with AH electrophysiology displayed Dogiel type II morphology. The AH neuron (with corresponding electrophysiology in A) has a smooth cell body with multiple, long axons (arrows) branching circumferentially (A’). S neurons fired fEPSPs (indicated by the arrow head). At RMP, control neurons fired either a single AP or tonically at upper depolarising current pulse amplitudes (B). S-type neurons had Dogiel type I morphology. The S neuron (corresponds to the first depolarising current pulse trace in B) has lamellar dendrites with a single orally-projecting axon (arrow) (B’).

155

Figure 6.3 Examples of cell bodies and projections of neurons in the mouse ileum. Micrographs depicting the morphologies of Dogiel type I (A and B) and Dogiel type II (B and C) cell bodies. Most Dogiel I neurons had lamellar dendrites, all Dogiel type II neurons had smooth cell bodies. Closed arrows denote axons, open arrow head denotes expansion bulb. All AH/type II neurons had an average of 3 axons/axonal branches projecting in a circumferential direction. The neuron depicted in (D) represents this typical projection pattern. Most S/type I neurons had a single axon projecting orally (E-G) The neuron represented in (E) had an orally projecting axon 650 µm in length that branched into the 156

circular muscle. Some uniaxonal neurons had orally-projecting axons that did not enter the muscle and were interrupted by the tissue dissection (F) (1000 µm projection length), (G) (1100 µm in length). Closed arrow heads denote cell bodies.

157

Figure 6.4 Myenteric neurons of CT-incubated ileal segments displayed increased spontaneous synaptic input. Stimulus-evoked fEPSPs were no different in amplitude or duration in control (A) and CT- treated preparations (B). Micrographs of control (C) and CT (12.5 µg/ml) pre-treated (D) S- type neurons with their corresponding electrophysiological recordings. Following CT exposure, S-type neurons showed an increased incidence of spontaneous firing activity (single and bursts of fast EPSPs indicated by arrow heads). Arrows denote axons.

158

Table 6.1 Electrophysiological parameters of neurons recorded from control and CT- treated ileal loops.

Parameter S-type AH-type

Control CT Control CT

RMP (mV) -46 ± 4 -52 ± 4 -60 ± 2 -53 ± 5

n = 6 n = 3 n = 3 n = 3

Threshold 160 ± 10 150 183 ± 42 100 ± 25 (pA) n = 5 n = 1 n = 3 n = 2

Max # of 3.5 ± 2.5 3 2.5 ± 0 5.3 ± 1.3 Action APs potential n = 3 n = 1 n = 3 n = 2

Firing Incidence n/a n/a 0% (0/2) 100% (3/3) Anode n = 2 n = 3 Break AP

Stim-evoked fEPSP 22.1 ± 2.9 28.6 ± 2.6 n/a n/a amplitude (mV)# n = 13 n = 12

Stim-evoked fEPSP 77 ± 10 104 ± 12 n/a n/a duration (ms)# n = 13 n =1 2

Spontaneous 15 % (3/20) 61.1% (11/18) 0% (0/3) 50% (2/4) activity# n = 20* n = 18* n = 3 n = 4

# recorded at hyperpolarised membrane potentials of -90 mV, * P < 0.01, Fisher’s exact test

159

4. DISCUSSION

In vivo pre-incubation of CT in the mouse ileal loop did not produce significant changes the calcium activity of myenteric neurons, but increased in c-Fos expression in myenteric neurons (Fung, Koussoulas et al., unpublished). Moreover, the MP was shown to be necessary for the maximal hypersecretory effect induced by CT in vivo (Jodal et al., 1993) and in Chapter 3 (Koussoulas et al., 2017) I demonstrated that CT in vitro produces a sustained hyperexcitability in myenteric sensory neurons of guinea pig jejunum. Thus, in this study I examined the potential effects of CT on myenteric neurons using intracellular recording. Furthermore, I correlated electrophysiological and morphological properties of enteric neurons in the murine small intestine. A striking sampling bias was revealed by electrophysiology, possibly arising from an anatomical basis that has important implications in the interpretation of CT’s effects on myenteric neurons. Thus, while the role of the MP may be underpowered in this study, I demonstrated that CT increased synaptic input to a proportion of myenteric neurons, which suggests that the MP plays at least a small part in contributing to sustained CT-induced hypersecretion in the mouse ileum.

4.1 Considerations of the general electrophysiology and morphology of myenteric neurons in the mouse ileum

Only a very small number of studies have examined the electrophysiology of enteric neurons in the mouse ileum and no correlations between different electrophysiological classes of neurons and their morphologies have been made. Mao et al. (2006) used intracellular recording to characterize myenteric AH neurons of mouse small intestine and confirmed that these neurons have Dogiel type II morphology via intracellular injection of Neurobiotin, but while S neurons were also recorded and noted to have Dogiel type I morphology, they were not included in the study. My study is novel in examining the membrane, synaptic, firing and morphological properties of both AH and S neurons.

AH neurons displayed phasic firing properties, in agreement with longstanding literature on other species including guinea pig (Bornstein et al., 1994), rat (Brookes et al., 1988) and human (Brookes et al., 1987) and with more recent reports on the firing properties of AH neurons in the mouse ileum (Mao et al., 2006), duodenum (Foong et al., 2012) and colon (Nurgali et al., 2004). The AP firing frequency of control neurons was also comparable to

160

reports in the guinea pig (Furness et al., 1998; Kunze et al., 1998) and mouse small intestine (Foong et al., 2012). AH neurons displayed the characteristic prolonged AHP following AP firing (Hirst et al., 1974; Bornstein et al., 1994) and the Ih-induced rectification of the membrane potential (Galligan et al., 1990; Rugiero et al., 2002b) with analogous features also observed by others in the mouse ileum (Bian et al., 2003; Ren et al., 2003; Mao et al., 2006) duodenum (Foong et al., 2012) and colon (Nurgali et al., 2004), thus the membrane and firing properties of AH neurons appear to be equivalent to those in the guinea pig and across gut regions. Accordingly, as demonstrated in the guinea pig, murine AH neurons could also form excitatory recurrent networks which are critical in determining functional plasticity and processing in the ENS, particularly in perturbed states. Networks of AH neurons are dependent on the pattern of input, contributed by highly-conserved features such as the AHP. This provides network stability and the capacity to give graded responses (Thomas et al., 2000). Altered firing of such networks via changes in these electrophysiological features, could certainly underlie any changes to circuit activity in the mouse ENS as occurs in the guinea pig. This may have implications in the mechanisms behind the neurogenic component of CT-induced hypersecretion in the small intestine in vivo (Thomas and Bornstein, 2003; Chambers et al., 2005; Chambers et al., 2014).

All neurons that displayed AH electrophysiology possessed Dogiel type II morphology in accordance with other studies in the mouse small intestine (Mao et al., 2006; Foong et al., 2012). In fact, Dogiel type II neurons with AH electrophysiology exist in the small intestine of other species such as rat (Mann et al., 1998), guinea pig (Bornstein et al., 1991a) and in the colon of the guinea pig (Lomax et al., 1999) and mouse (Nurgali et al., 2004). Since functionally classified intrinsic sensory neurons (ISNs) in the guinea pig small intestine that are mechanosensitive and chemosensitive (Bertrand et al., 1997; Kunze et al., 1995,1998, 2000) exhibit a strong correlation between AH electrophysiology and Dogiel type II morphology (Furness et al., 2004a), it is likely that AH neurons in the mouse ileum and in other intestinal regions and species are ISNs. Indeed, as observed in the guinea pig small intestine, Mao et al. (2006) demonstrated that AH neurons in the mouse ileum are mechanosensitive (Kunze et al., 1998, 2000) and so it is likely that they may also be chemosensitive.

161

S-type neurons fired either tonically (1 neuron) or fired a single AP with maximum pulse amplitude during depolarising current steps (2 neurons). The tonically-firing neuron had a maximum instantaneous firing frequency of 21 Hz, which is within the range of firing frequencies for guinea pig S-type neurons (Bornstein et al., 1991a, 1994). Under control conditions, some S neurons fired fEPSPs spontaneously as has been previously reported in the mouse colon (Furukawa et al., 1986; Nurgali et al., 2004).

S neurons possessed Dogiel type I morphology, as reported previously in the mouse small intestine (Foong et al., 2012) and colon (Nurgali et al., 2004). In the guinea pig S-type/Dogiel I neurons are interneurons and motor neurons (Brookes, 2001); it is likely that S neurons in the small intestine of the mouse have similar functions. Among the orally-projecting S type neurons, 2 projected into the circular muscle and 1 into the longitudinal muscle. These neurons were most likely motor neurons innervating the circular and longitudinal muscle much like in the guinea pig small intestine (Bornstein et al., 1991a; Brookes, 2001) and mouse colon (Nurgali et al., 2004). Two S-type neurons had orally projecting axons that could not be traced to the muscle that ran for greater distances without branching, only to be interrupted by the dissection of the tissue. It is possible that these may include ascending interneurons, however the full extent of their projection was not preserved due to their length so this can only be speculated. Indeed some interneurons run long distances without branching before supplying innervation to other neurons (Brookes, 2001) including those in the mouse colon (Nurgali et al., 2004).

A striking feature of the current study was an obvious under-sampling of certain neuronal populations, including both AH neurons and some S neurons. Indeed, it seems that relative to the commonly studied guinea pig model, AH neurons in the mouse small intestine are rare (Bian et al., 2003; Ren et al., 2003) and are even less common in the colon (Furukawa et al., 1986). This may indicate that normal intestinal reflex behaviour in the mouse requires less sensory input or input possibly via S-type neurons that have been found to have mechanosensitive function in the MP of mouse ileum (Mazzuoli and Schemann, 2012). In addition there was an obvious under-sampling of anally-projecting uniaxonal neurons in this study. The low occurrence of both AH neurons and anally projecting S-type neurons could reflect a neuronal size differential between species or in the somata of oral vs anal-projecting uniaxonal neurons and thus affect likelihood of impalement.

162

Unpublished data from our lab suggests that some neurons in the mouse intestine are small and hence may be difficult to impale. These include nNOS-immunoreactive (IR) neurons, which are typically descending interneurons and inhibitory motor neurons and hence presumably anally-projecting uniaxonal neurons, and calretinin-IR neurons which also include sensory AH/Dogiel II neurons in the mouse (Sang and Young, 1996; Qu et al., 2008). Both populations of neuronal subtypes contain neurons with areas as small as 50 µm2, whereas in the guinea pig, inhibitory motor neurons have average areas of 1023± 43µm2 in the gastric corpus (Brookes et al., 1998) and show the same pattern in the small intestine (Brookes et al., 1991b). Further to this, cell body sizes of Dogiel type II neurons in the mouse colon have been reported to be smaller than Dogiel type II neurons of the guinea pig colon (Nurgali et al., 2003b, 2004). Further to this Nurgali et al. (2004) found that in the mouse colon, of the uniaxonal neurons impaled, the majority of circular (10/13) and longitudinal muscle motor neurons (4/4) projected orally as did neurons presumed to be interneurons (40/65). Whether this sampling bias of uniaxonal neurons is due to neuronal size differences is unknown as are any functional implications, but the electrophysiological properties in orally- projecting neurons vs anally-projecting neurons were not found to be different in the study of Nurgali et al. (2004). Such sampling biases would certainly have implications in data interpretation and indeed anally-projecting S-neurons were a class of neuron that showed c- Fos expression after incubation with CT (Fung, Koussoulas et al., unpublished) and so in this instance intracellular recording does not give a comprehensive picture of circuit activity.

4.2 CT enhances secretion and increases spontaneous activity in myenteric S neurons

Incubations of CT in the ileal loop produced obvious distention only in the loop region due to the hypersecretion occurring during the incubation period. In our parallel study, tissue from CT-treated segments was confirmed to show enhanced secretion as indicated by an increase in basal short-circuit current (ISC) (Fung, Koussoulas et al., unpublished). Much of the sustained increase in ISC in the mouse ileum was found to be due to direct action of CT at the mucosa since TTX did not attenuate the CT-induced increase in basal ISC. The sustained hypersecretion was also mediated in part by a contribution from submucosal, but not myenteric neurons. In contrast, in the guinea pig in vitro, hyperexcitability of submucosal and myenteric circuits appear to drive secretion (Gwynne et al., 2009; Chapter 3). Therefore I further examined any sustained activity of the MP in detail with intracellular recording. It

163

should be noted that intracellular recording from the SMP in the murine ileum has not been successful before due to great technical difficulty and so was not pursued in this study.

At the neuronal level, I found that CT induced an increase in sustained spontaneous synaptic activity in S neurons in this tissue several hours after the incubation had ceased. Our parallel study found that CT activates some myenteric neurons at the time of incubation as indicated by an increase in c-Fos expression in a proportion of myenteric neurons (mainly nNOS-IR interneurons and motor neurons but also calretinin-IR inter/motor neurons) but does not produce sustained increases in circuit activity measured during Ca2+ imaging (Fung, Koussoulas et al., unpublished). There is a possibility that the increased c-Fos expression in the MP of CT loops may be due to distension produced by hypersecretion.

Further I did not observe any obvious sustained changes in electrophysiological properties such as RMP when examining myenteric S-type neurons with intracellular recording. However due to the under-sampling of a large population of S-type neurons (anally projecting interneurons and motor neurons), any excitability changes in this population would not have been detected.

Taken together; c-Fos expression indicated that descending interneurons and inhibitory motor neurons were active during incubation, while Ca2+ imaging did not indicate they display sustained excitation after CT exposure. However, any small increases in sustained excitation such as the synaptic activity I observed in orally-projecting neurons, would be missed due to the intracellular recording bias.

Thus in light of the experimental bias I can only speculate that it is likely CT produced only subtle changes to the circuit activity of this plexus during these experiments that did not appear to affect AP firing at RMP. In agreement with this, the spontaneous activity demonstrated in the present study, possibly due to an increase drive from other myenteric neurons (see below), was apparent at hyperpolarised membrane potentials where the driving potential is sufficient to reveal small synaptic potentials. The spontaneous firing of synaptic potentials may not have been detected in calcium imaging due to experiments recorded only at RMP and an insufficient sampling rate. The sampling rate of 1 Hz used in these experiments is likely not fast enough to examine a calcium event closely associated with a single AP and detailed membrane events such as fEPSPs (Michel et al., 2011) since Ca2+ signals measure cellular events secondary to primary electrical changes. While Ca2+ signals

164

reflect AP firing, the source of Ca2+ is still not exactly clear. This could occur via voltage- gated Ca2+ channels which are expressed by AH neurons (Kunze et al., 1994), additionally Ca2+ responses can occur via Ca2+ -permeable nAChRs as well as Ca2+release from intracellular stores (Vanden Berghe et al., 2002; Michel et al., 2011).

4.3 Effects of CT on AH neurons

My data do not allow definitive conclusions about whether CT produced sustained changes in the firing properties of the AH neuronal population, due to their small population in this gut region (Bian et al., 2003; Ren et al., 2003). This was compounded by the technical difficulty in obtaining sufficient stable impalements to properly investigate this. Out of 22 experiments under a third (7) provided stable enough conditions to impale AH neurons at a rate of 1 per experiment and only 5/7 neurons had the firing properties characterised in full. Similarly, out of 38 successful S neuron impalements, only 4 were stable enough for firing property characterisation. The limited number of AH neurons, however, demonstrated that their maximum firing during depolarising current pulses was doubled by CT treatment. Further, a unique finding in this study was that after saline incubation, AH neurons were quiescent without stimulation, but following CT incubations spontaneous firing of proximal process potentials was exhibited by 2 of 4 AH neurons. Such spontaneous firing has not previously been reported in the murine small intestine under control conditions (0 /32 AH neurons) (Mao et al., 2006), (0/26 neurons) (Foong et al., 2012) and only very rarely in mouse colon (1/23 neurons) (Nurgali et al., 2004).

In the guinea pig, in vitro pre-incubation of CT induced prolonged hyperexcitability at several sites within the enteric circuitry including changes in excitability of myenteric AH neurons, well after the incubation period (Chapter 3). We postulated that this was probably due to increased activity of these neurons during the incubation period, as they form recurrent excitatory networks with each other and interact via slow excitatory transmission (Bertrand and Galligan, 1995; Alex et al., 2001; Thomas and Bornstein, 2003; Johnson and Bornstein, 2004). As demonstrated by computer modelling, increases in the excitability of such networks increases the drive in the circuit output and in doing so, amplifies sensory input into enteric neural pathways. This potentially irreversibly enhances firing within the network (Thomas et al., 2000; Thomas and Bornstein, 2003; Chambers et al., 2005). Whether recurrent sensory networks exist in the mouse remains to be established, although the 165

projection patterns of AH neurons reported by Nurgali et al. (2004) and Foong et al. (2012) strongly suggest that they do.

In the MP of the mouse ileum, AH neurons all display sEPSPs (Mao et al., 2006; Foong et al., 2012) and it appears that AH/type II neurons project within myenteric ganglia and supply varicose terminals to other myenteric neurons (Sang and Young, 1996; Qu et al., 2008). Hence any changes in their firing, including spontaneous firing as observed in this study may also affect other elements of the circuit. This might well be the source of the spontaneous fEPSPs in S neurons observed in this study and the sustained increase in Ca2+ activity of submucosal neurons in our parallel study, since neurons in the MP of the mouse small intestine also project to the SMP (Sang and Young, 1996). Additionally, the scarcity of ileal AH neurons relative to other species such as the guinea pig, could translate into signal transduction from fewer functional sensory fields at the mucosa (Bertrand et al., 1998) to the underlying circuitry. Minimal excitatory reinforcement compared to guinea pig, within any myenteric sensory networks may account for the small changes in circuit activity at level of the cell membrane that were detectable by electrophysiology, but not higher throughput analyses. However, sensory input into murine enteric networks could also be contributed by S-type neurons with a sensory function (Mazzuoli and Schemann, 2012).

4.4 In context of the circuit

The data in this study suggest the possibility that myenteric AH neurons may also demonstrate sustained excitation. This could provide a source of ongoing excitatory input to the submucosa to amplify activity in secretomotor circuits. Indeed in the guinea pig in vitro, submucosal secretomotor neurons show sustained CT-induced excitation attributed in part, to ongoing excitatory input from myenteric AH neurons (Gwynne et al., 2009; Koussoulas et al., 2017). A point of consideration arises, however, as electrophysiological and Ca2+ imaging experiments were performed on dissected preparations of isolated plexus, which would remove a number of synaptic and mucosal inputs. In the guinea pig, mucosal mediators may have different effects on the afferent and efferent limbs of the neurogenic pathways excited by CT (Koussouals et al., 2017; Chapter 3) so all possible pathways recruited by CT may not be represented in these studies.

166

Any increase in excitation in myenteric AH neurons may be also a source of drive to other neurons since S neurons displayed an increase in spontaneous fast EPSPs after CT-treatment. However, as in guinea pig, (Koussoulas et al., 2017 Chapter 3) the overall firing and membrane properties of S neurons in vitro were most likely unchanged. It is possible that S neurons only play a small role in sustained CT-induced effects where spontaneous fast transmission could well be a source of background drive into the circuit with a role of increasing excitability in response to incoming stimuli. Additionally, the possibility that the MP is involved in the induction of CT-induced hypersecretion cannot be excluded; as indicated by expression of c-Fos in activated interneurons and motor neurons. This would certainly correlate with immediate motility effects observed following CT exposure in the guinea pig and mouse (Fung et al., 2010; Balasuriya et al., 2016). Since a significant portion of these neurons were not sampled via intracellular recording, whether they also show sustained changes in excitation is unknown.

4.5 Conclusions and future directions

This chapter demonstrated that in the mouse small intestine correlations between electrophysiological and morphological neuronal properties are analogous to those in the guinea pig. Further, this study examined the contribution of the ENS following CT incubations in the mouse ileum in vivo and revealed effects on myenteric neurons that were not found in Ca2+-imaging experiments. It was demonstrated that the MP might play a small part in contributing to the sustained excitation of secretomotor circuits which ultimately drive secretion across the mucosal epithelium. The effects of CT on the firing properties of myenteric S and AH neurons require a more extensive examination since intracellular recording was not sufficient in sampling from large enough neuronal numbers and from a physiologically representative range of neurons; more animals and time would be needed to adequately address this which is outside the scope of this study. The sampling bias and exclusion of largely inhibitory descending interneurons and motor neurons from the data set means that significant effects in excitability may have been missed. For this reason in the next chapter, the effects of TcdA in vivo were examined by the broader sampling technique of Ca2+ imaging. In light of the present work and our parallel study, while CT’s main site of action is at the mucosa, it is likely that multiple levels of the enteric circuitry contribute to CT-induced hypersecretion. Whether or not in addition to their activation during incubation,

167

inhibitory interneurons and motor neurons display sustained excitation after CT is an important point to address in future studies.

168

CHAPTER 7: EFFECTS OF IN VIVO LUMINAL INCUBATION OF TCDA ON THE MOUSE ILEUM

ABSTRACT

TcdA has both an inflammatory and extrinsic neural component in its mechanism, but the role of the enteric nervous system in its pathogenesis remains unclear. In Chapters 3 and 4 I demonstrated that intrinsic sensory neurons in the MP are a common target of the actions of enterotoxins and associated bacterial metabolites. In this this study I examined the broad enteric neural activity in both plexuses making use of activity dependent markers and Ca2+ imaging following in vivo incubation of TcdA in the mouse ileum. Sustained effects of the toxin included a reduction in spontaneous and electrically-induced calcium responses in the SMP and unchanged activity in the MP. Expression of activity dependent markers revealed that activity was unchanged during the incubation period at the site of toxin exposure, but enteric neurons in regions off-target to the toxin-affected area showed increased expression. It was concluded that the reduction in sustained excitability might have been due to damage produced by the toxin at the mucosa and underlying plexuses, while activation of off-target enteric neurons may have been due to release of inflammatory cytokines into the circulation or activation of extrinsic neural pathways.

1. INTRODUCTION

C.difficile toxin A (TcdA) like CT, stimulates a large movement of water and electrolytes across the gut mucosa into the lumen during in vivo incubations in rodent ileal loops, (Triadafilopoulos et al., 1987; Pothoulakis et al., 1994; Castagliuolo et al., 1998a). However, in contrast to CT, TcdA has an inflammatory component in its mechanism; ileal loop incubations with TcdA evoke mast cell activation, neutrophil recruitment and tissue necrosis (Castagliuolo et al., 1994; Kelly et al., 1994; Pothoulakis et al., 1998). After binding to carbohydrate cell surface receptors on the apical membrane of enterocytes, TcdA alters the tight-junction permeability of these cells due to disruption of the actin cytoskeletal framework via inhibition of the ADP-ribosylation of GTP-binding Rho proteins (Hecht et al.,

169

1988; Just et al., 1995). This causes a decrease in epithelial cell resistance, an increase in tight junction permeability, and mucosal damage and oedema which is characteristic of C.difficile infection (CDI) (Pothoulakis et al., 1998). TcdA also has a neuronal mechanism; the intestinal response to TcdA is reduced by the neuronal blockers lidocaine and hexamethonium and by functionally ablating extrinsic primary afferent nerves through capsaicin treatment (Castagliuolo et al., 1994).

In Chapter 4 I showed that TcdA incubated luminally in vitro, increases the excitability of myenteric sensory neurons which appear to be key points of convergence for enterotoxins, including CT (Chapter 3). This is consistent with previous suggestions that TcdA also requires activation of the enteric nervous system (ENS) to elicit a response (Pothoulakis et al., 1994; Xia et al., 2000). Myenteric sensory neurons project to and synapse with most other myenteric and submucosal neurons in the guinea pig (Bornstein et al., 1987; Furness et al., 2004a) while, in the mouse, myenteric sensory neurons project within the ganglion suppling varicose terminals to other myenteric neurons and to the SMP (Sang and Young, 1996; Qu et al., 2008). Hence, whether or not TcdA incubated luminally excites other elements of the enteric circuitry is unknown.

In this study, TcdA was incubated luminally in vivo using the mouse ileal loop model since in Chapter 4 I found that effects of TcdA and bacterial metabolites such as GABA on myenteric neuronal excitability likely involved indirect effects, and the ileal loop model better preserves extrinsic and mucosal pathways that may contribute to neuronal responses. For this reason the ileal loop model is the gold standard model for examining TcdA effects (Castagliuolo et al., 1998a; Savidge et al., 2011) allowing for more physiologically relevant examination of C.diff pathology. The sampling bias revealed by intracellular recording when examining the effects of CT in vivo using the same incubation model in the mouse small intestine (Chapter 6), resulted in the exclusion of significant functional neuron populations in the data set. Thus intracellular recording was not used in this study. Instead the broad activity and contribution of both enteric plexuses after TcdA incubation was examined by making use of Wnt1- Cre;R26R-GCaMP3 mice. Ca2+ imaging was implemented as a high throughput approach to identify any sustained effects induced by the toxin and expression of activity-dependent markers was examined to determine if changes to the circuitry occurred during the incubation. I found that TcdA produced an overall reduction in excitability in the submucosal

170

plexus from the TcdA exposed ileal loop region, but appeared to activate submucosal and myenteric neurons in regions proximal and distal to the site of acute toxin exposure.

2. METHODS

2.1 Mice

Experiments were performed using male mice on a C57BL/6 background including Wnt1- Cre;R26R-GCaMP3 mice that express the fluorescent calcium indicator GCaMP3 in all enteric neurons and glia, aged 8-12 weeks (see Chapter 2).

2.2 Ileal loop surgery

Ileal loop surgery was performed on mice as described in Chapter 2 (section 2). Either physiological saline or toxin in saline (TcdA: 60 µg/mL, a kind gift from Tor Savidge, Baylor College of Medicine Texas and Charalabos Pothoulakis, David Geffen School of Medicine UCLA ) was injected into the ileum filling the region (200 µL of incubating solution). A commercial supply of toxin was trialled initially (List Biological Laboratories, California) but did not yield satisfactory results. Therefore in using the gifted stock only a limited number of experiments were possible before the supply was exhausted. During a 2.5-3.5 hr incubation period the animal was monitored closely and culled immediately if signs of pain and discomfort were exhibited (which sometimes occurred as early as 2.5 hrs into the incubation), identified using the mouse grimace scale (Langford et al., 2010; Matsumiya et al., 2012). After the incubation period, the animal was killed by cervical dislocation as approved by the University of Melbourne Animal Experimentation Ethics Committee. The ileal loop segment was dissected from the abdomen for in vitro experimentation. Off-target tissues including jejunum, ileum 1-2 cm proximal to the ileal loop and proximal colon were also collected for immunohistochemical and histological analyses.

2.3 Validation of the viability of TcdA

The TcdA was shipped from Houston on dry ice. Upon its arrival the viability of the toxin was first confirmed by performing a cytotoxicity assay, testing the toxin on cultured Vero

171

cells (Sigma-Aldrich, Castle Hill, NSW, AUS) (Lyras et al., 2009; Carter et al., 2015) as described in Chapter 2, section 4.

By 24 hours, TcdA induced a cytotoxic response in all Vero cells so they appeared round, detached and a number were floating free in the medium (Figure 7.1).

Figure 7.1 TcdA (1µg/ml) cytotoxicity assay. Vero cells showing 100% rounding after 24 hours of exposure to TcdA, validating the cytotoxicity of TcdA.

172

2.4 Histological processing and imaging of ileal loop and off-target tissue

To assess the integrity of the mucosa, cryosections (20 μm) (see Chapter 2, section 3) of ileal loop tissue incubated with either saline or TcdA in saline and off-target jejunal tissues were processed with Hematoxylin and Eosin (H&E) staining. The protocol for staining is outlined in Chapter 2, Table 2.1. Preparations of H&E-stained tissue were viewed through a 20x objective and imaged using an Axio Imager D.1 microscope with an AxioCam MRc5 camera and AxioVision software (version 4.8.2.0) (all from Zeiss, Australia).

2.5 Measurement of short-circuit current (ISC) in vitro using Ussing chambers

Full thickness preparations of ileal loop tissue, with all intestinal layers intact, were mounted in Ussing chambers (see Chapter 2, section 7). Following a 30 min equilibration period, - short-circuit current (ISC), an index of electrogenic Cl secretion across the intestinal wall (Clarke, 2009), was measured throughout the experiment.

Data collection and analysis were performed using AcqKnowledge 3.9.0 software (BIOPAC Systems, Inc., SDR Clinical Technology, Middle Cove, NSW, Australia). The maximum change in ISC from baseline (ΔISC) was measured and compared between control and treatment groups.

2.6 Ca2+ imaging and analysis

Saline and TcdA-treated ileal loops from Wnt1-Cre;R26R-GCaMP3 were used for Ca2+ imaging experiments. Once removed from the abdomen, they were placed in physiological saline, where the tissue was prepared for imaging as described in Chapter 2, section 6.1 with up to 2 preparations each of submucosal plexus (SMP) and myenteric plexus (MP) from each ileal loop segment were obtained. Ganglia from each plexus were imaged at 1 Hz without stimulation to assess spontaneous activity and were also electrically stimulated as described in Chapter 2 section 6.2. The electrical stimulation regime was performed once on 3 different ganglia for each plexus (MP and SMP). To assess spontaneous activity, 3 different ganglia per plexus were recorded for 2 2+ mins and spontaneous [Ca ]i transients were analysed after recording.

Analyses were performed using custom-written directives in IGOR Pro (WaveMetrics, Lake Oswego, Oregon, USA). Regions of interest were drawn over a selected area of the 173

2+ cytoplasm for each neuron. The amplitude of each ([Ca ]i) transient signal was calculated 2+ and expressed as the maximum increase in [Ca ]i from the baseline signal (ΔFi/F0) (as described in Chapter 2, section 6.3 ). A minimum of 3 animals were examined for each condition.

Data are presented as the mean ΔFi/F0 ± SEM where n = number of neurons examined. Statistical analyses were performed using unpaired t-tests with P < 0.05 considered statistically significant. Comparisons were performed using GraphPad Prism 5.0 (GraphPad Softwares, San Diego California).

2.7 Immunohistochemistry for activity dependent markers

Wholemounts of ileal loop and off-target tissue (as outlined above) from mice on a C57BL/6 background, including Wnt1-Cre;R26R-GCaMP3 mice, that had undergone ileal loop surgery using either saline or TcdA and saline incubations were double or triple-labelled with primary antisera (Table 7.1) including antisera raised against the activity dependent markers c-Fos (rabbit anti c-Fos) and pCREB (rabbit anti pCREB) for 72 hours at 4°C. Preparations were washed 3 times for 10 mins with PBS and incubated with biotinylated donkey α rabbit IgG (1:100; Jackson Immuno Labs, West Grove, Pennsylvania, USA) for 2 hours at room temperature. After another 3 x 10 min washes, preparations were incubated with strepavadin AF594 (1:200; Molecular Probes, Eugene, Oregon, USA) and secondary antisera (Table 7.1) for 2 hours and 15 mins, rinsed again in PBS 3 times and mounted on slides with Dako fluorescent mounting medium (Carpinteria, California, USA).

2.8 Imaging and analysis of activity-dependent markers

Myenteric and submucosal preparations were imaged and analysed as described in Chapter 2 section 5. The data are expressed as mean ± SEM and n = the number of cells examined. Statistical analyses were performed using unpaired t-tests and Fisher’s exact tests with P < 0.05 considered statistically significant. Comparisons were performed using GraphPad Prism 5.0 (GraphPad Softwares, San Diego California).

174

Table 7.1: Primary and secondary antisera used for immunostaining.

Primary Antisera Raised in Dilution factor Source pCREB Rabbit 1:1000 Millipore c-Fos Rabbit 1:5000 Oncogene

Hu Human 1:5000 Gift from Dr V. Lennon nNOS Sheep 1:1000 Gift from P. Emson

Secondary Antisera Raised in Dilution factor Source

Anti-human AF 647 Donkey 1:500 Jackson Immuno Labs

Anti-sheep AF 488 Donkey 1:400 Molecular Probes

175

3. RESULTS

3.1 Incubations of TcdA

TcdA was trialled in vivo at various concentrations up to 10x lower than that used in this study, without producing hypersecretion and distention in the loop. The concentration implemented in this chapter was selected as it produced obvious secretion and mucosal damage as reported by others in the mouse ileum (Castagliuolo et al., 1998a). In addition, administration of TcdA required a greater of volume of saline in the loop (compared to CT) to generate its hypersecretion; 200 µL of vehicle was injected into TcdA loops compared to 100 µL in CT loops (Chapter 6). As a result, under control conditions loops in this study were more distended in comparison to those in the CT study (see Figure 6.1 A Chapter 6, Figure 7.2 A below).

3.2 TcdA in the ileum in vivo produces gross mucosal damage and a sustained increase in basal short circuit current

Segments of ileum incubated with TcdA (60 µg/mL) for 2.5-3.5 hours showed an obvious distention in the loop region compared to saline controls, presumably from fluid accumulation during the incubation period (Figure 7.2 A, B, D, E). In animals with TcdA- treated loops, the loop, plus regions oral and distal to the loop and the mesentery displayed signs of vasodilation, specifically increased redness, compared to controls, (Figure 7.2 B-B’, E-E’). Redness of the tissue has been reported in rodent models of inflammation including models of CDI (McLarren et al., 2011; Best et al., 2012) and is likely due to extensive vasodilation in the blood supply. This was not observed in CT-incubated loops (see Figure 6.1 Chapter 6).

To assess the effect of TcdA on mucosal integrity, ileal loops were processed using Haematoxylin & Eosin (H&E) staining. Toxin-treated loops (n = 2) showed gross mucosal damage, including a complete breakdown of the mucosal epithelium with total destruction of villus and crypt architecture, including crypt loss, compared to saline-treated tissues (n = 2) (Figure 7.2 C, F). Such changes to epithelial structure are defined as ‘marked’ using histomorphological scores for intestinal inflammation in rodent models (Erben et al., 2014).

176

To then confirm that this model of toxin-incubation produced a sustained increase in basal secretion, full thickness preparations of either control or TcdA-treated ileal loop tissue were placed in Ussing chambers where short-circuit current (ISC) was measured. Basal ISC was significantly higher in TcdA-treated preparations (98.7 ± 5.3 µA/cm2, n = 2 mice) compared to controls (64.5 ± 3.2 µA/cm2, n = 3 mice, P < 0.01) (Figure 7.2 G).

Figure 7.2 Control vs. TcdA-treated ileal loops following 2.5-3.5 h incubation in vivo. Images of ileal loops (marked with yellow asterisks) following 2.5-3.5 hour incubation in vivo with either saline (A, B) or TcdA (60 µg/mL) (D, E) injected into the lumen. The ileal loop and other areas of the gastrointestinal tract including the mesentery (white arrow, E’), areas proximal (white arrow head) and distal (white arrow head outlined) to the loop region (E) from TcdA-treated animals showed signs of vasodilation. Scale bars = 1 cm. H&E staining shows severe damage to the mucosal epithelium in TcdA loops compared to controls (C, F), black arrow head marks villus of the mucosal epithelium. Scale bar = 100 µm. Basal

ISC measured in full thickness TcdA-incubated ileal loop tissue (n = 2) was significantly higher than that of saline controls (n = 3; P < 0.01) (G).

177

3.3 TcdA does not alter pCREB expression in submucosal neurons, but submucosal neurons display a sustained reduction in spontaneous Ca2+ transients

To examine whether TcdA affected neural activity within the SMP during the incubation period, the expression of the activity dependent marker pCREB (Sheng and Greenberg, 1990) was examined by co-expression with the pan-neuronal marker Hu. SMP preparations were also processed for the activity-dependent marker c-Fos (Hunt et al., 1987; Morgan and Curran, 1989; Sheng and Greenberg, 1990), but c-Fos expression in the mouse SMP has not been compellingly verified (Bjerknes and Cheng, 2001) and I did not detect c-Fos expression in the SMP in my experiments. Accordingly, I only quantified pCREB expression for the SMP.

The proportion of Hu-immunoreactive (IR) cell bodies in the SMP that expressed pCREB did not differ between control preparations and those from TcdA-treated ileal loops (Control: 123 pCREB-IR neurons, n = 139 Hu-IR neurons in total; TcdA: 106 pCREB-IR neurons, n =128 Hu-IR neurons in total, P > 0.2, Figure 7.3). Additionally the proportion of non-Hu-IR cell bodies (presumably glial targets for the actions of TcdA) in the SMP that expressed pCREB was unchanged by toxin treatment (control: 0.6 ± 0.1 cells/ganglion, n = 20 ganglia; TcdA: 1.0 ± 0.3 cells/ganglion, n = 15 ganglia, P > 0.3).

Submucosal ganglia from ileal loop tissue in Wnt1-Cre;R26R-GCaMP3 mice were imaged 2+ for [Ca ]i without stimulation to assess spontaneous activity within the circuitry. Fewer 2+ neurons in TcdA-treated ganglia exhibited spontaneous [Ca ]i transients than in saline controls. Under control conditions, 12/70 neurons were spontaneously active, while 5/98 TcdA-treated neurons were spontaneously active (Fisher’s exact test P = 0.017).

Ganglia were also electrically stimulated to assess any changes to the circuitry. Single pulse- 2+ evoked [Ca ]i transients did not differ in amplitude between the two treatment groups, but 2+ train-evoked [Ca ]i transients were smaller in TcdA-treated preparations than in saline

(single pulse ΔFi/F0 control: 0.09 ± 0.01 n = 41; ΔFi/F0 TcdA: 0.09 ± 0.01 n = 50 P > 0.05; 20 pulse: ΔFi/F0 control: 0.40 ± 0.02 n = 68 ΔFi/F0 TcdA: 0.30 ± 0.01 n = 87, P < 0.001) (Figure 7.4)

178

Figure

7.3 Submucosal neurons do not show a change in pCREB expression following TcdA exposure. TcdA did not produce any changes in pCREB expression during incubation period. Images of representative submucosal ganglia from the ileal loop illustrating neurons stained for pCREB (A-B) and merged images of pCREB and Hu (A’-B’). Merged images demonstrate colocalisation of pCREB with Hu-IR neurons in control and TcdA-treated preparations. Scale bars = 20 μm. Filled arrows indicate colocalisation of pCREB with the pan neuronal marker. Histogram showing no change in the expression of pCREB in the SMP of the ileal loop (C).

179

Figure 7.4 Submucosal neurons show an overall sustained reduction in excitability. Fluorescence images of neurons responding to 20 pulse (20 Hz) electrical stimulation with 2+ [Ca ]i transients in TcdA-treated preparations (A-A’). Scale bars = 20 µm. In TcdA 2+ preparations, trains of electrical stimulation produced [Ca ]i responses that were reduced in amplitude compared to controls. Numbers of neurons examined are displayed within each histogram (B).

180

3.4 TcdA does not induce changes in c-Fos expression or sustained changes in excitability in the MP

The co-expression of Hu-IR cell bodies with c-Fos was examined in the MP to determine if TcdA induced immediate effects during incubation. The proportion of Hu-IR cell bodies that expressed c-Fos in TcdA- treated loops did not differ from that observed in saline loops (Control: 283 c-Fos-IR neurons, n =781 Hu-IR neurons in total; TcdA: 254 c-Fos-IR neurons, n = 684 Hu-IR neurons in total, P> 0.7, Figure 7.5). Additionally co-expression of c-Fos with the neuronal subtype marker nNOS was examined to determine if any changes in the activity of a specific neuronal subtype occurred. No difference in the proportion of expression between the saline and TcdA-treated preparations was observed (control: 140 c-Fos-IR neurons, n = 200 nNOS-IR neurons in total; TcdA: 117 c-Fos-IR neurons, n =181 nNOS-IR neurons in total, P > 0.05). The proportion of non-Hu-IR cell bodies (presumably glia) in the MP that expressed c-Fos was unchanged by toxin treatment (control: 4.4 ± 1.1 cells/ganglion, n = 32 ganglia; TcdA: 4.1 ± 0.9 cells/ganglion, n = 20 ganglia, P > 0.8). The expression of pCREB was also examined, but was not quantified due to excessive non-specific staining produced by this this marker.

Sustained effects on control and TcdA-treated myenteric ganglia were examined using Ca2+ imaging. No differences in spontaneous activity were observed between the two groups (control: 26/243 neurons were spontaneously active, TcdA: 21/233 neurons were spontaneously active, Fisher’s exact test P = 0.65)

2+ Similarly, electrically-induced [Ca ]i transient responses were unaffected by TcdA treatment

(single pulse ΔFi/F0 control: 0.15 ± 0.01 n = 87; ΔFi/F0 TcdA: 0.14 ± 0.01 n = 166 P > 0.05;

20 pulse: ΔFi/F0 control: 0.57 ± 0.02 n = 179; ΔFi/F0 TcdA: 0.56 ± 0.02 n = 275, P> 0.05) (Figure 7.6).

181

Figure 7.5 Myenteric neurons do not show a change in c-Fos expression following TcdA exposure. TcdA did not produce any changes in c-Fos expression during the incubation. Images of representative myenteric ganglia from the ileal loop illustrating neurons stained for c-Fos (A- B) and merged images of c-FOS, nNOS and Hu (A’-B’). Merged images demonstrate colocalisation of c-Fos, with nNOS-IR and Hu-IR neurons in control and TcdA-treated preparations. No difference in co-expression of c-Fos with the neuronal subtype marker nNOS was observed. Scale bars = 20 μm. Filled arrows indicate colocalisation of c-Fos with the pan neuronal marker and nNOS. Histogram showing no change in the expression of c-Fos in the MP of ileal loop (C).

182

Figure 7.6 MP shows no sustained changes in excitability following TcdA exposure in vivo. 2+ Fluorescence images of neurons responding to 20 pulse electrical stimulation with [Ca ]i transients in TcdA-treated preparations (A-A’). Scale bars = 20 µm. Electrically-evoked 2+ [Ca ]i responses were unaffected by TcdA. Numbers of neurons examined are displayed within each histogram (B).

183

3.5 TcdA induced epithelial damage and increased neuronal pCREB and c-Fos expression in off-target regions

Since signs of vasodilation and inflammation were observed in areas proximal and distal to the confined ileal loop region, with the most obvious effects observed in the proximal region, jejunal tissue was collected and processed using H&E staining to assess the mucosal integrity of this proximal region of gut. Jejuna proximal to toxin-treated loops showed gross mucosal damage compared to controls (n = 2). Unlike the damage observed within the ileal loop, disruption of the epithelial architecture was more confined to the villi and would be defined as ‘moderate’ in the histomorphological scale mentioned above (Figure 7.7).

In the SMP preparations, TcdA induced an increase in expression of pCREB in the jejunum (control: 174 pCREB-IR neurons, n = 248 Hu-IR neurons in total; TcdA: 239 pCREB-IR neurons, n = 253 Hu-IR neurons in total, P < 0.0001, Figure 7.8), a reduction in pCREB expression in the region proximal to the loop (control: 211 pCREB-IR neurons, n = 236 Hu- IR neurons in total; TcdA: 166 pCREB-IR neurons, n = 202 Hu-IR neurons in total, P < 0.05, Figure 7.8) and no change in expression in the proximal colon, (control: 494 pCREB-IR neurons, n = 535 Hu-IR neurons in total; TcdA: 453 pCREB-IR neurons, n = 475 Hu-IR neurons, in total P > 0.05, Figure 7.8).

In the MP, TcdA induced an increase in c-Fos expression in all off-target tissue including the jejunum (control: 16 c-Fos-IR neurons, n = 540 Hu-IR neurons in total; TcdA: 68 c-Fos-IR neurons, n = 489 Hu-IR neurons in total, P < 0.0001, Figure 7.8), ileum proximal to the loop (control: 25 c-Fos-IR neurons n = 754 Hu-IR neurons in total; TcdA: 70 c-Fos-IR neurons, n = 715 Hu-IR neurons in total, P < 0.0001) and the proximal colon (control: 26 c-Fos-IR neurons, n = 1654 Hu-IR neurons in total; TcdA: 187 c-Fos-IR neurons, n = 1222 Hu-IR neurons in total, P < 0.0001, Figure 7.8).

184

Figure 7.7 TcdA-induced epithelial damage in the jejunum. H&E staining shows healthy mucosal epithelium in jejunum of saline-treated animals (A) and severe damage of the mucosal epithelium in the jejunum of TcdA-treated animals (B). Scale bar = 100 µm.

185

Figure legend opposite 186

Figure 7.8 Expression of activity-dependent markers in off-target regions of the GIT. Images of representative submucosal ganglia from the jejunum and proximal colon illustrating neurons stained for pCREB (A-D) and merged images of pCREB and Hu (A’-D’). Merged images demonstrate colocalisation of pCREB with Hu-IR neurons, filled white arrows denote colocalisation of pCREB with the pan neuronal marker, open arrows denote no colocalisation. Images of representative myenteric ganglia illustrating neurons stained for c- Fos (E-H) and merged images of c-Fos and Hu (E’-H’). Merged images demonstrate colocalisation of c-Fos with some Hu-IR neurons in TcdA-treated preparations in both regions of the gut. Filled arrows indicate colocalisation of c-Fos with the pan neuronal marker. Scale bars = 20 μm. Histograms showing a TcdA-induced increase in the expression of pCREB in the SMP of jejunal tissue, but not in other off-target regions (I) and an increase in the expression of c-Fos in the MP was found in all off-target regions of gut proximal and distal to TcdA-treated loops (D).

187

4. DISCUSSION

In this study, TcdA in the ileal loop was shown to evoke signs of inflammation, gross mucosal damage and increased secretion within the loop as has been described previously (Castagliuolo et al., 1998a). At the level of the enteric circuitry and well after the incubation period, TcdA produced an overall sustained reduction in the activity of the SMP as indicated by spontaneous and electrically-evoked activity during Ca2+ imaging. Activity dependent markers were employed as a measure of neural changes occurring at the time of incubation and TcdA appeared to activate submucosal neurons proximal to the site of acute toxin exposure and myenteric neurons in regions both proximal and distal to the toxin-treated site. Therefore the key off-target area activated by TcdA appears to be the jejunum.

4.1 TcdA produces secretion, mucosal damage and inflammation

TcdA characteristically produces hypersecretion, mucosal damage and inflammation when incubated in vivo across a number of rodent models, including the mouse (Lyerly et al.. 1982; Castaglioulo et al., 1994, 1998). Consistent with this, preparations from TcdA-treated loops in this study exhibited enhanced secretion observed as increased fluid accumulation in the loop and measured as a sustained increase in basal electrogenic secretion. The mucosa from the loop region showed gross epithelial villus and crypt damage. While the increase in basal secretion might be due to hypersecretion, since mucosal transepithelial resistance was not examined there is also the possibility that the increase in basal current was due to breakdown of the mucosa so it is a less effective barrier to current flow. Additionally, there was a ‘flushing’ or redness of the loop tissue and connected mesentery in TcdA-treated animals which was probably due to vasodilation; a sign of generalised inflammation in mouse models (McLarren et al., 2011; Best et al., 2012). Vasodilation and increased fluid movement into the lumen of the loop probably serves a protective function in assisting the elimination of the toxin. Vasodilation and erosion of the mucosa is likely due to TcdA-induced activation of extrinsic primary afferents and release of substance P (SP) on vascular cells to produce vasodilation (Lembeck and Holzer, 1979; Louis et al., 1989; Pothoulakis et al., 1994) and SP –mediated erosion via activation of mast cells (Castagliuolo et al., 1994; Wershil et al., 1998) and subsequent recruitment of neutrophils (Kelly et al., 1994; Kurose et al., 1994). In agreement, chemical ablation of extrinsic afferents and SP antagonists inhibit secretion of

188

fluid into the lumen and erosion of the mucosa (Castagliuolo et al., 1994; Pothoulakis et al., 1994).

4.2 TcdA evokes a sustained reduction in the excitability of the SMP and no change in excitability in the MP within the ileal loop

The inflammatory and hypersecretory effects of TcdA as described above, are known to be partially neurally mediated by extrinsic primary afferents. Although evidence of the contribution of the ENS in the response has been published, studies have been limited (Mantyh et al., 1996; Xia et al., 2000). This study is in its preliminary stage, but it is the first to broadly survey the involvement of the ENS following in vivo exposure of TcdA.

The administration of TcdA required a greater of volume of saline in the loop to generate its hypersecretion than CT (in Chapter 6). This resulted in a more distended control loops in this study which is likely to have activated neural pathways in the muscle coat and mucosa of the loops leading to a higher basal level of activity dependent maker expression between the controls in this study and those in the CT study (Fung, Koussoulas et al., 2017, unpublished). Since the basal control levels within the ileal loop region for pCREB and c-Fos expression was already high, this might not be a useful assay for neuron activation, but this is not the case for examination of off –target regions (see below).

The lack of any change in expression of the activity dependent marker pCREB suggests that there was no change in neuronal activity within the plexus during the incubation. Activity of submucosal neurons during incubation cannot be excluded however because most neurons expressed pCREB even under control conditions, possibly due to the lower activation threshold of pCREB compared to c-Fos (Fields et al., 1997). The high values of expression made it challenging to detect any changes induced by the toxin, as such this finding is not definitive, but does allow a conclusion that there was no decrease in pCREB expression.

I found that the excitability of the SMP was reduced following TcdA exposure, as indicated by a reduction in the number of spontaneously active neurons and a decrease in the amplitude of electrically-evoked calcium responses. This taken together with the lack of effect on the level of pCREB staining, although not definitive, may suggest that hyperactivity during the incubation is unlikely.

189

Despite extensive vasodilation, it is puzzling that an increase in the activity of submucosal neurons did not occur. Indeed, when applied directly to the cell bodies of submucosal enteric neurons of the guinea pig, TcdA is excitatory (Xia et al., 2000). The reduction in sustained excitation within the SMP however, may simply be due to damage produced by TcdA at the mucosa and possibly the underlying SMP, leading to reduced excitability.

Concentrations used by Xia et al. (2000) were higher (0.6 mg/mL) but were further diluted by an unknown amount when injected into the recording chamber during the intracellular study and so exact concentrations were not determined. Thus it is possible that in this study, while the concentration of toxin produced well known effects on the mucosa characteristic of C.diff, this was too damaging for underlying neurons. Further, in the electrophysiological study the toxin was not in prolonged contact with the neurons, while this study used incubations of up to 3.5 hours.

Future experiments should include precise determination of effective concentrations of TcdA for use in prolonged incubations when examining enteric neuronal activity. In my preliminary experiments, concentrations 10-fold lower were incubated in the ileal loop without activating any enteric circuitry, hence the concentration range will likely be narrow.

The toxin did not appear to activate myenteric neurons at the time of incubation as indicated by c-Fos expression in this plexus, nor did it produce any sustained changes measured by Ca2+ imaging. Since TcdA induced an inflammatory response, and inflammation per se enhances the firing of some myenteric (Palmer et al., 1998; Linden et al., 2003; Nurgali et al., 2007) and submucosal neurons (Lomax et al., 2005, 2006) it is somewhat surprising that TcdA did not alter the activity of enteric neurons in this study. Changes to sensory neurons in inflamed states are also associated with changes in membrane and synaptic properties including reductions in afterhyperpolarising potentials (AHP), augmentation of hyperpolarisation-activated cation (Ih) currents and changes in fast synaptic input to these neurons (Linden et al., 2003, Lomax et al., 2005, 2006) as well changes to slow excitatory responses and increases in spontaneous firing of fEPSPs in S-type (motor) neurons (Lomax et al., 2005, 2006). It might be that any sustained firing, membrane or synaptic changes to neurons were events too subtle to be detected by Ca2+ imaging at the sampling rate of 1Hz implemented in this study. Subtle sustained synaptic changes in the myenteric circuitry were observed using intracellular recording following CT incubation in the ileal loop that were not

190

detected via Ca2+ imaging (Chapter 6). Thus, further detailed investigation into the contribution of the MP using electrophysiology would be required, but due to the difficulty of acquiring the toxin, its limited shelf life and the time consuming nature of intracellular recording, this was not pursued.

Further, any such changes during incubation could have occurred transiently and with insufficient intensity to induce c-Fos activation, which depends on stimulus frequency (Sheng et al., 1993; Fields et al., 1997). Additionally, lack of myenteric activity may also be a consequence of the gross damage produced by the toxin, which underlies the need to find a lower effective dose.

4.3 Potential role for extrinsic neural activity within the ileal loop

Since TcdA is known to increase primary afferent activity this would also be expected to contribute to activation of enteric neurons. When mesenteric nerves entering the ENS are stimulated in the guinea pig ileum, sEPSPs are evoked in myenteric sensory neurons (Takaki and Nakayama, 1990). Further, submucosal ganglia in guinea pig small intestine receive inputs from SP-containing extrinsic nerve fibres (Costa et al., 1981) and others have shown that stimulating SP-containing axons excites submucosal secretomotor neurons in the ileum (Vanner and MacNaughton, 1995; MacNaughton et al., 1997). However since no increases in activity were observed in either plexus and primary afferents supply collateral branches in sympathetic prevertebral ganglia (Furness et al., 1998; Szurszewski et al., 2002), activation of intestino-intestinal reflexes and therefore inhibitory sympathetic outflow via the spinal cord should also be considered.

Sympathetic noradrenergic fibres ramify within myenteric ganglia and cause a presynaptic inhibition of cholinergic transmission to S neurons (Hirst and McKirdy, 1974b; Furness et al., 1998), while noradrenergic innervation of submucosal ganglia and stimulation of sympathetic noradrenergic neurons produces inhibitory post-synaptic potentials in secretomotor neurons (Furness and Costa, 1974; North and Surprenant, 1985; Bornstein et al., 1988). Therefore it may be that reflex sympathetic activation inhibits myenteric and submucosal pathways to counteract any excitability increases within the loop region.

191

4.4 Some enteric neurons in off-target regions may be activated during TcdA incubation

I found that areas of the GIT oral and distal to the loop including the mesentery, exhibited vasodilation and hence signs of inflammation (McLarren et al., 2011) in animals treated with TcdA. Additionally, jejuna proximal to toxin-treated loops showed epithelial damage while c- Fos and pCREB expression indicated that myenteric and submucosal neurons in jejuna and myenteric neurons in all other off-target tissues (ileum proximal to loop and proximal colon) were activated during the toxin incubation phase. A point to note, however, is that under control conditions, pCREB expression in the jejunum was lower than within the loop and the other off target regions. This could be due to jejuna being furthest from disturbance of the mucosa from saline-evoked distention within the loop, therefore less basal activity in this proximal region occurred, but this is as yet unknown. Nonetheless this casts some doubt on TcdA- increased effects seen in submucosal neurons in the jejunum.

The increase in neural off-target activity and off-target mucosal damage could be due to at least three possible mechanisms: leakage of TcdA, circulation of inflammatory mediators, or neural mechanisms.

Firstly, activation of enteric neurons and mucosal damage in off-target regions could be due to the toxin leaking out of the loop during the incubation and into adjacent regions of the bowel, but this is unlikely since using equivalent incubations I performed with CT (Chapter 6), colleagues observed no indication of neuronal activation outside the loop region.

Alternatively, several inflammatory cytokines, including prostaglandin E2 (PGE2), leukotrienes, tumor necrosis factor (TNF-α) and interleukins are released across the mucosa following TcdA exposure (Pothoulakis and Lamont 2001), and agents that reduce cytokine synthesis reduce effects of TcdA (Triadafilopoulos et al., 1989). It is therefore likely that such mediators infiltrated the gut wall and entered the circulation to activate inflammatory pathways and off-target enteric neurons. Indeed, such mediators can directly influence the behaviour of enteric neurons (Dekkers et al., 1997a; Liu et al., 2003; Lakhan and Kirchgessner, 2010). One further possible circulating mechanism could be TcdA itself crossing through the damaged mucosa into the circulation to gut other regions triggering mucosal inflammation and activating enteric neurons.

A third mechanism could be that since TcdA operates via extrinsic afferents, activation of these extrinsic pathways may account for the changes in marker expression observed outside 192

of the loop region, primarily in the MP. Aligned with this, chemical ablation of extrinsic fibres does not prevent CT –induced hypersecretion in vivo (Castagliuolo et al., 1994; Pothoulakis et al., 1994) and CT studies by colleagues demonstrated no change in c-Fos expression in the MP of off -target tissue (Fung et al., 2017, unpublished). Indeed, evidence of extrinsic innervation of the MP plexus in the mouse exists. In the mouse small intestine labelled extrinsic nerve fibres in the MP have been identified apposing sensory, interneuron/motor neurons (Tan et al., 2010), while in the colon direct anterograde tracing from DRG neurons has labelled spinal afferent endings densely innervating myenteric ganglia (Spencer et al., 2014). Since myenteric neurons in the mouse small intestine probably also project to the SMP as they are known to in the guinea pig (Furness and Costa, 1982; Furness et al., 1990b; Sang and Young, 1996), they may alter other elements in the circuitry and account for the increase in pCREB expression observed in the SMP of the jejunum. Further to this, direct extrinsic innervation of submucosal ganglia in the guinea pig small intestine has been shown (Costa et al., 1981).

Moreover, SP released from extrinsic primary afferents following in vivo incubation of TcdA in the rat has been shown to interact with SP binding sites on enteric neurons of the submucosal and myenteric plexuses (Mantyh et al., 1996). It is postulated that stimulated enteric neurons in turn initiate a signalling cascade via yet undefined mechanisms. This could well be a source of enteric neuron activation in the present study. While afferents innervating the mouse jejunum do in fact express SP (Tan et al., 2008), the individual neurons that supply the jejunum probably do not have collaterals that supply the ileum, thus it might be that primary afferent axon reflexes do not play a role in the off target effects observed in this study. However it is possible that ileal-jejunal reflexes may occur via vagal efferents where vagal efferent fibres form predominately excitatory cholinergic synaptic connections with enteric neurons including those in the MP (Kirchgessner and Gershon, 1989; Browning and Travagli, 2014).

4.5 Conclusions and future directions

The study is the first to look at the broad involvement of enteric neurons in TcdA-induced intestinal effects. While the toxin produced a reduction in the excitability of the SMP, it activated enteric circuitry in regions off-target the site of acute toxin exposure, possibly via release of inflammatory cytokines into the circulation or extrinsic neural pathways. It is 193

unknown if this activation of enteric neurons is a sustained effect and what the identities of the activated neurons are; these will be an essential focus of future studies. Other future studies should include determining effective concentrations of TcdA for examining effects on enteric neurons and the possibility of TcdA in inducing neuronal cell death could be tested using markers of neuronal cell death such as active caspase-3. An examination of the functional effects of incubating TcdA and GABA together in the loop in light of GABA’s emerging role in the pathogenesis of CDI would be apposite. Further, the actions of GABA receptors I identified in chapter 5 on any functional effects could be examined.

194

CHAPTER 8: CONCLUSIONS

Perturbed reflex control of secretion underlies pathogenic conditions such CT and TcdA- induced diarrhoeal disease, which remain major healthcare concerns in both the developing and developed world. Secretion is regulated by neurons in the wall of the GIT, with the output of secretion-controlling pathways located in the SMP. There is evidence however, that the MP is involved in CT –induced hypersecretion but the enteric pathways remain surprisingly undefined (See Chapter 1). It is well documented that TcdA activates extrinsic neural pathways to elicit its actions whereas enteric neural input has only been partially implicated. Further, microbial-derived metabolites such as GABA are linked with CDI susceptibility and present major complications in treatment of infection. The interaction of luminal GABA with the ENS and GABAergic transmission under physiological conditions are ambiguities that need to be resolved in order to gain an understanding of their putative role in pathological states (Chapter 1). My thesis aimed to address these issues highlighted in the literature to assist in the design of treatments to combat these often fatal hypersecretory states. First I examined the effects of luminal incubations with enterotoxins and GABA in vitro in the guinea pig small intestine, on the excitability of intrinsic sensory neurons in the MP using intracellular recording (Chapters 3 and 4). Secondly, I further defined the role of endogenous GABA in the ENS as a means to elucidate the mechanisms through which microbial metabolites may act by using immunohistochemical techniques and Wnt1- Cre;R26R-GCaMP3 mice in which enteric neurons and glia express a fluorescent calcium indicator, for calcium imaging (Chapter 5). Finally I extended the examination of the effects of enterotoxins on the enteric circuitry by incubating CT (Chapter 6) and TcdA (Chapter 7) in an in vivo ileal loop and assessed the effects on the enteric circuitry via intracellular recording and calcium imaging respectively.

In carrying out these studies a number of findings were established that contribute significantly to our understanding of the neural pathways underlying enterotoxin-induced hypersecretion. The work also raised additional, yet important, questions that are beyond the scope of this thesis and need to be addressed in future studies. Thus, this chapter addresses the implications of these novel findings and outlines experiments that should be undertaken in the future.

195

1. CT INDUCES SUSTAINED HYPEREXCITABILITY IN MYENTERIC AH NEURONS

A key and novel finding of this thesis is that CT increases excitability of myenteric AH neurons in the guinea pig small intestine (Chapter 3: Koussoulas et al., 2017). This finding extends previous work that demonstrates CT induces a sustained activity of secretomotor neurons (Gwynne et al., 2009) thus amplifying the efferent arm of the secretomotor reflex pathway. While secretion is predominantly under the control of transmission within the SMP, my results demonstrate that CT augments firing elsewhere in the circuitry specifically the afferent limb of the enteric secretory reflex in the MP in agreement with evidence that the MP is required for a response to CT (Jodal et al., 1993). I found that the effects on firing involved a 5-HT3 receptor-independent pathway, which indicates that CT induces sustained excitation at multiple sites within the circuitry and not simply at the mucosal epithelium where CT produces a large release of 5-HT to activate underlying secretomotor pathways (Nilsson et al., 1983; Bearcroft et al., 1997; Bertrand et al., 2000; Farthing, 2000). This finding is interesting since 5-HT3 antagonists block the increased excitability of secretomotor neurons (Gwynne et al., 2009), hypersecretion seen in vivo (Beubler et al., 1989b; Mourad et al., 1995; Turvill and Farthing, 1997) and alter CT-induced motility effects (Fung et al., 2010; Balasuriya et al., 2016). Thus mediators other than 5-HT released at the mucosa are likely to excite the system. Indeed an extensive array of mediators including ATP and peptide hormones such as cholecystokinin, secretin, glucagon-like peptides (GLP) 1 and 2 , neurotensin, peptide YY (PYY) are known to be contained in enteroendocrine cells in the mucosa (Cooke et al., 2003; Engelstoft et al., 2013; Gribble and Reimann, 2016; Fothergill et al., 2017). The results of Chapter 3 provide an insight into the complexity of neural mechanisms activated by CT and highlight the need to identify mediators responsible for activation of a 5-HT3-independent pathway. This would be desirable as a means to target other potential therapeutic sites that play a significant role in the reflex amplification.

In contrast to sustained activation of myenteric AH neurons, the data indicate that CT does not produce long-term changes in the firing of other myenteric neurons (S-type) including most interneurons in pathways regulating secretion or motility and the output neurons of motility pathways. We cannot rule out significant effects on some myenteric S neuron though due to the wide range of functional subtypes. Additionally the possibility that these neurons

196

undergo short-term changes, such as activation during the induction of CT’s response cannot be excluded. Certainly, in the mouse small intestine in vivo expression of activity dependent marker c-Fos indicates that myenteric neurons are likely to be active during the incubation itself (Fung et al., 2017, unpublished). It is probable that this activity coincides with early motility changes that CT is known to produce (Fung et al., 2010). Therefore it is likely that CT excites multiple pathways within the ENS to produce both short term and long term changes in neuronal firing corresponding to the induction of the response and the sustained effects of CT. Indeed it has been proposed that the induction and maintenance of CT-induced hypersecretion in the rat small intestine encompass different neural mechanisms (Kordasti et al., 2006).

2. MYENTERIC AH NEURONS ARE A COMMON PATHWAY THROUGH WHICH CT, TCDA AND GABA ACT

A major finding of this thesis is that myenteric AH neurons in the guinea pig small intestine are made hyperexcitable by the luminal application of TcdA and GABA in addition to CT (Chapters 3 and 4). While direct application of CT, TcdA and GABA excites enteric cell bodies (Jiang et al., 1993; Cherubini and North, 1984a; Xia et al., 2000), luminal incubation of these agents demonstrates the underlying enteric pathways that they activate converge on the same population of myenteric AH neurons highlighting a generality in the actions of toxins and associated bacterial metabolites in the gut. This, therefore, greatly assists our understanding of the general neural mechanisms underlying bacterially-induced diarrhoea. Further since each agent is known to induce motility changes in the rodent small intestine (Tonini et al., 1989b; Burakoff et al., 1995; Kordasti et al., 2006; Fung et al., 2010; Auteri et al., 2015) it is likely that altered firing in myenteric AH neurons also contributes to enterotoxin and GABA-induced motility effects. The identification of this key integration site for incoming luminal signals may assist in making predictions about the effects other enterotoxins will have on enteric circuitry and secretomotor output.

Whether the pathways activated by bacterial metabolites such as GABA ‘prime’ the ENS and potentiate enterotoxin responses is not yet known. Since GABA is linked to CDI susceptibility, co-administration TcdA with GABA in vitro and examining firing activity of

197

ISNs is a necessary future study. In addition, the effects of GABA on enterotoxin-induced changes in secretion and motor output in vivo would form a desirable line of experimentation as this would help determine whether any neural interaction between bacterial metabolites and enterotoxins extends to altered functional output to exacerbate the effects of toxins in whole organ responses.

3. ENDOGENOUS GABA IS INVOLVED IN SYNAPTIC TRANSMISSION

A very interesting novel finding in this thesis is that GABAC receptors have a synaptic function and modulate synaptic responses to trains of stimuli which have been shown to produce sEPSPs in the murine circuitry (Nurgali et al., 2004; Mao et al., 2006), where GABA acting at GABAC receptors tonically inhibits the system in the small intestine (Chapter 5).

While I observed that GABAA, GABAB and GABAC receptors are expressed in the ENS via direct application of GABA, consistent with data showing their presence in mouse GIT from immunohistochemical and functional studies (Zizzo et al., 2007; Casanova et al., 2009; Seifi et al., 2014; Auteri et al., 2015), when endogenous GABA is released it acts at GABAC receptors. This study is the first to demonstrate a synaptic function of GABAC receptors in the ENS.

In contrast, GABAA and GABAB receptors were surprisingly not found to have a synaptic function since GABAA and GABAB antagonists had no effect on electrically-stimulated calcium responses despite functional studies indicating that inhibiting GABAA and GABAB receptors can modify motility and reflexes in the mouse (Sanger et al., 2002; Auteri et al.,

2014). For instance, in the mouse colon GABAA-activated increases in peristaltic activity are blocked by bicuculline, while GABAB- mediated inhibition of peristalsis is antagonized by phaclofen (Auteri et al., 2014).

It may be that any endogenous GABA does not reach the relevant receptors and possibly produces only small responses. Intracellular recording experiments could be implemented to examine this. GABA-mediated post synaptic responses could be investigated via application of GABA receptor antagonists following electrical stimulation of murine ganglia. But given the technical difficulty of intracellular recording in the mouse and the relatively small 198

population of GABA-responding neurons (Chapter 5) these may prove to be challenging experiments.

GABAergic neurons were identified via immunohistochemistry as interneurons in Chapter 5 where GABA-IR varicosities were found to make close contacts with calretinin-IR neurons and nNOS-IR neurons. Thus, they are likely to be interneurons contacting motor neuron subtypes. As interneurons, it is highly probable that they participate in motor reflexes as has been shown in mouse colon. This is consistent with previous studies in the MP of the rat and murine colon and guinea pig ileum that have reported that GABAergic neurons probably function as interneurons (Jessen et al., 1986; Furness et al., 1989; Sang and Young, 1996; Krantis, 2000; Seifi et al., 2014).

Examining the physiological role of GABA in the ENS deepens our understanding of endogenous GABAergic transmission in the ENS and assists in highlighting potential mechanisms whereby luminal GABA confers disease susceptibility to CDI (Dann et al.,

2015). For instance, GABA acting at GABAC receptors, which may inhibit neural circuits as this study demonstrates, could act to decrease motility enabling prolonged contact and exposure of TcdA in the gut lumen, allowing more time for stimulation of pathogenic effects.

Blocking GABAC receptors might evacuate contents and offer a therapeutic effect. Functional motility experiments following ileal loop incubations of TcdA with GABAC antagonists would be worthwhile in investigating this.

4. THE INVOLVEMENT OF THE ENS IN CT-INDUCED HYPERSECRETION IN VIVO

In Chapter 6 of this thesis, I used intracellular recording to examine correlations between electrophysiological and morphological properties of enteric neurons in the murine small intestine and found that these are analogous to those in the widely studied guinea pig model (Bornstein et al., 1991a, 1994; Brookes et al., 2001; Furness et al., 2004a), and hence the neurons are likely to share similar functions. Following in vivo incubations of CT in the mouse ileal loop, I found that CT induces an increase in firing of spontaneous fast synaptic potentials in myenteric S neurons.

199

This study was conducted alongside another in our group that used Wnt1-Cre;R26R- GCaMP3 mice to examine the excitability of the general enteric circuitry following in vivo CT exposure with Ca2+ imaging (Fung et al., 2017, unpublished). It was found that in the mouse ileum, CT acts primarily at the mucosa to produce sustained hypersecretion and in part by inducing a sustained increase in the activity of submucosal secretomotor neurons (Fung et al., 2017, unpublished). This is in accordance with a study in the guinea pig by Gwynne et al., 2009 showing that CT in vitro produces sustained excitation of secretomotor neurons in the SMP. However, in Chapter 3, I demonstrated that equivalent in vitro incubations produce excitation in myenteric AH neurons, while in the mouse Ca2+ imaging showed an overall reduction in myenteric activity.

Thus it appears that in both animal models CT produces excitation within enteric neural networks that in part likely drives the ongoing hypersecretion, but the neural contributions in the two models differ with a strong mucosal effect observed in mice which have less ENS input compared to the guinea pig.

While the properties of enteric neurons are certainly similar between the two species, differences in circuit connectivity and hence signal transduction and network excitation may and account for the differences in neural contribution. In agreement with other electrophysiological studies in the mouse small intestine (Bian et al., 2003; Ren et al., 2003), and in contrast to guinea pig, AH neurons were rarely encountered by electrophysiology in mouse myenteric plexus and their low numbers may have implications for excitatory reinforcement of recurrent network activity (Thomas et al., 2000; Chambers et al., 2005). Further to this AH/type II neurons are not known to be present in the SMP of the mouse (Mongardi Fantaguzzi et al., 2009; Foong et al., 2014). Moreover, neural contribution in the mouse may have been under-represented in Chapter 6. While the spontaneous fEPSPs recorded in S neurons are probably too small to contribute to much more than a background effect, this observation implies heightened activity in other neurons elsewhere in the circuitry. While still inconclusive, changes in AH neuronal firing certainly suggest this but sampling challenges restrict this conclusion. Further the sampling bias excluded descending interneurons and inhibitory motor neurons from the data set and possibly discounted significant effects in their excitability. Whether or not interneurons and inhibitory motor neurons display sustained excitation after CT is an important point to address in future studies. This bias could be overcome by deliberately impaling specific subsets of neurons e.g.

200

nNOS neurons that express GFP. Moreover, since incubations in mice were performed in vivo while those in guinea pig were in vitro, differences in neural contributions of the ENS to the hypersecretion may be due these technical differences. Ultimately, in vivo incubations would have been required in guinea pigs to determine this.

In addition, other experiments to further define the neural mechanisms excited by CT in the mouse could include those that examine the mode of action of enhanced activity of secretomotor neurons in the SMP. Whether these neurons are activated by mucosal mediators such as 5-HT as shown in the guinea pig (Gwynne et al., 2009) could be determined by luminal incubations of CT with antagonists such as granisetron.

5. THE INVOLVEMENT OF THE ENS IN TCDA-INDUCED HYPERSECRETION IN VIVO

The findings in Chapter 7 demonstrate that TcdA incubated in vivo in Wnt1-Cre;R26R- GCaMP3 mice produces an overall reduction in the excitability of the SMP, but activates enteric circuitry in regions away from the site of acute toxin exposure.

TcdA evokes its intestinal effects via a strong extrinsic component (Castagliuolo et al., 1994), as such its neural mechanism of action may be more complex than that of CT. Others have demonstrated direct excitatory effects of the toxin on enteric neurons (Xia et al., 2000) but this is the first study to survey the enteric nervous system after luminal exposure to TcdA. From this work it is evident that enteric neurons are activated by TcdA; while this work requires significant additional investigation, it sets a framework for future inquiry and in doing so has also highlighted some salient experimental considerations for future researchers.

At the ileal site of acute exposure, the TcdA produced a reduction in neural activity in the SMP and no change in myenteric activity despite producing extensive signs of inflammation, which in and of itself is known to excite enteric neurons in both plexuses (Linden et al., 2003; Lomax et al., 2005, 2006). I attributed this potentially to the gross damage produced by TcdA at the mucosa leading to a reduction in excitability or lack of change in activity in the underlying plexuses. While the concentration of toxin used in this study may trigger extrinsic neural and inflammatory responses, my results emphasise the need to identify a lower and

201

more operative concentration for use when examining enteric neural activity with prolonged incubations.

Due to the systemic inflammation observed in Chapter 7, TcdA exposure might have produced widespread effects in other organs. Indeed C. difficile is known to produce systemic disease in conjunction with gastrointestinal effects. Systemic complications include pleural effusion, cardiopulmonary arrest and acute respiratory distress in humans and in animal models (Jacobs et al., 2001; Steele et al., 2012). The exact mechanisms by which C. difficile induces these systemic effects is not well characterised, but may be due to systemic toxin uptake as well as elevations in circulating inflammatory cytokines (Steele et al., 2012). This underscores the importance of a more effective dose and is an important consideration when interpreting data in TcdA animal models

Enteric neurons were activated by TcdA as indicated by expression of activity dependent markers in regions off-target the site of acute toxin exposure. Off-target regions also displayed significant signs of inflammation. Since TcdA is known to release inflammatory cytokines across the mucosal epithelium (Triadafilopoulos et al., 1989; Pothoulakis and Lamont 2001), their entry into the general circulation and subsequent activation of inflammation or of enteric neurons themselves (Dekkers et al., 1997a; Liu et al., 2003; Lakhan and Kirchgessner, 2010) is quite possible. Ileal loop incubations of TcdA with anti- inflammatory agents such as cyclooxygenase (COX) inhibitors or with agents that reduce cytokine synthesis should be performed to determine the nature of enteric activation.

Further it would be desirable to ascertain the identities of the activated off-target neurons by immunohistochemically identifying neuronal subtypes and whether the activation of enteric neurons is a sustained effect. This could be achieved by examining any ongoing activity of off -target tissue using Ca2+ imaging.

6. A FINAL WORD

This thesis has presented studies that contribute to our understanding of the neural mechanisms underlying enterotoxin-induced hypersecretion in the MP of the guinea pig ileum (Chapters 3 and 4). Work on the role of GABA in enteric transmission as a means to

202

shed light on the interactions of bacterial metabolites with the ENS has provided major insights into the function of GABA and its receptors in the MP (Chapter 5). Elements of the enteric circuitry involved in CT and TcdA- induced pathogenic effects in vivo have been identified, while additional investigation will undoubtedly be required, these studies have established foundations for future lines of research.

203

LIST OF REFERENCES

Aktories, K. and Just, I. (1995). Monoglucosylation of low-molecularmass GTP-binding Rho proteins by clostridial cytotoxin. Trends Cell Biol 5: 441–443. Al-Barrak, A., Embil, J., Dyck, B., Olekson, K., Nicoll, D., Alfa, M. and Kabani, A. (1999). An outbreak of toxin A negative, toxin B positive Clostridium difficile-associated diarrhea in a Canadian tertiary-care hospital. Canada communicable disease report Relevé des maladies transmissibles au Canada 25: 65. Alex, G., Kunze, W., Furness, J. and Clerc, N. (2001). Comparison of the effects of neurokinin-3 receptor blockade on two forms of slow synaptic transmission in myenteric AH neurons. Neuroscience 104: 263-269. Alfa, M., Kabani, D., Lyerly, S., Moncrief, L., Neville, A., Al-Barrak, G., Harding, B., Olekson, D. and Embil, J. (2000). Characterization of a toxin A-negative, toxin B- positive strain of Clostridium difficile responsible for a nosocomial outbreak of Clostridium difficile-associated diarrhea. J Clin Microbiol 38: 2706–2714. Ameen, N., Ardito, T., Kashgarian, M. and Marino, C. (1995). A unique subset of rat and human intestinal villus cells express the cystic fibrosis transmembrane conductance regulator. Gastroenterology 108: 1016-1023. Amin, J. and Weiss, D. (1994). Homomeric ρ1 GABA channels: activation properties and domains. Receptors and Channels 2: 227–236. Ammerman, N., Beier-Sexton, M. and Azad, A. (2008). Growth and Maintenance of Vero Cell Lines. Curr Protoc Microbiol APPENDIX: Appendix–4E. Anitha, M., Vijay-Kumar, M., S.V. Sitaraman, S., Gewirtz, A. and Srinivasan, S. (2012). Gut microbial products regulate murine gastrointestinal motility via Toll-like receptor 4 signaling Gastroenterology 143: 1006-1016. Auteri, M., Zizzo, M., Mastropaolo, M. and Serio, R. (2014). Opposite role played by GABA A and GABA B receptors in the modulation of peristaltic activity in mouse distal colon. European journal of pharmacology 731: 93-99. Auteri, M., Zizzo, M. and Serio, R. (2015). GABA and GABA receptors in the gastrointestinal tract: from motility to inflammation. Pharmacological research 93: 11-21.

204

Balasuriya, G., Hill-Yardin, E., Gershon, M. and Bornstein, J. (2016). A sexually dimorphic effect of cholera toxin: rapid changes in colonic motility mediated via a 5-HT3 receptor-dependent pathway in female C57Bl/6 mice. Journal of Physiology 594: 4325-4338. Banks, M., Farthing, M., Robberecht, P. and Burleigh, D. (2005). Antisecretory actions of a novel vasoactive intestinal polypeptide (VIP) antagonist in human and rat small intestine. British Journal of Pharmacology 144: 994-1001. Barber, D., Buchan, A., Walsh, J. and Soll, A. (1986). Regulationof neurotensin release from canine enteric primary cell cultures. Amer. J. Physiol 250: G385–G390. Barreto, A., Cavalcante, I., Castro, M., Junqueira, A., Vale, M., Ribeiro, R. and Brito, G. (2008). Fucoidin prevents Clostridium difficile toxin-A-induced ileal enteritis in mice. Digestive diseases and sciences 53: 990-996. Barrett, E., Ross, R., O’toole, P., Fitzgerald, G. and Stanton, C. (2012). γ-Aminobutyric acid production by culturable bacteria from the human intestine. J. Appl. Micro 113: 411– 417. Bartho, L., Sebok, B. and Szolcsanyi, J. (1982). Indirect evidence for the inhibition of enteric substance P neurones by opiate antagonists but not by capsaicin. Eur J Pharmacol 77: 273-279. Bayer, S., Jellali, A., Crenner, F., Aunis, D. and Angel, F. (2003). Functional evidence for a role of GABA receptors in modulating nerve activities of circular smooth muscle from rat colon in vitro. LifeSci 72: 1481–1493. Bearcroft, C., Andre, E. and Farthing, J. (1997). In vivo effects of the 5-HT3 antagonist alosetron on basal and cholera-induced secretion in the human jejunun: a segmental perfusion study. Aliment. Pharmacol. Ther 11: 1109–1114. Bearcroft, C., Perret, D. and Farthing, M. (1996). 5-Hydroxytryptamine release into human jejunum by cholera toxin. Gut 39: 528-531. Bertrand, P. and Bornstein, J. (2002). ATP as a putative sensory mediator: activation of intrinsic sensory neurons of the myenteric plexus via P2X receptors. J Neurosci 22: 4767–4775. Bertrand, P. and Galligan, J. (1994). Contribution of chloride conductance increase to slow EPSC and tachykinin current in guinea pig myenteric neurones. J. Physiol. (Lond.) 481: 47-60.

205

Bertrand, P. and Galligan, J. (1995). Signal-transduction pathways causing slow synaptic excitation in guinea pig myenteric AH neurones. Am. J. Physiol 269: 710-720. Bertrand, P., Kunze, W., Bornstein, J. and Furness, J. (1998). Electrical mapping of the projections of intrinsic primary afferent neurons to the mucosa of the guinea pig small intestine. Neurogastroenterol. Motil 10: 533-541. Bertrand, P., Kunze, W., Bornstein, J., Furness, J. and Smith, M. (1997). Analysis of the responses of myenteric neurons in the small intestine to chemical stimulation of the mucosa. Am. J. Physiol 273: 422-435. Bertrand, P., Kunze, W., Furness, J. and Bornstein, J. (2000). The terminals of myenteric intrinsic primary afferent neurons of the guinea pig ileum are excited by 5- hydroxytryptamine acting at 5-hydroxytryptamine-3 receptors. Neuroscience 101: 459-469 Bertrand, P. and Thomas, E. (2004). Multiple levels of sensory integration in the intrinsic sensory neurons of the enteric nervous system. Clin. Exp. Pharmacol. Physiol 31: 745-755. Best, E., Freeman, J. and Wilcox, M. (2012). Models for the study of Clostridium difficile infection. Gut microbes 3: 145-167. Beubler, E., Kollar, G., Saria, A., Bukhave, K. and Raskmadsen, J. (1989b). Involvement of 5-Hydroxytryptamine, prostaglandin E2, and cyclic adenosine-monophosphate in cholera toxin-induced fluid secretion in the small intestine of the rat in vivo. Gastroenterology 96: 368-376. Bhat, R., Axtell, R., Mitra, A., Miranda, M., Lock, C., Tsien, R. and Steinman, L. (2010). Inhibitory role for GABA in autoimmune inflammation. Proceedings of the National Academy of Sciences of the United States of America 107: 2580–2585. Bian, X., Ren, J., Vries, M., Schnegelsberg, B., Cockayne, D., Ford, A. and Galligan, J. (2003). Peristalsis is impaired in the small intestine of mice lacking the P2X3 subunit. The Journal of physiology 551: 309-322. Bjerknes, M. and Cheng, H. (2001). Modulation of specific intestinal epithelial progenitors by enteric neurons. Proceedings of the National Academy of Sciences of the United States of America 98: 12497-12502. Boesmans, W., Martens, M., Weltens, N., Hao, M., Tack, J., Cirillo, C. and Vanden Berghe, P. (2013). Imaging neuron-glia interactions in the enteric nervous system. Front Cell Neurosci 7: 183. 206

Bornstein, J., Costa, M. and Furness, J. (1986). Synaptic inputs to immunohistochemically identified neurones in the submucous plexus of the guinea-pig small intestine. Journal of Physiology 381: 465-482. Bornstein, J., Costa, M., Furness, J. and Lees, G. (1984). Electrophysiology and enkephalin immunoreactivity of identified myenteric plexus neurones of guinea pig small intestine. J. Physiol. (Lond.) 351: 313-325. Bornstein, J., Costa, M. and Grider, J. (2004). Enteric motor and interneuronal circuits controlling motility. Neurogastroenterology and Motility 16 Suppl 1: 34-38. Bornstein, J., Costa, M. and Jb, F. (1988). Intrinsic and extrinsic inhibitory synaptic inputs to submucous neurones of the guinea-pig small intestine. Journal of Physiology 398: 371-390. Bornstein, J., Furness, J. and Costa, M. (1987). Sources of excitatory synaptic inputs to neurochemically identified submucous neurons of guinea-pig small intestine. Journal of the Autonomic Nervous System 18: 83-91. Bornstein, J., Furness, J. and Costa, M. (1989). An electrophysiological comparison of substance P-immunoreactive neurons with other neurons in the guinea pig submucous plexus. J.Auton. Nerv. Syst 26: 113-120. Bornstein, J., Furness, J. and Kunze, W. (1994). Electrophysiological characterization of myenteric neurons: how do classification schemes relate? Journal of the Autonomic Nervous System 48: 1-15. Bornstein, J., Furness, J., Kunze, W. and Bertrand, P. 2002. Enteric reflexes that influence motility. In: Nervous control of the gastrointestinal tract, London, UK, Taylor & Francis. Bornstein, J., Furness, J., Smith, T. and Trussell, D. (1991a). Synaptic responses evoked by mechanical stimulation of the mucosa in morphologically characterized myenteric neurons of the guinea-pig ileum. Journal of Neuroscience 11: 505-518. Bornstein, J., Gwynne, R. and Sjövall, H. 2012. Enteric Neural Regulation of Mucosal Secretion, Boston: Academic Press. Bornstein, J., Hendriks, R., Furness, J. and Trussel, D. (1991b). Ramifications of the axons of AH neurons injected with the intracellular marker biocytin in the myenteric plexus of the guinea pig small intestine. J. Comp. Neurol 314: 437-451.

207

Bornstein, J., Low, A. and Furness, J. (1992). Electrophysiological identification of calbindin immunoreactive myenteric neurons in the guinea-pig distal colon. Proc. Aust. Neurosci. Soc 3: 189. Brehmer, A. (2007). The value of neurofilament-immunohistochemistry for identifying enteric neuron types—special reference to intrinsic primary afferent (sensory) neurons. In: Arlen RK (ed) New research on neurofilament proteins. Nova Science, New York: 99-114. Brookes, S. (2001). Classes of enteric nerve cells in the guinea pig small intestine. Anat. Rec. 262: 58-70. Brookes, S. and Costa, M. 2002. Cellular organisation of the mammalian enteric nervous system.

Brookes, S., Ewart, W. and Wingate, D. (1987). Intracellular recordings from myenteric neurones in the human colon. The Journal of physiology 390: 305-318.

Brookes, S., Ewart, W. and Wingate, D. (1988). Intracellular recordings from cells in the myenteric plexus of the rat duodenum. Neuroscience 24: 297-307. Brookes, S., Hennig, G. and Schemann, M. (1998). Identification of motor neurons to the circular muscle of the guinea pig gastric corpus. J Comp Neurol 397: 268–280. Brookes, S., Steele, P. and Costa, M. (1991a). Calretinin immunoreactivity in cholinergic motor neurones, interneurones and vasomotor neurones in the guinea-pig small intestine. Cell Tissue Res 263: 471–481. Brookes, S., Steele, P. and Costa, M. (1991b). Identification and immunohistochemistry of cholinergic and non-cholinergic circular muscle motor neurons in the guinea-pig small intestine. Neuroscience 42: 863–878. Brown, D. and Passmore, G. (2009). Neural KCNQ (Kv7) channels. British Journal of Pharmacology 156: 1185–1195. Browning, K. and Lees, G. (1996). Myenteric neurons of the rat descending colon: electrophysiological and correlated morphological properties. Neuroscience 73: 1029–1047. Browning, K. and Travagli, R. (2014). Central Nervous System Control of Gastrointestinal Motility and Secretion and Modulation of Gastrointestinal Functions. Comprehensive Physiology 4: 1339–1368.

208

Brugger, F., Olpe, H., Froestl, W. and Mickel, S. (1993). The action of new potent GABAB receptor antagonists in the hemisected spinal cord preparation of the rat. European Journal of Pharmacology 235: 153-155 Burakoff, R., Zhao, L., Celifarco, A., Rose, K., Donovan, V., Pothoulakis, C. and Percy, W. (1995). Effects of Purified Clostridium difficile Toxin A on Rabbit Distal Colon. Gastroenterology 109: 348-354. Burleigh, D. and Banks, M. (2007). Stimulation of intestinal secretion by vasoactive intestinal peptide and cholera toxin. Autonomic Neuroscience: Basic and Clinical 133: 64-75. Burrows, W. and Musteikis, G. (1966). Cholera Infection and Toxin in the Rabbit Ileal Loop. The Journal of Infectious Diseases 116: 183-190. Caren, J., Meyer, J. and Grossman, M. (1974). Canine intestinal secretion during and after rapid distension of the small bowel. Am. J. Physiol 227: 183-188. Carey, H. and Cooke, H. (1986). Submucosal nerves and cholera toxin-induced secretion in guinea pig ileum in vitro. Digestive Diseases and Sciences 31: 732-736. Carter, G., Chakravorty, A., Nguyen, T., Mileto, S., Schreiber, F., Li, L. and Cheknis, A. (2015). Defining the roles of TcdA and TcdB in localized gastrointestinal disease, systemic organ damage, and the host response during Clostridium difficile infections. MBio 6: e00551-15. Casanova, E., Guetg, N., Vigot, R., Seddik, R., Julio-Pieper, M., Hyland, N., Cryan, J., Gassmann, M. and Bettler, B. (2009). Mouse model for visualization of GABAB receptors. Genesis 47: 595-602. Cassuto, J., Fahrenkrug, J. and Jodal, M. (1981b). The release of vasoactive intestinal polypeptide from the cat small intestine exposed to cholera toxin. Gut 22: 958-963. Cassuto, J., Jodal, M. and Lundgren, O. (1982a). The effect of nicotinic and muscarinic receptor blockade on cholera toxin induced secretion in rats and cats. Acta. Physiol. Scandinavica 114: 573-577. Cassuto, J., Jodal, M., Tuttle, R. and Lundgren, O. (1981a). On the role of intramural nerves in the pathogenesis of cholera toxin-induced intestinal secretion. Scandinavian Journal of Gastroenterology 16: 377-384. Castagliuolo, I., Lamont, J., Letourneau, R., Kelly, C., O'keane, J., Jaffer, A., Theoharides, T. and Pothoulakis, C. (1994). Neuronal involvement in the intestinal effects of

209

Clostridium difficile toxin A and Vibrio cholerae enterotoxin in rat ileum. Gastroenterology 107: 657-665. Castagliuolo, I., Riegler, M., Pasha, A., Nikulasson, S., Lu, B., Gerard, C., Gerard, N. and Pothoulakis, C. (1998a). Neurokinin-1 (NK-1) receptor is required in Clostridium difficile-induced enteritis. Journal of Clinical Investigation 10: 1547. Castelucci, P., Robbins, H., Poole, D. and Furness, J. (2002). The distribution of purine P2X2 receptors in the guinea pig enteric nervous system. Histochem. Cell Biol 117: 415– 422. CDC (2013). Antibiotic Resistance Threats in the United States. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention. www.cdc.gov/drugresistance. Chambers, J., Bornstein, J., Gwynne, R., Koussoulas, K. and Thomas, E. (2014). A detailed, conductance-based computer model of intrinsic sensory neurons of the gastrointestinal tract. Am. J. Physiol Gastrointest Liver Physiol 307: 517-532. Chambers, J., Bornstein, J., Sjövall, H. and Thomas, E. (2005). Recurrent networks of submucous neurons controlling intestinal secretion: a modeling study. Am. J. Physiol. Gastrointest. Liver Physiol 288: 887-896. Chavas, J. and Marty, A. (2003). Coexistence of excitatory and inhibitory GABA synapses in the cerebellar interneuron network Journal of Neuroscience 23: 2019-2031. Chen, Z., Fei, Y., Anderson, C., Wake, K., Miyauchi, S., Huang, W., Thwaites, D. and Ganapathy, V. (2003). Structure, function and immunolocalization of a proton- coupled amino acid transporter (hPAT1) in the human intestinal cell line Caco-2. J. Physiol 546: 349–361. Chen, Z., Suntres, Z., Palmer, J., Guzman, J., Javed, A., Xue, J., Yu, J., Cooke, H., Awad, H., Hassanain, H. and Cardounel, A. J. (2007). Cyclic AMP signaling contributes to neural plasticity and hyperexcitability in AH sensory neurons following intestinal Trichinella spiralis-induced inflammation. . International Journal for Parasitology 37: 743-761. Cherubini, E. and North, R. (1984a). Actions of γ‐aminobutyric acid on neurones of guinea‐ pig myenteric plexus. British journal of pharmacology 82: 93-100. Cho, H., Robinson, E., Rivera, L., Mcmillan, P., Testro, A. and Nikfarjam, M. (2014). Glucagon-like peptide 1 and peptide YY are in separate storage organelles in enteroendocrine cells. Cell Tissue Res 357. 210

Christofi, F., Wunderlich, J., Yu, J., Wang, Y., Xue, J., Guzman, J., Javed, N. and Cooke, H. (2004). Mechanically evoked reflex electrogenic chloride secretion in rat distal colon is triggered by endogenous nucleotides acting at P2Y1, P2Y2, and P2Y4 receptors. The Journal of Comparative Neurology 469: 16-36. Clerc, N., Furness, J., Bornstein, J. and Kunze, W. (1998). Correlation of electrophysiological and morphological characteristics of myenteric neurons of the duodenum in the guinea-pig Neuroscience 82: 899–914. Clerc, N., Furness, J., Kunze, W., Thomas, E. and Bertrand, P. (1999). Long term effects of synaptic activation at low frequency on excitability of myenteric AH neurons. Neuroscience 90: 279-289. Cloud, J., Noddin, L., Pressman, A., Hu, M. and Kelly, C. (2009). Clostridium difficile strain NAP-1 is not associated with severe disease in a nonepidemic setting. Clin Gastroenterol Hepatol 7: 868–873. Cooke, H. (1998). Enteric tears”: chloride secretion and its neural regulation. Physiology 13: 269-274. Cooke, H., Sidhu, M. and Wang, Y. (1997a). 5-HT activates neural reflexes regulating secretion in the guinea-pig colon. Neurogastroenterol. Motil 9: 181–186. Cooke, H., Sidhu, M. and Wang, Y. (1997b). Activation of 5-HT1P receptors on submucosal afferents subsequently triggers VIP neurons and chloride secretion in the guinea-pig colon. J. Auton. Nerv. Syst 66: 105–110. Cooke, H., Wunderlich, J. and Christofi, F. (2003). "The force be with you": ATP in gut mechanosensory transduction. News in Physiological Sciences 18: 43-49. Cooke, H., Xue, J., Yu, J., Wunderlich, J., Wang, Y., Guzman J, Javed, N. and Christofi, F. (2004). Mechanical stimulation releases nucleotides that activate P2Y1 receptors to trigger neural reflex chloride secretion in guinea pig distal colon. The Journal of Comparative Neurology 469: 1-15. Cooper, E. (2011). Made for "anchorin": Kv7.2/7.3 (KCNQ2/KCNQ3) channels and the modulation of neuronal excitability in vertebrate axons. Semin Cell Dev Biol 22: 185–192. Copel, C., Osorio, N., Crest, M., Gola, M., Delmas, P. and Clerc, N. (2009). Activation of neurokinin 3 receptor increases Na(v)1.9 current in enteric neurons. J. Physiol. 587: 1461-1467.

211

Cornelissen, W., De Laet, A., Kroese, A., Van Bogaert, P., Scheuermann, D. and Timmermans, J. (2000). Electrophysiological features of morphological Dogiel type II neurons in the myenteric plexus of pig small intestine. J Neurophysiol 84: 102–111. Cornelissen, W., De Laet, A., Kroese, A., Van Bogaert, P., Scheuermann, D. and Timmermans, J. (2001). Excitatory synaptic inputs on myenteric Dogiel type II neurones of the pig ilem. J. Comp. Neurol 432: 137-154. Corthier, G., Muller, M., Wilkins, T., Lyerly, D. and L’haridon, R. (1991). Protection against experimental pseudomembranous colitis in gnotobiotic mice by use of monoclonal antibodies against Clostridium difficile toxin A. Infect Immun 59: 1192–1195. Costa, M., Brookes, S. and Hennig, G. (2000). Anatomy and physiology of the enteric nervous system. Gut 47(suppl 4): iv15-iv19. Costa, M., Brookes, S., Steele, P., Gibbins, I., Burcher, E. and Kandiah, C. (1996). Neurochemical classification of myenteric neurons in the guinea-pig ileum. Neuroscience 75: 949-967. Costa, M., Furness, J., Cuello, A., Verhofstad, A., Steinbusch, H. and Elde, R. (1982). Neurons with 5-hydroxytryptamine-like immunoreactivity in the enteric nervous system: their visualization and reactionsto drug treatment. Neuroscience 7: 351–363. Costa, M., Jb, F., Llewellyn Smith, I. and Cuello, A. (1981). Projections of substance P- containing neurons within the guinea-pig small intestine. Neuroscience 6: 411–424. Dann, S., Aitken, S., Ross, C., Tessier, M., Loeffelholz, M., Koussoulas, K., Bornstein, J., Feng, H., Versalovic, J., Pothoulakis, C., Garey, K. and Savidge, T. (2015). Zolpidem Confers Disease Susceptibility to Clostridium difficile Infection. Gastroenterology 148: S727–S728. Davenger, S., Hjelle, O., Babaie, E., Larsson, L., Hougaard, D., Storm-Mathisen, J. and Ottersen, O. (1994). Colocalization of gamma-aminobutyrate and gastrin in the rat antrum: an immunocytochemical and in situ hybridization study. Gastroenterology 107: 137–148. De Araújo Junqueira, A., Dias, A., Vale, M., Spilborghs, G., Bossa, A., Lima, B. and Brito, G. (2011). Adenosine deaminase inhibition prevents Clostridium difficile toxin A- induced enteritis in mice. Infection and immunity 79: 653-662. Dekkers, J., Akkermans, L. and Kroese, A. (1997a). Effects of the inflammatory mediator prostaglandin E2 on myenteric neurons in guinea pig ileum. Am J Physiol Gastrointest Liver Physiol 272: G1451–G1456. 212

Del Carlo, B., Pellegrini, M. and Pellegrino, M. (2003). Modulation of Ca2+-activated K+ channels of human erythrocytes by endogenous protein kinase C. Biochim. Biophys. Acta 1612: 107-116. Dogiel, A. (1895b). Zur Frage uber die Ganglion der Darmgefl echte bei den Saugetieren. Anat. Anz 10: 517–528. Dogiel, A. (1899). Über den Bau der Ganglien in den Geflechten des Darmes und der Gallenblase des Menschen und der Säugetiere. Arch Anat Physiol Leipzig Anat Abt Jg 1899: 130-158. Drudy, D., Harnedy, N., Fanning, S., O’mahony, R. and Kyne, L. (2007). Isolation and characterisation of toxin A-negative, toxin B-positive Clostridium difficile in Dublin, Ireland. Clin Microbiol Infect 13: 298–304. Dupont, H. (2014). Evidence for the role of gut microbiota in irritable bowel syndrome and its potential influence on therapeutic targets. Alimentary pharmacology & therapeutics 39: 1033-1042. Eastwood, K., Else, P., Charlett, A. and Wilcox, M. (2009). Comparison of Nine Commercially Available Clostridium difficile Toxin Detection Assays, a Real-Time PCR Assay for C. difficile TcdB, and a Glutamate Dehydrogenase Detection Assay to Cytotoxin Testing and Cytotoxigenic Culture Methods. J. Clin. Microbiol 47: 3211- 3217. Eklund, S., Jodal, M., Lundgren, O. and Sjöqvist, A. (1979). Effects of vasoactive intestinal polypeptide on blood flow, motility and fluid transport in the gastrointestinal tract of the cat. Acta Physiol. Scand 105: 461–468. Eklund, S., Karlström, L., Rökaeus, A., Theodorsson, E., Jodal, M. and Lundgren, O. (1989). Effects of cholera toxin, Escherichia coli heat stable toxin and sodium deoxycholate on neurotensin release from the ileum in vivo. . Regulatory Peptides 26. Ellis, M., Chambers, J., Gwynne, R. and Bornstein, J. (2013). Serotonin and cholecystokinin mediate nutrient-induced segmentation in guinea pig small intestine. American Journal of Physiology Gastrointestinal and Liver Physiology 304: G749-761. Engelstoft, M., Egerod, K., Lund, M. and Schwartz, T. (2013). Enteroendocrine cell types revisited. Current Opinion in Pharmacology 13: 912-921. Enz, R. (2001). GABA(C) receptors: a molecular view. Biol Chem 382: 1111-22. Erben, U., Loddenkemper, C., Doerfel, K., Pieckermann, S., Haller, D., Heimesaat, M. and Kühl, A. (2014). A guide to histomorphological evaluation of intestinal inflammation 213

in mouse models. International journal of clinical and experimental pathology 7: 4557. Erdo, S. and Bowery, N. 1986. GABAergic Mechanisms in the MammalianPeriphery, New York, Raven.

Erdö, S., Ezer, E., Matuz, J., Wolff, J. and Amenta, F. (1989). GABAA receptors in the rat stomach may mediate mucoprotective effects. European journal of pharmacology 165: 79-86.

Evans, R., Jiang, M. and Surprenant, A. (1994). Morphological properties and projections of electrophysiologically characterized neurons in the guinea-pig submucosal plexus. Neuroscience 59: 1093-1110. Farthing, M. (2000). Enterotoxins and the enteric nervous system: a fatal attraction. Int. J. Med. Microbiol 290: 491-496. Farthing, M. (2002). Novel targets for the control of secretory diarrhoea. Gut 50: Suppl 3:Iii15-18. Fields, R., Eshete, F., Stevens, B. and Itoh, K. (1997). Action potential-dependent regulation of gene expression: temporal specificity in Ca2+, cAMP-responsive element binding proteins, and mitogen-activated protein kinase signaling. Journal of Neuroscience 17: 7252-7266. Fletcher, E., Clark, M. and Furness, J. (2002). Neuronal and glial localization of GABA transporter immunoreactivity in the myenteric plexus. Cell Tissue Res 308: 339–346. Fletcher, E., Clark, M., Senior, P. and Furness, J. (2001). Gene expression and localization of GABA(C) receptors in neurons of the rat gastrointestinal tract. Neuroscience 107: 181–189. Fletcher, E., Koulen, P. and Wässle, H. (1998). GABAA and GABAC receptors on mammalian rod bipolar cells J. Comp. Neurol 396: 351-365. Foltz, M., Ansems, P., Schwarz, J., Tasker, M., Lourbakos, A. and Gerhardt, C. (2008). Protein hydrolysates induce CCK release from enteroendocrine cells and act as partial agonists of the CCK1 receptor. Journal of agricultural and food chemistry 56: 837- 843. Foong, J. and Bornstein, J. (2009). mGluR1 receptors contribute to non-purinergic slow excitatory transmission to submucosal VIP neurons of guinea-pig ileum. Frontiers in neuroscience 3. 214

Foong, J., Hirst, C., Hao, M., Mckeown, S., Boesmans W, Young, H., Bornstein Jc and Vanden Berghe, P. (2015). Changes in nicotinic neurotransmission during enteric nervous system development. Journal of Neuroscience 35: 7106-7115. Foong, J., Nguyen, T., Furness, J., Bornstein, J. and Young, H. (2012). Myenteric neurons of the mouse small intestine undergo significant electrophysiological and morphological changes during postnatal development. Journal of Physiology 590: 2375-2390. Foong, J., Parry, L., Gwynne, R. and Bornstein, J. (2010b). 5-HT(1A), SST(1), and SST(2) receptors mediate inhibitory postsynaptic potentials in the submucous plexus of the guinea pig ileum. American Journal of Physiology Gastrointestinal and Liver Physiology 298: G384-394. Foong, J., Tough, I., Cox, H. and Bornstein, J. (2014). Properties of cholinergic and non- cholinergic submucosal neurons along the mouse colon. Journal of Physiology 592: 777-793. Fothergill, L., Callaghan, B., Hunne, B., Bravo, D. and Furness, J. (2017). Co-storage of enteroendocrine hormones evaluated at the cell and subcellular levels in male mice. Endocrinology 158: 2113–2123. Frieling, T., Hj, C. and Wood, J. (1994). Neuroimmune communication in the submucous plexus of guinea-pig colon after sensitization to milk antigen. Am J Physiol 267: G1087–93. Frigo, G., Galli, A., Lecchini, S. and Marcoli, M. (1987). A facilitatory effect of bicuculline on the enteric neurones in the guinea-pig isolated colon. Br J Pharmacol 90: 31–41. Fung, C., Ellis, M. and Bornstein, J. (2010). Luminal cholera toxin alters motility in isolated guinea pig jejunum via a pathway independent of 5-HT3 receptors. Front. Neurosci 4: 1-8. Furness, J. (2000). Types of neurons in the enteric nervous system. . Journal of the Autonomic Nervous System 81. Furness, J. (2006). The Enteric Nervous System. Blackwell Publishing, Boston. Furness, J., Alex, G., Clark, M. and Lal, V. (2003a). Morphologies and projections of defined classes of neurons in the submucosa of the guinea-pig small intestine. Anat. Rec 272A: 475–483. Furness, J., Callaghan, B., Rivera, L. and Cho, H. (2014). The enteric nervous system and gastrointestinal innervation: integrated local and central control. Advances in Experimental Medicine and Biology 817: 39-71. 215

Furness, J. and Costa, M. (1974). The adrenergic innervation of the gastrointestinal tract. Ergeb. Physiol 69: 1-51. Furness, J. and Costa, M. (1980). Types of nerves in the enteric nervous system. Neuroscience 5: 1-20. Furness, J. and Costa, M. (1982). Neurons with 5-hydroxytryptamine like immunoreactivity in the enteric nervous system: their projections in the guinea-pig small intestine. Neuroscience 7: 341–349. Furness, J. and Costa, M. (1987). The Enteric Nervous System. Churchill Livingstone, Edinburgh. Furness, J., Costa, M., Gibbins, I., Llewellyn-Smith, I. and Oliver, J. (1985). Neurochemically similar myenteric and submucous neurons directly traced to the mucosa of the small-intestine. Cell and Tissue Research 241: 155-163. Furness, J., Jones, C., Nurgali, K. and Clerc, N. (2004a). Intrinsic primary afferent neurons and nerve circuits within the intestine. Progress in Neurobiology 72: 143-164. Furness, J., Kunze, W., Bertrand, P., Clerc, N. and Bornstein, J. (1998). Intrinsic primary afferent neurons of the intestine. Prog. In Neurobiol 54: 1-18. Furness, J., Kunze, W. and Clerc, N. (1999b). Nutrient tasting and signalling mechanisms in the gut II. The intestine as a sensory organ: neural, endocrine, and immune responses. Am. J. Physiol 277: G922– G928. Furness, J., Lloyd, K., Sternini, C. and Walsh, J. (1990a). Projections of substance P, vasoactive intestinal peptide and tyrosine hydroxylase immunoreactive nerve fibres in the canine intestine, with special reference to the innervation of the circular muscle. Arch. Histol. Cytol 53: 129-140. Furness, J., Robbins, H., Xiao, J., Stebbing, M. and Nurgali, K. (2004b). Projections and chemistry of Dogiel type II neurons in the mouse colon. Cell Tissue Res 317: 1-12. Furness, J., Trussell, D., Pompolo, S., Bornstein, J., Maley, B. and Storm-Mathisen, J. (1989). Shapes and projections of neurons with immunoreactivity for gamma- aminobutyric acid in the guinea-pig small intestine. Cell and tissue research 256: 293-301. Furness, J., Trussell, D., Pompolo, S., Bornstein, J. and Smith, T. (1990b). Calbindin neurons of the guinea pig small intestine: quantitative analysis of their numbers and projections. Cell. Tissue. Res 260: 261-272.

216

Furukawa, K., Taylor, G. and Bywater, R. (1986). An intracellular study of myenteric neurons in the mouse colon. J Neurophysiol 55: 395–1406. Gabella, G. and Trigg, P. (1984). Size of neurons and glial cells in the enteric ganglia of mice, guinea-pigs, rabbits and sheep. Journal of Neurocytology 13: 49-71. Galligan, J., North, R. and Tokimasa, T. (1989b). Muscarinic agonists and potassium currents in guinea-pig myenteric neurones. Br. J. Pharmacol 96: 193–203. Galligan, J., Suprenant, A., Tonini, M. and North, R. (1988). Differential localization of 5- HT1 receptors on myenteric and submucosal neurons. Am J Physiol 255: G603-611. Galligan, J., Tatsumi, H., Shen, K., Suprenant, A. and North, R. (1990). Cation current activated by hyperpolarisation (IH) in guinea pig enteric neurons. Am. J. Physiol 259: 966-972. Gentilini, G., Franchi-Micheli, S., Pantalone, D., Cortesini, C. and Zilletti, L. (1992). GABAB receptor-mediated mechanisms in human intestine in vitro. . Eur. J. Pharmacol 217: 9–14. Gershon, M. (1999). The Second Brain: A Groundbreaking New Understanding of Nervous Disorders of the Stomach and Intestine. Harper Collins. Gershon, M. and Kirchgessner, A. (1991). Identification, characterisation and projections of intrinsic primary afferent neurones of the submucosal plexus: Activity-induced expression of c-fos immunoreactivity. J. Auton. Nerv. Syst 33: 185-187. Giotti, A., Luzzi, S., Spagnesi, S. and Zilletti, L. (1983). GABAA and GABAB receptor- mediated effects in guinea-pig ileum. Br. J. Pharmacol 78: 469–478. Girdhar, A., Dhumal, V., Gulati, O., Bhavsar, V. and Hemavathi, G. (1981). Effect of gamma-aminobutyric acid on the rabbit isolated intestine. J Pharm Pharmacol 33: 614–615. Goorhuis, A., Legaria, M., Van Den Berg, R., Harmanus, C., Klaassen, C., Brazier, J., Lumelsky, G. and Kuijper, E. (2009). Application of multiple-locus variable-number tandem-repeat analysis to determine clonal spread of toxin A-negative Clostridium difficile in a general hospital in Buenos Aires, Argentina. Clin Microbiol Infect 15: 1080–1086. Grafe, P., Mayer, C. and Wood, J. (1980). Synaptic modulation of calciumdependent potassium conductance in myenteric neurones in the guinea pig. J. Physiol. 305: 235–248.

217

Greenwood, B. and Davison, J. (1987). The relationship between gastrointestinal motility and secretion. Am. J. Physiol 252: 1-7. Gribble, F. and Reimann, F. (2016). Enteroendocrine Cells: Chemosensors in the Intestinal Epithelium. Annual Review of Physiology 78: 277-299. Grider, J. (2003). Neurotransmitters mediating the intestinal peristaltic reflex in the mouse. J Pharmacol Exp Ther 307: 460–7. Grider, J., Kuemmerle, J. and Jin, J. (1996). 5-HT released by mucosal stimuli initiates peristalsis by activating 5-HT4/5-HT1p receptors on sensory CGRP neurons. Am. J. Physiol 270: G778–G82. Grider, J. and Makhlouf, G. (1992). Enteric GABA: mode of action and role in the regulation of the peristaltic reflex. Am. J. Physiol 262: G690–G694. Guichard, A., Moreno, B., Aguilar, B., Van Sorge, N., Kuang, J., Kurkciyan, A. and Bier, E. (2013). Cholera toxin disrupts barrier function by inhibiting exocyst-mediated trafficking of host proteins to intestinal cell junctions. Cell Host & Microbe 14: 294– 305. Gulledge, A. and Stuart, G. (2003). Excitatory actions of GABA in the cortex Neuron 37: 299-309. Guo, Y., Thompson, J. and Singh, P. (1989). Effect of γ-aminobutyric acid on bombesin- evoked release of somatostatin and gastrin from isolated rat stomach. Regul Pept 24: 179–186. Gwynne, R. and Bornstein, J. (2007a). Mechanisms underlying nutrient-induced segmentation in isolated guinea pig small intestine. American Journal of Physiology Gastrointestinal and Liver Physiology 292: G1162-1172. Gwynne, R. and Bornstein, J. (2007b). Synaptic transmission at functionally identified synapses in the enteric nervous system: roles for both ionotropic and metabotropic receptors. Curr. Neuropharmacol 5: 1-17. Gwynne, R. and Bornstein, J. (2009). Electrical stimulation of the mucosa evokes slow EPSPs mediated by NK1 tachykinin receptors and by P2Y1 purinoceptors in different myenteric neurons. Am J Physiol Gastrointest Liver Physiol. 297: G179-186. Gwynne, R., Clarke, A., Furness, J. and Bornstein, J. (2014). Both exogenous 5-HT and endogenous 5-HT, released by fluoxetine, enhance distension evoked propulsion in guinea-pig ileum in vitro. Neuroscience 8.

218

Gwynne, R., Ellis, M., Sjovall, H. and Bornstein, J. (2009). Cholera toxin induces sustained hyperexcitability in submucosal secretomotor neurons in guinea pig jejunum. Gastroenterology 136: 299-308. Hansen, M. (2003). The enteric nervous system II: gastrointestinal functions The enteric nervous system I: organisation and classification. Pharmacol. Toxicol 92: 249-257. Hao, M., Boesmans, W., Van Den Abbeel, V., Jennings, E., Bornstein, J., Young, H. and Vanden Berghe, P. (2011). Early emergence of neural activity in the developing mouse enteric nervous system. J Neurosci 31: 15352–15361. Hardcastle, J., Hardcastle, P. and Mathias, W. (1991). The influ¬ence of the gamma-amino butyric acid (GABA) antagonist bicuculline on transport processes in rat small intestine. J. Pharm. Pharmacol 43: 128–130. Harrington, A., Hutson, J. and Southwell, B. (2005). Immunohistochemical localization of substance P NK1 receptor in guinea pig distal colon. Neurogastroenterol Motil, 17: 727-737. Hecht, G., Koutsouris, A., Pothoulakis, C., Lamont, J. and Madara, J. (1992). Clostridium difficile toxin B disrupts the barrier function of T84 Monolayers. Gastroenterol 102: 416-423. Hecht, G., Pothoulakis, C., Lamont, J. and Madara, J. (1988). Clostridium difﬁcile toxin A perturbs cytoskeletal structure and junction permeability in cultured humanepithelial cells. J. Clin. Invest 82: 1516–1524. Hills, J., Jessen, K. and Mirsky, R. (1987). An immunohistochemical study of the distribution of enteric GABA-containing neurons in the rat and guinea-pig intestine. Neuroscience 22: 301-312. Hillsey, K., Kenyon, J. and Smith, T. (2000). Ryanodine-sensitive stores regulate the excitability of AH neurons in the myenteric plexus of guinea pig ileum. J. Neurophysiol 84: 2777-2785. Hirst, C., Foong, J., Stamp, L., Fegan, E., Dent, S., Cooper, E., Lomax, A., Anderson, C., Bornstein, J., Young, H. and Mckeown, S. J. (2015). Ion channel expression in the developing enteric nervous system. PloS one 10: p.e0123436. Hirst, G., Holman, M. and Spence, I. (1974). Two types of neurones in the myenteric plexus of duodenum in the guinea pig. J. Physiol. (Lond.) 236: 303-336. Hirst, G., Johnson, S. and Van Helden, D. (1985a). The calcium current in a myenteric neurone of the guinea pig ileum. J. Physiol. (Lond.) 361: 297-314. 219

Hirst, G., Johnson, S. and Van Helden, D. (1985b). The slow calcium-dependent potassium current in a myenteric neurone of the guinea pig ileum. J. Physiol. (Lond.) 361: 315- 337. Hirst, G. and Mckirdy, H. (1974b). Presynaptic inhibition at a mammalian peripheral synapse? . Nature 250: 430–431. Hirst, G. and Mckirdy, H. (1975). Synaptic potentials recorded from neurones of the submucous plexus of guinea-pig small intestine. Journal of Physiology 249: 369-385. Hirst, G. and Spence, I. (1973). Calcium action potentials in mammalian peripheral neurons. Nature 243: 54-56. Hodgkiss, J. and Lees, G. (1984). Slow intracellular potentials in AH neurons of the myenteric plexus evoked by repetitive activation of synaptic inputs. Neuroscience 11: 255–261. Holmgren, J., Lönnroth, I., Månsson, J. and Svennerholm, L. (1975). Interaction of cholera toxin and membrane GM1 ganglioside of small intestine. Proceedings of the National Academy of Sciences 72: 2520-2524. Hu, H., Gao, N., Zhu, M., Liu, S., Ren, J., Gao, C., Xia, Y. and Wood, J. (2003). Slow excitatory synaptic transmission mediated by P2Y1 receptors in the guinea pig enteric nervous system. J. Physiol 550: 493-504. Hubel, K. (1985). Intestinal nerves and ion transport: stimuli, reflexes and responses. Am. J. Physiol 248: 261-271. Hull, C., Li, G. and Von Gersdorff, H. (2006). GABA transporters regulate a standing GABAC receptor-mediated current at a retinal presynaptic terminal. J Neurosci 26: 6979–6984. Hunt, S., Pini, A. and Evan, G. (1987). Induction of c-fos-like protein in spinal cord neurons following sensory stimulation Nature 328: 632 - 634 Husebye, E., Hellström, P., Sundler, F., Chen, J. and Midtvedt, T. (2001). MidtvedtInfluence of microbial species on small intestinal myoelectric activity and transit in germ-free rats. rAm. J. Physiol. Gastrointest. Liver Physiol 280: G368-G380. Hyland, N. and Cryan, J. (2010). Gut Feeling about GABA: Focus on GABA(B) Receptors. Front. Pharmacol 1: 124. Inagaki, A., Yamaguchi, S., Takahashi-Iwanaga, Iwanaga, T. and Ishikawa, T. (2010). Functional characterization of a ClC-2-like Cl− conductance in surface epithelial cells of rat rectal colon. Journal of Membrane Biology 235: 27-41. 220

Iyer, V., Bornstein, J., Costa, M., Furness, J., Takahashi, Y. and Iwanaga, T. (1988). Electrophysiology of guinea pig myenteric neurons correlated with immunoreactivity for calcium binding proteins. J. Auton. Nerv. Syst. 22: 141-150. Jacobs, A., Barnard, K., Fishel, R. and Gradon, J. (2001). Extracolonic manifestations of Clostridium difficileinfections. Presentation of 2 cases and review of the literature. Medicine (Baltimore) 80: 88–101. Jenkinson, K., Morgan, J., Furness, J. and Southwell, B. (1999). Neurons bearing NK(3) tachykinin receptors in the guinea-pig ileum revealed by specific binding of fluorescently labelled agonists. Histochemistry and Cell Biology 112: 233-246. Jessen, K., Hills, J. and Saffrey, M. (1986). Immunohistochemical demonstration of GABAergic neurons in the enteric nervous system. J Neurosci 6: 1628–1634. Jessen, K., Mirsky, R., Dennison, M. and Burnstock, G. (1979). GABA may be a neurotransmitter in the vertebrate peripheral nervous system. Nature 281: 71-74. Jiang, M., Kirchgessner, A. and Gershon, M. (1993). Cholera toxin–sensitive neurons in guinea pig submucosal plexus. Am J Physiol 264: G86–G94. Jin, Y. and Blikslager, A. (2015). ClC-2 regulation of intestinal barrier function: Translation of basic science to therapeutic target. Tissue Barriers 3: e1105906. Jodal, M., Holmgren, S., Lundgren, O. and Sjöqvist, A. (1993). Jodal M, Holmgren S, Lundgren O, Sjöqvist A Gastroenterology 105: 1286-1293. Johnson, P. and Bornstein, J. (2004). Neurokinin-1 and -3 receptor blockade inhibits slow excitatory synaptic transmission in myenteric neurons and reveals slow inhibitory input. Neuroscience 126: 137-147. Johnson, S., Katayama, Y., Morita, K. and North, R. (1981). Mediators of slow synaptic potentials in the myenteric plexus of the guinea pig ileum. J. Phyisol. (Lond.) 320: 175-86. Johnson, S., Katayama, Y. and North, R. (1980). Slow synaptic potentials in neurones of the myenteric plexus. J Physiol 301: 505–516. Johnston, G. (2002). Medicinal chemistry and molecular pharmacology of GABA-C receptors. Current topics in medicinal chemistry 2: 903-913. Just, I., Selzer, J., Von Eichel-Streiber, C. and Aktories, K. (1995). The low molecular mass GTP-binding protein Rho is affected by toxin A from Costridium difficile. J Clin Invest 95: 1026–1031.

221

Kanwar, S., Wallace, J., Befus, D. and Kubes, P. (1994). Nitric oxide synthesis inhibition increases epithelial permeability via mast cells. Gastrointest. Liver Physiol 29: G222–G229. Katayama, Y., Lees, G. and Pearson, G. (1986). Electrophysiological and morphological characteristics of vasoactive intestinal peptide-immunoreactive neurones in the guinea-pig myenteric plexus. J. Physiol 378: 1–11. Katayama, Y. and North, R. (1978). Does substance P mediate slow synaptic excitation within the myenteric plexus? . Nature 274: 387- 388. Kato, H., Kato, N., Katow, S., Maegawa, T., Nakamura, S. and Lyerly, D. (1999). Deletions in the repeating sequences of the toxin A gene of toxin A-negative, toxin B-positive Clostridium difficile strains. FEMS Microbiol Lett 175: 197-203. Keane, F., Di Magno, E., Dozois, R. and Go, W. (1980). Relationships amongst canine interdigestive exocrine pancreatic and biliary flow, duodenal motor activity, plasma pancreatic polypeptide, and motilin. Gastroenterology 78: 726-731. Keast, J., Furness, J. and Costa, M. (1985b). nvestigations of nerve populations influencing ion transport that can be stimulated electrically, by serotonin and by a nicotinic agonist. Naunyn-Schmiedeberg's Archives of Pharmacology 331: 260-266. Keates, A., Castagliuolo, I., Qiu, B., Nikulasson, S., Sengupta, S. and Pothoulakis, C. (1998). CGRP upregulation indorsal root ganglia and ileal mucosa during Clostridiumdifﬁcile toxin A-induced enteritis. Am. J. Physiol (Gastrointest. Liver Physiol) 274: G196– G202. Kelly, C., Becker, S., Linevsky, J., Joshi, M., O’keane, J., Dickey, B., Lamont, J. and Pothoulakis, C. (1994). Neutrophil recruitment in Clostridium difﬁcile toxin A enteritis in the rabbit. J. Clin. Invest 93: 1257–1265. Kelly, C. and Kyne, L. (2011). The host immune response to Clostridium difficile. J Med Microbiol 60: 1070–1079. Kilbinger, H., Ginap, T. and Erbelding, D. (1999). GABAergic inhibition of nitric oxide- mediated relaxation of guinea-pig ileum. Naunyn-Schmiedeberg's Archives of Pharmacology 359: 500-504. Kim, M., Cooke, H., Javed, N., Carey, H., Christofi, F. and Raybould, H. (2001). D-Glucose releases 5-hydroxytryptamine from human BON cells as a model of enterochromaffin cells. Gastroenterology 121: 1400–1406.

222

Kim, P., Iaconis, J. and Rolfe, R. (1987). Immunization of adult hamsters against Clostridium difficile-associated ileocecitis and transfer of protection to infant hamsters. Infect Immun 55: 2984–2992. Kirchgessner, A. and Gershon, M. (1988a). Projections of submucosal neurons to the myenteric plexus of the guinea of intestine: in vitro tracing of microcircuits by retrograde and anterograde transport. J. Comp. Neurol 277: 487-498. Kirchgessner, A. and Gershon, M. (1988b). Projections of submucosal neurons to the myenteric plexus of the guinea pig intestine: in vitro tracing of microcircuits by retrograde and anterograde transport. J.Comp. Neurol 277: 487–498. Kirchgessner, A. and Gershon, M. (1989). Identification of vagal efferent fibers and putative target neurons in the enteric nervous system of the rat. J Comp Neurol 285: 38–53. Kirchgessner, A. and Liu, M. (1998). Immunohistochemical localization of nicotinic acetylcholine receptors in the guinea pig bowel and pancreas. The Journal of Comparative Neurology 390: 497-514. Kirchgessner, A., Liu, M. and Gershon, M. (1996). In situ identification and visualisation of neurons that mediate enteric and enteropancreatic reflexes. J. Comp. Neurol 371: 270-286. Kirchgessner, A., Tamir, H. and Gershon, M. (1992). Identification and stimulation by serotonin of intrinsic sensory neurons of the submucosal plexus of the guinea pig gut: activity-induced expression of Fos immunoreactivity. Journal of Neuroscience 12: 235-248. Kokkotou, E., Torres, D., Moss, A., O'brien, M., Grigoriadis, D., Karalis, K. and Pothoulakis, C. (2006). Corticotropin-releasing hormone receptor 2-deficient mice have reduced intestinal inflammatory responses. J. Immunol 177: 3355–3361. Konowalchuk, J., Speirs, J. and Stavric, S. (1977). Vero Response to a Cytotoxin of Escherichia coli. Innfect Immun 18: 775-9. Kordasti, S., Sapnara, M., Thomas, E., Lindstrom, E., Forsman, M., Bornstein, J. and Sjövall, H. (2006). Effects of cholera toxin on the potential difference and motor responses induced by distension in the rat proximal small intestine in vivo. Am. J. Physiol. Gastrointest. Liver Physiol 290: 948-958. Koussoulas, K., Gwynne, R. M., Foong, J. P. P. and Bornstein, J. C. (2017). Cholera Toxin Induces Sustained Hyperexcitability in Myenteric, but Not Submucosal, AH Neurons in Guinea Pig Jejunum. Frontiers in Physiology 8. 223

Krantis, A. (2000). GABA in the mammalian enteric nervous system. Physiology 15: 284- 290. Krantis, A. and Harding, R. (1987). GABA-related actions in isolated in vitro preparations of the rat small intestine. Eur J Pharmacol 141: 291–8. Krantis, A., Mattar, K. and Glasgow, I. (1998). Rat gastroduodenal motility in vivo: interaction of GABA and VIP in control of spontaneous relaxations. Gastrointest LiverPhysiol 38: G897–903. Krantis, A., Nichols, K., De Blas, A. and Staines, W. (1994). Demonstration of benzodiazepine receptors in submucosal neurons of the gastrointestinal tract. Neurosci Lett 176: 32–36. Krantis, A., Shabnavard, L., Nichols, K., De Blas, L. and Staines, W. (1995). Localization of GABAA receptor immunoreactivity in NO synthase positive myenteric neurones. Journal of the autonomic nervous system 53: 157-165. Kuijper, E., De Weerdt, J., Kato, H., Kato, N., Van Dam, A., Van Der Vorm, E., Weel, J., Van Rheenen, C. and Dankert, J. (2001). Nosocomial outbreak of Clostridium difficile-associated diarrhoea due to a clindamycin-resistant enterotoxin A-negative strain. Eur J Clin Microbiol Infect Dis 20: 528–534. Kunze, W., Bornstein, J. and Furness, J. (1995). Identification of sensory nerve cells in a peripheral organ, the intestine of a mammal. Neuroscience 66: 1-4. Kunze, W., Bornstein, J., Furness, J., Hendriks, R. and Stephenson, D. (1994). Charybdotoxin and iberiotoxin but not apamin abolish the slow after hyperpolarisation in myenteric plexus neurons. Pflugers Arch. Eur. J. Physiol 428: 300-306. Kunze, W., Clerc, N., Bertrand, P. and Furness, J. (1999). Contractile activity in intestinal muscle evokes action potential discharge in guinea-pig myenteric neurons. J Physiol (Lond). 517: 547-561. Kunze, W., Clerc, N., Furness, J. and Gola, M. (2000). The soma and neurites of primary afferent neurons in the guinea pig intestine respond differentially to deformation. J. Physiol. (Lond.) 526: 375-385. Kunze, W. and Furness, J. (1999). The enteric nervous system and regulation of intestinal motility. Annu. Rev. Physiol 61: 117-142. Kunze, W., Furness, J., Bertrand, P. and Bornstein, J. (1998). Intracellular recording from myenteric neurons of the guinea pig ileum that respond to stretch. J. Physiol. (Lond.) 506: 827-842. 224

Kunze, W., Furness, J. and Bornstein, J. (1993). Simultaneous intracellular recordings from enteric neurons reveal that myenteric AH neurons transmit via slow excitatory post synaptic potentials. Neuroscience: 685-694. Kurose, I., Pothoulakis, C., Lamont, J., Anderson, D., Paulson, J., Miyasaka, M. and Granger, D. (1994). Clostridium difficile toxin A-induced microvascular dysfunction. Role of histamine. Journal of Clinical Investigation 94: 1919. Kyne, L., Hamel, M., Polavaram, R. and Kelly, C. (2002). Health Care Costs and Mortality Associated with Nosocomial Diarrhea Due to Clostridium difficile. Clinical Infectious Diseases 34: 346-53. Lakhan, S. and Kirchgessner, A. (2010). Neuroinflammation in inflammatory bowel disease. Journal of Neuroinflammation 7: 37. Langford, D., Bailey, A., Chanda, M., Clarke, S., Drummond, T., Echols, S., Glick, S., Ingrao, J., Klassen-Ross, T., Lacroix-Fralish, M., Matsumiya, L., Sorge, R., Sotocinal, S., Tabaka, J., Wong, D., Van Den Maagdenberg, A., Ferrari, M., Craig, K. and Mogil, J. (2010). Coding of facial expressions of pain in the laboratory mouse. Nature Methods 7: 447-449. Larsson, M., Sapnara, M., Thomas, E., Bornstein, J., Lindstrom, E., Svensson, D. and Sjovall, H. (2008). Pharmacological analysis of components of the change in transmural potential difference evoked by distension of rat proximal small intestine in vivo. American Journal of Physiology-Gastrointestinal and Liver Physiology 294: G165- G173. Lembeck, F. and Holzer, P. (1979). Substance P as neurogenic mediator of antidromic vasodilation and neurogenic plasma extravasation. Naunyn-Schmiedeberg's archives of pharmacology 31: 175-183. Lepard, K., Messori, E. and Galligan, J. (1997). Purinergic fast excitatory postsynaptic potentials in myenteric neurons of guinea pig: Distribution and pharmacology. Gastroenterology 113: 1522-1534. Lessa , F. C., Mu , Y., Bamberg , W. M., Beldavs , Z. G., Dumyati , G. K., Dunn , J. R., Farley , M. M., Holzbauer , S. M., Meek , J. I., Phipps , E. C., Wilson , L. E., Winston , L. G., Cohen , J. A., Limbago , B. M., Fridkin , S. K., Gerding , D. N. and Mcdonald , L. C. (2015). Burden of Clostridium difficile Infection in the United States. New England Journal of Medicine 372: 825-834.

225

Li, Z., Chalazonitis, A., Huang, Y., Mann, J., Margolis, K., Yang, Q., Kim, D., Côté, F., Mallet, J. and Gershon, M. (2011). Essential roles of enteric neuronal serotonin in gastrointestinal motility and the development/survival of enteric dopaminergic neurons. Journal of Neuroscience, 31: 8998-9009. Liddle, R. (1997). Cholecystokinin cells. Annu. Rev. Physiol 59: 221-242. Limaye, A., Turgeon, D., Cookson, B. and Fritsche, T. (2000). Pseudomembranous colitis caused by a toxin A− B+ strain of Clostridium difficile. Journal of clinical microbiology 38: 1696-1697. Linden, D., Sharkey, K. and Mawe, G. (2003). Enhanced excitability of myenteric AH neurones in the inflamed guinea pig distal colon. J. Physiol. (Lond.) 547: 589-601. Liu, M., Rothstein, J., Gershon, M. and Kirchgessner, A. (1997). Glutamatergic enteric neurons. Journal of Neuroscience 17: 4764-4784. Liu, S., Hu, H., Gao, C., Gao, N., Wang, G., Wang, X., Gao, X., Xia, Y. and Wood, J. (2003). Actions of cysteinyl leukotrienes in the enteric nervous system of guinea-pig stomach and small intestine. European journal of pharmacology 459: 27-39. Liu, S., Ji, T., Ren, W., Qu, M., Zhu, J. and Wood, J. (2013). Role of Na-K-2Cl symporter in GABA-evoked excitation in rat enteric neurons. FASEB J, (meeting abstract supplement) 27: 1160. Lomax, A., Bertrand, P. and Furness, J. (1998). Identification of the populations of enteric neurons that have NK1 tachykinin receptors in the guinea-pig small intestine. Cell and Tissue Research 294: 27-33. Lomax, A., Mawe, G. and Sharkey, K. (2005). Synaptic facilitation and enhanced neuronal excitability in the submucosal plexus during experimental colitis in guinea-pig. J Physiol 564: 863–875. Lomax, A., O'hara, J., Hyland, N., Mawe, G. and Sharkey, K. (2006). Persistent alterations to enteric neural signalling in guinea pig colon following resolution of colitis. Am J Physiol Gastrointest Liver Physiol 292: 482–491. Lomax, A., Sharkey, K., Bertrand, P., Low, A., Jc, B. and Furness, J. (1999). Correlation of morphology, electrophysiology and chemistry of neurons in the myenteric plexus of the guinea-pig distal colon. J. Auton. Nerv. Syst 76: 45–61. Louis, S., Jamieson, A., Russel, N. and Dockray, G. (1989). The Role of Substance P and Calcitonin Gene-Related Peptide in Neurogenic Plasma Extravasation and Vasodilatation in the Rat. Neuroscience 581-586. 226

Lukasiewicz, P. and Shields, C. (1998). Different combinations of GABAA and GABAC receptors confer distinct temporal properties to retinal synaptic responses. 79: 3157– 3167. Lundgren, O. (1998). 5-Hydroxytryptamine, enterotoxins, and intestinal secretion. Gastroenterology 115: 1009-1012. Lundgren, O. (2002). Enteric nerves and diarrhoea. Pharmacol. Toxicol 90: 109-120. Lundgren, O. and Jodal, M. (1997). The enteric nervous system and cholera toxin-induced secretion. Comparative Biochemistry and Physiology Part A: Physiology 118: 319- 327. Luzzi, S., Franchi-Micheli, S., Ciuffi, M., Rosi, E. and Zilletti, L. (1987). Effect of various GABA-receptor agonists and antagonists on anaphylactic histamine release in the guinea-pig ileum. Agents Actions 20: 181–184. Lyerly, D., Lockwood, D., Richardson, S. and Wilkins, T. (1982). Biological activities of toxins A and B of Clostridium difficile. Infect. Immun 35: 1147-1150.

Lyerly, D., Saum, K., Mcdonald, D. and Wilkins, T. (1985). Effects of C. difficile toxins given intragastrically to animals. Infect. Immun 47: 349- 352. Lyras, D., O’connor, J., Howarth, P., Sambol, S., Carter, G., Phumoonna, T. and Rood, J. (2009). Toxin B is essential for virulence of Clostridium difficile. Nature 458: 1176– 1179. Macnaughton, W., Moore, B. and Vanner, S. (1997). Cellular pathways mediating tachykinin-evoked secretomotor responses in guinea pig ileum. American Journal of Physiology-Gastrointestinal and Liver Physiology 273: G1127-G1134. Macnaughton, W., Pineau, B. and Krantis, A. (1996). Gamma-Aminobutyric acid stimulates electrolyte transport in the guinea pig ileum in vitro. Gastroenterology 110: 498–507. Mahida, Y., Makh, S., Hyde, S., Gray, T. and Borriello, S. (1996). Effect of Clostridium difficile toxin A on human intestinal epithelial cells: induction of interleukin 8 production and apoptosis after cell detachment. Gut 38: 337-347. Mahony, D., Gilliatt, E., Dawson, S., Stockdale, E. and Lee, S. (1989). Vero Cell Assay for Rapid Detection of Clostridium perfringens Enterotoxin. Applied and Environmental Microbiology. 55: 2141-2143.

227

Mann, P., Furness, J. and Southwell, B. (1998). Choline acetyltransferase immunoreactivity of putative intrinsic primary afferent neurons in the rat ileum. Cell Tissue Res 297: 241–248. Mann, P., Southwell, B., Young, H. and Furness, J. (1997). Appositions made by axons of descending interneurons in the guinea-pig small intestine, investigated by confocal microscopy. Journal of Chemical Neuroanatomy 12: 151-164. Mantyh, C., Pappas, T., Lapp, J., Washington, M., Neville, L., Ghilardi, J., Rogers, S., Mantyh, P. and Vigna, S. (1996). Substance P activation of enteric neurons in response to intraluminal Clostridium difficile toxin A in the rat ileum. Gastroenterology 111: 1272–1280. Mao, Y., Wang, B. and Kunze, W. (2006). Characterization of myenteric sensory neurons in the mouse small intestine. Journal of neurophysiology 96: 998-1010. Marcoli, M., Scarrone, S., Maura, G., Bonanno, G. and Raiteri, M. (2000). A subtype of the gamma-aminobutyric acid (B) receptor regulates choliner¬gic twitch response in the guinea pig ileum. J. Pharmacol. Exp. Ther 293: 42–47. Mathias, J., Carlson, G., Dimarino, A., Bertiger, G., Morton, H. and Cohen, S. (1976). Intestinal myoelectric activity in response to live Vibrio cholerae and cholera enterotoxin. J. Clin. Invest 58: 91–9610. Matsumiya, L., Sorge, R., Sotocinal, S. and Tabaka, J. (2012). Using the Mouse Grimace Scale to reevaluate the efficacy of postoperative analgesics in laboratory mice. Journal of the American Association for Laboratory Animal Science 51: 42-49. Matthews, G., Ayoub, G. and Heidelberger, R. (1994). Presynaptic inhibition by GABA is mediated via two distinct GABA receptors with novel pharmacology. J Neurosci 14: 1079–1090. Mawe, G., Brabchek, T. and Gershon, M. (1986). Peripheral neural serotonin receptors: Identification and characterisation with specific antagonists and agonists. Proc. Natl. Acad. Sci USA 83: 9799-9803. Mawe, G., Strong, D. and Sharkey, K. (2009). Plasticity of enteric nerve functions in the inflamed and post-inflamed gut. NeurogastroMot 21: 481-491. Mazzoli, R. and Pessione, E. (2016). The Neuro-endocrinological Role of Microbial Glutamate and GABA Signaling. Frontiers in Microbiology 7: 1934.

228

Mazzuoli, G. and Schemann, M. (2009). Multifunctional rapidly adapting mechanosensitive enteric neurons (RAMEN) in the myenteric plexus of the guinea pig ileum. The Journal of Physiology 587: 4681-4694. Mazzuoli, G. and Schemann, M. (2012). Mechanosensitive enteric neurons in the myenteric plexus of the mouse intestine. PLoS ONE 7: e39887. Mclarren, K., Cole, A., Weisser, S., Voglmaier, N., Conlin, V., Jacobson, K. and Sly, L. (2011). SHIP-deficient mice develop spontaneous intestinal inflammation and arginase-dependent fibrosis. The American journal of pathology 179: 180-188. Mcvey Neufeld, K., Mao, Y., Bienenstock, J., Foster, J. and Kunze, W. (2013). The microbiome is essential for normal gut intrinsic primary afferent neuron excitability in the mouse. Neurogastroenterol.Motil 25: e183–e188. Mellander, A., Järbur, K. and Sjövall, H. (2000). Pressure and frequency dependent linkage between motility and epithelial secretion in human proximal small intestine. Gut 46: 376-384. Messenger, J. and Furness, J. (1990). Projections of chemically specified neurons in the guinea-pig colon. Arch. Histol. Cytol 53: 467-495. Methé, B., Nelson, K., Pop, M., Creasy, H., Giglio, M., Huttenhower, C. and White, O. (2012). A framework for human microbiome research. Nature 486: 215–221. Meyer, G., Neetz, A., Brandes, G., Tsikas, D., Butterfield, J., Just, I. and Gerhard, R. (2007). Clostridium difficile toxins A and B directly stimulate human mast cells. Infect. Immun 75: 3868–3876. Michel, K., Michaelis, M., Mazzuoli, G., Mueller, K., Vanden Berghe, P. and Schemann, M. (2011). Fast calcium and voltage-sensitive dye imaging in enteric neurones reveal calcium peaks associated with single action potential discharge. J Physiol 589: 5941– 5947. Mihara, S., Nishi, S., North, R. and Surprenant, A. (1987). A non-adrenergic, non-cholinergic slow inhibitory post-synaptic potential in neurones of the guinea-pig submucous plexus. Journal of Physiology 390: 357-365. Mitchell, T., Ketley, J., Haslam, S., Stephen, D., Burdon, D., Candy, A. and Daniel, R. (1986). Effect of toxins A and B on Clostridium difficile on rabbit ileum and colon. Gut 27: 78-85.

229

Mongardi Fantaguzzi, C., Thacker, M., Chiocchetti, R. and Furness, J. (2009). Identification of neuron types in the submucosal ganglia of the mouse ileum. Cell and Tissue Research 336: 179-189. Monro, R., Bertrand, P. and Bornstein, J. (2004). ATP participates in three excitatory postsynaptic potentials in the submucous plexus of the guinea pig ileum. Journal of Physiology-London 556: 571-584. Monro, R., Bornstein, J. and Bertrand, P. (2005). Slow excitatory post-synaptic potentials in myenteric AH neurons of the guinea-pig ileum are reduced by the 5- hydroxytryptamine7 receptor antagonist SB 269970. Neuroscience 134: 975-986. Monro, R., Bornstein, J. and Bertrand, P. (2008). Synaptic transmission from the submucosal plexus to the myenteric plexus in Guinea-pig ileum. Neurogastroenterology and Motility 20: 1165-1173. Moore, B. and Vanner, S. (2000). Properties of synaptic inputs from myenteric neurons innervating submucosal S neurons in guinea pig ileum. American Journal of Physiology Gastrointestinal and Liver Physiology 278: G273-280. Morgan, J. and Curran, T. (1989). Stimulus-transcription coupling in neurons: role of cellular immediate-early genes. Trends Neurosci 12: 459–462. Morita, K., North, R. and Tokimasa, T. (1982a). The calcium-activated potassium conductance in guinea pig myenteric neurones. J.Physiol. (Lond.) 329: 341-354. Mourad, F. and Nassar, C. (2000). Effect of vasoactive intestinal polypeptide (VIP) antagonism on rat jejunal fluid and electrolyte secretion induced by cholera and Escherichia coli enterotoxins. Gut 47: 382-386. Mourad, F., O’donnell, L. and Dias, J. (1995). Role of 5-hydroxytryptamine type 3 receptors in rat intestinal fluid and electrolyte secretion induced by cholera and Escherichia coli enterotoxins. Gut 37: 350-345. Nacher, A., Polache, A., Moll-Navarro, M., Pla-Delfina, J. and Merino, M. (1994). Intestinal absorption pathway of γ-aminobutyric acid in rat small intestine. Biopharm. Drug Disposition 15: 359–371. Nakajima, K., Tooyama, I., Kuriyama, K. and Kimura, H. (1996). Immunohistochemical demonstration of GABAB receptors in the rat gastrointestinal tract. Neurochem Res 21: 211–215. Neal, K. and Bornstein, J. (2007a). Mapping serotonin inputs to enteric neurons of the guinea pig small intestine. Neuroscience 145: 556-567. 230

Neal, K. and Bornstein, J. (2008). Targets of myenteric interneurons in the guinea pig small intestine Neurogastroenterol Motil 20: 566-575. Nemeth, P., Ort, C. and Wood, J. (1984). Intracellular study of effects of histamine on electrical behavior of myenteric neurones in guinea-pig small intestine. J. Physiol. (Lond.) 355: 411–425. Neunlist, M. and Schemann, M. (1997). Projections and neurochemical coding of myenteric neurons innervating the mucosa of the guinea pig proximal colon. Cell Tissue Res 287: 119-125. Neylon, C., Nurgali, K., Hunne, B., Robbins, H., Moore, S., Chen, M. and Furness, J. (2004). Intermediate-conductance calcium-activated potassium channels in enteric neurones of the mouse: pharmacological, molecular and immunochemical evidence for their role in mediating the slow after hyperpolarisation. J. Neurochem 90. Nguyen, T., Poole, D., Harvey, J., Stebbing, M. and Furness, J. (2005). Investigation of PKC isoform-specific translocation and targeting of the current of the late after hyperpolarising potential of myenteric AH neurons. Eur. J. Neurosci 21: 905-913. Nichols, K., Staines, W., Wu, J.-Y. and Krantis, A. (1995). Immunopositive GABAergic neural sites display nitric oxide synthase-related NADPH diaphorase activity in the human colon. J Auton Nerv Syst 50: 253–262. Nilsson, O., Cassuto, J., Larsson, P. and Jodal, M. (1983). 5-Hydroxytryptamine and cholera secretion: A histochemical and physiological study in cats. Gut 24: 542–548. Nishi, S. and North, R. (1973a). Intracellular recording from the myenteric plexus of the guinea pig ileum. J. Physiol 231: 471-491. North, R. (1973). The calcium-dependent slow after hyperpolarisation in myenteric plexus neurone with tetrodotoxin-resistant action potentials. Br. J. Pharmacol 49: 709-711. North, R. and Nishi, S. (1974). Properties of the ganglion cells off the myenteric plexus of the guinea pig ileum determined by intracellular recording. Proc. 4th Int. Symp. on Gastrointestinal Motility: 667-676. North, R., Slack, B. and Surprenant, A. (1985). Muscarinic M1 and M2 receptors mediate depolarization and presynaptic inhibition in guinea‐pig enteric nervous system. The Journal of Physiology 368: 435-452. North, R. and Surprenant, A. (1985). Inhibitory synaptic potentials resulting from α2 adrenoceptor activation in guinea-pig submucous plexus neurones. J. Physiol. (Lond.) 358: 17-33. 231

North, R. and Tokimasa, T. (1987). Persistent calcium-sensitive potassium current and the resting properties of guinea pig myenteric neurones. J. Physiol. (Lond.) 386: 333- 353. Nurgali, K., Furness, J. and Stebbing, M. (2003b). Correlation of electrophysiology, shape and synaptic properties of myenteric AH neurons of the guinea-pig distal colon. Auton Neurosci 103: 50–64. Nurgali, K., Nguyen, T., Matsuyama, H., Thacker, M., Robbins, H. and Furness, J. (2007). Phenotypic changes of morphologically identified guinea-pig myenteric neurons following intestinal inflammation. The Journal of Physiology 583: 593–609. Nurgali, K., Stebbing, M. and Furness, J. (2004). Correlation of electrophysiological and morphological characteristics of enteric neurons in the mouse colon. J Comp Neurol 468: 112–124. Ong, J. and Kerr, D. (1987). Comparison of GABA-induced responses in various segments of the guinea-pig intestine. Eur. J. Pharmacol 134: 349–353. Palmer, J., Wong-Riley, M. and Sharkey, K. (1998). Functional alterations in jejunal myenteric neurons during inflammation in nematode-infected guinea pigs. Am J Physiol 275: G922–935. Palmer, M. (2006). Functional segregation of synaptic GABAA and GABAC receptors in goldfish bipolar cell terminals. The Journal of Physiology 577: 45–53. Pan, H. and Gershon, M. (2000). Activation of intrinsic afferent pathways in submucosal ganglia of the guinea pig small intestine Journal of Neuroscience 20: 3295-3309. Pepin, J., Vo, T., Boutros, M., Marcotte, E., Dial, S., Dubé, S., Vasilevsky, C., Mcfadden, N., Patino, C. and Labbé, A. (2009). Risk factors for mortality following emergency colectomy for fulminant Clostridium difficile infection. Dis Colon Rectum 52: 400– 405. Pokusaeva, K., Johnson, C., Luk, B., Uribe, G., Fu, Y., Oezguen, N., Matsunami, R., Lugo, M., Major, A., Mori‐Akiyama, Y. and Hollister, E. (2017 ). GABA‐producing Bifidobacterium dentium modulates visceral sensitivity in the intestine. Neurogastroenterology & Motility 29. Pompolo, S. and Furness, J. (1995). Sources of inputs to longitudinal muscle motor neurons and ascending interneurons in the guinea-pig small intestine. Cell and Tissue Research 280: 549-560.

232

Pompolo, S. and Furness, J. (1998). Quantitative analysis of inputs to somatostatin immunoreactive descending interneurons in the myenteric plexus of the guinea pig small intestine. Cell Tissue Res 294: 216-226. Poole, D., Amadesi, S., Rozengurt, E., Thacker, M., Bunnet, N. and Furness, J. (2008). Stimulation of the Neurokinin 3 Receptor Activates Protein Kinase C Epsilon and Protein Kinase D in Enteric Neurons. Gastrointestinal and Liver Physiology 294: G1245-G1256 Portbury, A., Furness, J., Young, H., Southwell, B. and Vigna, S. (1996). Localisation of NK1 receptor immunoreactivity to neurons and interstitial cells of the guinea-pig gastrointestinal tract. The Journal of Comparative Neurology 367: 342-351. Portbury, A., Pompolo, S., Furness, J., Stebbing, M., Kunze, W., Bornstein, J. and Hughes, S. (1995). Cholinergic, somatostatin-immunoreactive interneurons in the guinea pig intestine: morphology, ultrastructure, connections and projections. Journal of anatomy 187: 303. Pothoulakis, C., Castagliuolo, I. and Lamont, J. (1998). Nerves and intestinal mast cells modulate responses to enterotoxins. NiPS 13: 58-63. Pothoulakis, C., Castagliuolo, I., Lamont, J., Jaffer, A., O'keane, J., Snider, R. and Leeman, S. (1994). CP-96,345, a substance P antagonist, inhibits rat intestinal responses to Clostridium difficile toxin A but not cholera toxin. Proc Natl Acad Sci, USA 91: 947- 951. Pothoulakis, C., Karmeli, F., Kelly, C., Eliakim, R., Joshi, M., O’keane, J., Castagliuolo, I., Lamont, J. and Rachmilewitz, D. (1993). Ketotifen inhibits Clostridium difficile toxin A-induced enteritis in rat ileum. Gastroenterology 105: 701–707. Pothoulakis, C. and Lamont , T. (2001). Microbes and Microbial Toxins: Paradigms for Microbial- Mucosal Interactions II. The integrated response of the intestine to Clostridium difficile toxins. American Journal of Physiology - Gastrointestinal and Liver Physiology 280: G178-G183 Pothoulakis, C., Sullivan, R., Melnick, D., Gadenne, A., Meshulan, T. and Lamont, J. (1988). Clostridium difficile toxin A stimulates intracellular calcium release and chemotactic response in human neutrophils. J Clin Invest 81: 1741-1745. Poulter, M., Singhal, R., Brown, L. and Krantis, A. (1999). GABA(A) receptor subunit messenger RNA expression in the enteric nervous system of the rat: implications for functional diversity of enteric GABA(A) receptors. Neuroscience: 1159–1165. 233

Qian, H. and Dowling, J. (1995). GABAA and GABAC receptors on hybrid bass retinal bipolar cells. J. Neurophysiol 74: 1920–1928. Qiu, B., Pothoulakis, C., Castagliuolo, I., Nikulasson, S. and Lamont, J. (1996). Nitric oxide inhibits rat intestinal secretion by Clostridium difficile toxin A, but not Vibrio cholerae enterotoxin. Gastroenterology 111: 409–418 Qu, Z., Thacker, M., Castelucci, P., Bagyánszki, M., Epstein, M. and Furness, J. (2008). Immunohistochemical analysis of neuron types in the mouse small intestine. Cell and Tissue Research 334: 147-161. Quartara, L. and Maggi, C. (1997). The tachykinin NK1 receptor. Part I: ligands and mechanisms of cellular activation. Neuropeptides 31: 537-63. Ragozzino, D., Woodward, R., Murata, Y., Eusebi, F., Overman, L. and Miledi, R. (1996). Design and in vitro pharmacology of a selective gamma-aminobutyric acid receptor antagonist. Molecular Pharmacology October 50: 1024-1030. Read, N. (1980). The migrating motor complex and spontaneous fluctuations of transmural potential difference in the human small intestine. In: Gastrointestinal motility: 299- 308. Reddix, R. and Cooke, H. (1992). Neurokinin 1 receptors mediate substance P-induced changes in ion transport in guinea-pig ileum. Regulatory Peptides 39: 215-225. Reed, D. and Vanner, S. (2001). Converging and diverging cholinergic inputs from submucosal neurons amplify activity of secretomotor neurons in guinea pig ileal submucosa. Neuroscience 107: 685-696. Ren, J., Bian, X., Devries, M., Schnegelsberg, B., Cockayne, D., Ford, A. and Galligan, J. (2003). P2X2 subunits contribute to fast synaptic excitation in myenteric neurons of the mouse small intestine. The Journal of physiology 552: 809-821. Ren, J., Hu, H., Liu, S., Xia, Y. and Wood, J. (2000). Glutamate receptors in the enteric nervous system: ionotropic or metabotropic? Neurogastroenterology & Motility 12: 257-264. Riegler, M., Sedivy, R., Pothoulakis, C., Hamilton, G., Zacherl, J., Bischof, E., Cosentini, W., Feil, R. S. and Lamont, J. (1995). Clostridium difficile toxin B is more potent than toxin A in damaging human colonic epithelium in vitro. Journal of Clinical Investigation 95: 2004–2011. Rolfe, R. (1991). Binding kinetics of Clostridium difficile toxins A and B to intestinal brush border membranes from infant and adult hamsters. Infect. Immun 59: 1223-1230. 234

Rolfe, R. and Song, W. (1993). Purification of a functional receptor for Clostridium difficile toxin A from intestinal brush border membranes of infant hamsters. Clin. Infect. Dis 16: 219-227. Rugiero, F., Gola, M., Kunze, W., Reynaud, J.-C., Furness, J. and Clerc, N. (2002a). Analysis of whole cell currents by patch clamp of guinea pig myenteric neurones in intact ganglia. J. Physiol. (Lond.) 538: 447-463. Rugiero, F., Gola, M., Kunze, W., Reynaud, J., Furness, J. and Clerc, N. (2002b). Analysis of whole cell currents by patch clamp of guinea pig myenteric neurones in intact ganglia. J. Physiol. (Lond.) 538: 447-463. Russell, J. (2000). Sodium–potassium–chloride cotransport. Physiological Reviews 80: 211– 276. Sage, D., Salin, P., Alcaraz, G., Castets, F., Giraud, P., Crest, M., Mazet, B. and Clerc, N. (2007). Nav1.7 and Nav1.3 are the only tetrodotoxin-sensitive sodium channels expressed by the adult guinea pig enteric nervous system. J Comp Neurol 504: 363– 378. Sang, Q., Williamson, S. and Young, H. (1997). Projections of chemically identified myenteric neurons of the small and large intestine of the mouse. J Anat 190: 209– 222. Sang, Q. and Young, H. (1996). Chemical coding of neurons in the myenteric plexus and external muscle of the small and large intestine of the mouse. Cell Tissue Res 284: 39–53. Sang, Q. and Young, H. (1998). The identification and chemical coding of cholinergic neurons in the small and large intestine of the mouse. Anat Rec 251: 185–199. Sanger, G., Munonyara, M., Dass, N., Prosser, H., Pangalos, M. and Parsons, M. (2002). GABAB recep¬tor function in the ileum and urinary bladder of wildtype and GABAB1 subunit null mice. Autonomic and Autacoid Pharmacology 22: 147-154. Sansonetti, P., Van Nhieu, G. and Égile, C. (1999). Rupture of the intestinal epithelial barrier and mucosal invasion by Shigella flexneri. Clinical infectious diseases 28: 466-475. Savidge, T. (2016). Epigenetic regulation of enteric neurotransmission by gut bacteria. Frontiers in cellular neuroscience 9: 503. Savidge, T., Pan, W., Newman, P., O'brien, M., Anton, P. and Pothoulakis, C. (2003). Clostridium difficile toxin B is an inflammatory enterotoxin in human intestine. Gastroenterology 125: 413-420. 235

Savidge, T., Urvil, P., Oezguen, N., Ali, K., Choudhury, A., Acharya, V., Pinchuk, I., Torres, A., English, R., Wiktorowicz, J., Loeffelholz, M., Kumar, R., Shi, L., Nie, W., Braun, W., Herman, B., Hausladen, A., Feng, H., Stamler, J. and Pothoulakis, C. (2011). Host S-nitrosylation inhibits clostridial small molecule-activated glucosylating toxins. Nature medicine 17: 1136-1141. Schemann, M. and Mazzuoli, G. (2010). Multifunctional mechanosensitive neurons in the enteric nervous system. Autonomic Neuroscience: Basic and Clinical 153: 21-25. Schemann, M. and Neunlist, M. (2004). The human enteric nervous system. Neurogastroenterol. Motil 16: 5-55. Schemann, M. and Wood, J. (1989a). Electrical behaviour of myenteric neurones in the gastric corpus of the guinea pig. J. Physiol 417: 501-518. Schwörer, H., Rackè, K. and Kilbinger, H. (1989). GABA receptors are involved in the modulation of the release of 5-hydroxytryptamine from the vascularly perfused small intestine of the guinea-pig. Eur J Pharmacol 165: 29–37. Seifi, M., Brown, J., Mills, J., Bhandari, P., Belelli, D., Lambert, J., Rudolph, U. and Swinny, J. (2014). Molecular and functional diversity of GABA-A receptors in the enteric nervous system of the mouse colon. Journal of Neuroscience 34: 10361-10378. Sharkey, K., Williams, R. and Dockray, G. (1984). Sensory substance P innervation of the stomach and pancreas. Demonstration of capsaicin- sensitive sensory neurons in the rat by combined immunohistochemistry and retrograde tracing. Gastroenterology 87: 914-921. Sharp, G. and Hynie, S. (1971). Stimulation of intestinal adenyl cyclase by cholera toxin. Nature 229: 266–269. Shen, K. Z. and Surprenant, A. (1993). Somatostatin-mediated inhibitory postsynaptic potential in sympathetically denervated guineapig submucosal neurones. J. Physiol. (Lond.) 470: 619-635. Sheng, H., Fields, R. and Nelson, P. (1993). Specific regulation of immediate early genes by patterned neuronal activity. J Neurosci Res 35: 459–467. Sheng, M. and Greenberg, M. (1990). The regulation and function of c-fos and other immediate early genes in the nervous system. Neuron 4: 477-485. Shreiner, A., Kao, J. and Young, V. (2015). The gut microbiome in health and in disease. Current Opinion in Gastroenterology 31: 69–75.

236

Sigel, E. and Steinmann, M. (2012). Structure, Function, and Modulation of GABAA Receptors. The Journal of Biological Chemistry 287: 40224–40231. Sjöqvist, A. (1991). Interaction between antisecretory opioid and sympathetic mechanisms in the rat small intestine. . Acta Physiologica Scandinavica 142: 127-132. Song, Z., Brookes, S. and Costa, M. (1991). Identification of myenteric neurons which project to the mucosa of the guinea pig small intestine. Neurosci Lett 129: 294-298. Song, Z., Brookes, S. and Costa, M. (1994). All calbindin-immunoreactive myenteric neurons project to the mucosa of the guinea pig small intestine. Neurosci. Lett 180: 219-222. Song, Z., Brookes, S. and Costa, M. (1996). Projections of specific morphological types of neurons within the myenteric plexus of the small intestine of the guinea-pig. Cell Tissue Res 285: 149–156. Song, Z., Brookes, S., Ramsay, G. and Costa, M. (1997). Characterization of myenteric interneurons with somatostatin immunoreactivity in the guinea-pig small intestine. Neuroscience 80: 907–923. Song, Z., Brookes, S., Steele, P. and Costa, M. (1992). Projections and pathways of submucous neurons to the mucosa of the guinea-pig small intestine. Cell Tissue Res 269: 87-98. Song, Z., Costa, M. and Brookes, S. (1998). Projections of submucous neurons to the myenteric plexus in the guinea pig small intestine. J Comp Neurol 399: 255–268. Sörensson, J., Jodal, M. and Lundgren, O. (2001). Involvement of nerves and calcium channels in the intestinal response to Clostridium difficile toxin A: an experimental study in rats in vivo. Gut 49: 56-65. Spencer, N., Kyloh, M. and Duffield, M. (2014). Identification of Different Types of Spinal Afferent Nerve Endings That Encode Noxious and Innocuous Stimuli in the Large Intestine Using a Novel Anterograde Tracing Technique. PLoS ONE 9: e112466. Spencer, T. and Smith, T. (2004). Mechanosensory S-neurons rather than AH-neurons appaer to generate a rythmic motor pattern in guinea-pig distal colon. J Physiol 558: 577- 596. Starodub, A. and Wood, J. (2000a). Histamine H₂ receptor activated chloride conductance in myenteric neurons from guinea pig small intestine. J. Neurophysiol 83: 1809-1816. Stebbing, M. and Bornstein, J. (1996). Electrophysiological mapping of fast excitatory synaptic inputs to morphologically and chemically characterised myenteric neurons of guinea pig small intestine. Neuroscience 73: 1017-1028. 237

Steele, J., Chen, K., Sun, X., Zhang, Y., Wang, H., Tzipori, S. and Feng, H. (2012). Systemic Dissemination of Clostridium difficile Toxins A and B Is Associated With Severe, Fatal Disease in Animal Models. The Journal of Infectious Diseases 205: 384–391. Sun, X., Savidge, T. and Feng, H. (2010). The enterotoxicity of Clostridium difficile toxins. Toxins 2: 1848-1880. Surprenant, A. (1984a). Slow excitatory synaptic potentials recorded from neurones of guinea-pig submucous plexus. Journal of Physiology 351: 343-361. Sutherland, K., Young, R., Cooper, N., Horowitz, M. and Blackshaw, L. (2007). Phenotypic characterization of taste cells of the mouse small intestine. Am. J. Physiol. Gastrointest. Liver Physiol 292: G1420–G1428. Sutton, P., Li, S., Webb, J., Solomon, K., Brazier, J. and Mahida, Y. (2008). Essential role of toxin A in C. difficile 027 and reference strain supernatant-mediated disruption of Caco-2 intestinal epithelial barrier function. Clin. Exp. Immunol 153: 439–447. Szurszewski, J., Ermilov, L. and Miller, S. (2002). Prevertebral ganglia and intestinofugal afferent neurones. Gut 51: i6-i10. Tack, J. and Wood, J. (1992). Electrical behaviour of myenteric neurones in the gastric antrum of the guinea‐pig. The Journal of Physiology 447: 49-66. Takaki, M., Mawe, G., Barasch, J. and Gershon, M. (1985). Physiological response of guinea pig myenteric neurons secondary to the release of endogenous serotonin by tryptamine. Neuroscience 16: 223-240. Takaki, M. and Nakayama, S. (1990). Electrical behavior of myenteric neurons induced by mesenteric nerve stimulation in the guinea pig ileum. Acta Med Okayama 44: 257- 261. Tamura, K. (1992). Morphology of electrophysiologically identified myenteric neurons in the guinea pig rectum. Am. J Physiol 262: G545-G552. Tamura, K. and Wood, J. (1992). Effects of prolonged exposure to histamine on guinea pig intestinal neurons. Digest Dis Sci 37: 1084. Tan, L., Bornstein, J. and Anderson, C. (2008). Distinct chemical classes of medium-sized transient receptor potential channel vanilloid 1-immunoreactive dorsal root ganglion neurons innervate the adult mouse jejunum and colon. Neuroscience 56: 334-343. Tan, L., Bornstein, J. and Anderson, C. (2010). The neurochemistry and innervation patterns of extrinsic sensory and sympathetic nerves in the myenteric plexus of the C57Bl6 mouse jejunum. Neuroscience 166: 564-79. 238

Tedeschi, L., Carai, M., Frison, G., Favretto, D., Colombo, G., Ferrara, S. and Gessa, G. (2003). Endogenous γ-hydroxybutyric acid is in the rat, mouse and human gastrointestinal tract. Life sciences 72: 2481-2488. Thelestam, M. and Florin, I. (1984). Cytopathogenic action of Clostridium difficile toxins. J. Toxicol. Toxin Rev 3: 139-180. Thomas, E., Bertrand, P. and Bornstein, J. (1999). Genesis and role of coordinated ﬁring in a feedforward network: a model study of the enteric nervous system. Neuroscience 93: 1525–1537. Thomas, E., Bertrand, P. and Bornstein, J. (2000). A computer simulation of recurrent excitatory networks of sensory neurons the gut in guinea pig. Neurosci. Lett 287: 137-140. Thomas, E. and Bornstein, J. (2003). Inhibitory cotransmission or after hyperpolarising potentials can regulate firing in recurrent networks with excitatory metabotropic transmission. Neuroscience 120: 133-352. Thomas, E., Sjövall, H. and Bornstein, J. (2004). Computational model of the migrating motor complex of the small intestine. Am. J. Physiol 286: 564-572. Thornton, P. and Bornstein, J. (2002). Slow excitatory synaptic potentials evoked by distension in myenteric descending interneurons of guinea-pig ileum. J. Physiol.(Lond.) 539: 589–602. Thwaites, D., Basterfield, L., Mccleave, P., Carter, S. and Simmons, N. (2000). Gamma- aminobutyric acid (GABA) transport across human intestinal epithelial (Caco-2) cell monolayers. British Journal of Pharmacology 129: 457-464. Tian, J., Chau, C., Hales, T. and Kaufman, D. (1999). GABA(A) receptors mediate inhibition of T cell responses. J Neuroimmunol 96: 21–28. Tian, L., Hires, S., Mao, T., Huber, D., Chiappe, M., Chalasani, S., Petreanu, L., Akerboom, J., Mckinney, S., Schreiter, E., Bargmann, C., Jayaraman, V., Svoboda, K. and Looger, L. (2009). Imaging neural activity in worms, flies and mice with improved GCaMP calcium indicators. Nat Methods 6: 875-881. Timmermans, J., Scheuermann, D. and Barbiers, M. (1992). Calcitonin gene-related peptide- like immunoreactivity in the human small intestine. Acta. Anat 143: 48-53. Tonini, M., Crema, A., Frigo, G., Rizzi, C., Manzo, L., Candura, S. and Onori, L. (1989a). An in vitro study of the relationship between GABA receptor function and propulsive motility in the distal colon of the rabbit. Br.J.Pharmacol 98: 1109–1118. 239

Tonini, M., De Petris, G., Onori, L., Manzo, L., Rizzi, C. and Crema, A. (1989b). The role of GABAA receptor function in peristaltic activity of the guinea-pig ileum: a comparative study with bicuculline, SR 95531 and picrotoxinin. Br J Pharmacol 97: 556 –562. Tonini, M., Onori, L., Rizzi, C., Perucca, E., Manzo, L. and Crema, A. (1987). Involvement of substance P in the excitatory action of GABAA agonists on cholinergic neurons in the guinea-pig ileum. Naunyn Schmiedebergs Arch Pharmacol 335: 629–35. Triadafilopoulos, G., Pothoulakis, C., O'brien, M. and Lamont, J. (1987). Differential effects of Clostridium difficile toxins A and B on rabbit ileum. Gastroenterology 93: 273- 279. Triadafilopoulos, G., Pothoulakis, C., Weiss, R., Giampaolo, C. and Lamont, J. (1989). Comparative study of Clostridium difficile toxin A and cholera toxin in rabbit ileum. Role of prostaglandins and leukotrienes. Gastroenterology 97: 1186–1192.

Turco, F., Sarnelli, G., Cirillo, C., Palumbo, I., De Giorgi, F., D'alessandro, A., Cammarota, M., Giuliano, M. and Cuomo, R. (2014). Enteroglial-derived S100B protein integrates bacteria-induced Toll-like receptor signalling in human enteric glial cells. Gut 63: 105-115

Turvill, J., Connor, P. and Farthing, M. (2000a). Neurokinin 1 and 2 receptors mediate cholera toxin secretion in rat jejunum. . Gastroenterology 119: 1037-1044.

Turvill, J. and Farthing, M. (1997). Effect of granisetron on cholera toxin-induced enteric secretion. .Lancet 349: 1293. Turvill, J., Mourad, F. and Farthing, M. (1998). Crucial role for 5-HT in cholera toxin but not Escherichia coli heat-labile enterotoxin-intestinal secretion in rats. Gastroenterology 115: 883-890. Vanden Berghe, P., Kenyon, J. and Smith, T. (2002). Mitochondrial Ca2+ uptake regulates the excitability of myenteric neurons. Journal of Neuroscience 22: 6962-6971. Vanner, S. and Macnaughton, W. (1995). Capsaicin-sensitive afferent neurons activate submucosal secretomotor neurons in guinea pig ileum. Am. J. Physiol. (Gastrointest. Liver Physiol) 269: G203–G209. Vanner, S. and Macnaughton, W. (2004). Submucosal secretomotor and vasodilator reflexes. Neurogastroenterology and Motility 16 Suppl 1: 39-43.

240

Vantrappen, G., Janssens, J., Peeters, T., Bloom, S., Christofides, N. and Hellemans, J. (1979). Motilin and the interdigestive migrating motor complex in man. Dig. Dis. Sci 24: 497-500. Varela, C., Blanco, R. and De La Villa, P. (2005). Depolarizing effect of GABA in rod bipolar cells of the mouse retina. Vision Res 45: 2659–2667. Vogalis, F., Harvey, J. and Furness, J. (2002a). TEA- and apamin –resistant KCa channels in guinea pig myenteric neurons: slow AHP channels. J. Physiol. (Lond.) 538: 421-433. Vogalis, F., Harvey, J. and Furness, J. (2003). PKA-mediated inhibition of a novel K⁺ channel underlies the slow after hyperpolarisation in enteric AH neurons. J. Physiol. (Lond.) 548: 801-814. Vogalis, F., Harvey, J., Lohman, R.-J. and Furness, J. (2002b). Action potential after depolarisation mediated by a Ca²⁺-activated cation conductance in myenteric AH neurons. Neuroscience 115: 375-393. Vogalis, F., Jb, F. and Kunze, W. (2001). After hyperpolarisation current in myenteric neurons of the guinea pig duodenum. J. Neurophysiol 85: 1941-1951. Washington, M., Murry, C., Raboin, S., Roberson, A., Mansour, M., Williams, C. and Sayegh, A. (2011). Cholecystokinin-8 activates myenteric neurons in 21-and 35-day old but not 4-and 14-day old rats. Peptides 32: 272-280. Weber, E., Neunlist, M., Schemann, M. and Frieling, T. (2001). Neural components of distension-evoked secretory responses in the guinea-pig distal colon. Journal of Physiology 536: 741-751. Weiser, M. and Quill, H. (1975). Intestinal villus and crypt cell responses to cholera toxin. . Gastroenterology 69: 479-482. Wershil, B., Castagliuolo, I. and Pothoulakis, C. (1998). Mast cell involvement in Clostridium difficile toxin A-induced intestinal fluid secretion and neutrophil recruitment in mice. Gastroenterology 114: 956-964.

WHO (2016). Antibiotics resistance Fact sheet October 2016, http://www.who.int/mediacentre/factsheets/fs107/en/

WHO (2017). Diarrhoeal disease fact sheet May 2017, http://www.who.int/mediacentre/factsheets/fs330/en/.

241

Williamson, S., Pompolo, S. and Furness, J. (1996). GABA and NOS immunoreactivities are colocalised in a subset of inhibitory motor neurons of the guinea-pig small intestine. Cell and Tissue Research 284: 29-37. Wilson, A., Llewellyn Smith, I., Furness, J. and Costa, M. (1987). The source of the nerve fibres innertvating the circular muscle and forming the deep muscular plexus in the guinea pig small intestine. Cell Tissue Res 247: 497-504. Wong, V., Blennerhassett, M. and Vanner, S. (2008). Electrophysiological and morphological properties of submucosal neurons in the mouse distal colon. Neurogastroenterol Motil 20: 725–734. Wood, J. 1994. Physiology of the enteric nervous system, New York, Raven press. Wood, J. (2006). Histamine, mast cells, and the enteric nervous system in the irritable bowel syndrome, enteritis, and food allergies. Gut 55: 445–447. Wood, J. and Mayer, C. (1978). Intracellular study of electrical activity of Auerbach’s plexus in guinea pig small intestine. Pflügers Arch. Eur. J. Physiol 374: 265-275. Wood, J. and Mayer, C. (1979a). Intracellular study of tonic-type enteric neurons in guinea pig small intestine. J. Neurophysiol 42: 569-581. Wu, C., Qin, X., Du, H., Li, N., Ren, W. and Peng, Y. (2017). The immunological function of GABAergic system. Frontiers in Bioscience 22: 162-1172. Xia, Y., Hu, H., Liu, S., Pothoulakis, C. and Wood, J. (2000). Clostridium difficile toxin A excites enteric neurones and suppresses sympathetic neurotransmission in the guinea pig. Gut 46: 481-486. Xiao, J., Nguyen, T., Ngui, K., Strijbos, P., Selmer, I., Neylon, C. and Furness, J. (2004). Molecular and functional analysis of hyperpolarisation-activated nucleotide gated (HCN) channels in the enteric nervous system. Neuroscience 129: 603–614. Xue, J., Askwith, C., Javed, N. and Cooke, H. (2007). Autonomic nervous system and secretion across the intestinal mucosal surface. Autonomic Neuroscience: Basic and Clinical 133: 55-63. Yamada, Y. and Mikoshiba, K. (2012). Quantitative comparison of novel GCaMP-type genetically encoded Ca(2+) indicators in mammalian neurons. Front. Cell. Neurosci 6: 41. Yano, J., Yu, K., Donaldson, G., Shastri, G., Ann, P., Ma, L., Nagler, C., Ismagilov, R., Mazmanian, S. and Hsiao, E. (2015). Indigenous bacteria from the gut microbiota regulate host serotonin biosynthesis. Cell 161: 264-276. 242

Zariwala, H., Borghuis, B., Hoogland, T., Madisen, L., Tian, L., De Zeeuw, C., Zeng, H., Looger, L., Svoboda, K. and Chen, T. (2012). A Cre-dependent GCaMP3 reporter mouse for neuronal imaging in vivo. Journal of Neuroscience 32: 3131-3141. Zhou, X. and Galligan, J. (1999). Synaptic activation and properties of 5- hydroxytryptamine(3) receptors in myenteric neurons of guinea pig intestine. The Journal of Pharmacology and Experimental Therapeutics 290: 803-810. Zhou, X. and Galligan, J. (2000). GABA(A) receptors on calbindin-immunoreactive myenteric neurons of guinea pig intestine. J Auton Nerv Syst 78: 122-135. Zizzo, M., Mule, F. and Serio, R. (2007). Functional evidence for GABA as modulator of the contractility of the longitudinal muscle in mouse duodenum: role of GABA (A) and GABA(C)receptors. Neuropharmacology 52: 1685–1690.

243

Minerva Access is the Institutional Repository of The University of Melbourne

Author/s: Koussoulas, Katerina

Title: Neural mechanisms involved in enterotoxin- induced intestinal hypersecretion

Date: 2017

Persistent Link: http://hdl.handle.net/11343/208989

File Description: Neural mechanisms involved in enterotoxin- induced intestinal hypersecretion

Terms and Conditions: Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the copyright owner. The work may not be altered without permission from the copyright owner. Readers may only download, print and save electronic copies of whole works for their own personal non-commercial use. Any use that exceeds these limits requires permission from the copyright owner. Attribution is essential when quoting or paraphrasing from these works.