Mechanisms of Axonal Pathophysiology in Diabetes and Chronic Kidney Disease
Tushar Issar
A thesis submitted in fulfilment of the requirement for the
degree of Doctor of Philosophy
Prince of Wales Clinical School
University of New South Wales
July 2020 Thesis/Dissertation Sheet Australia's Global University
Surname/Family Name lssar Given Name/s Tushar Abbreviation for degree as give in the University calendar PhD Faculty Medicine School Prince of Wales Clinical School Thesis Title Mechanisms of axonal pathophysiology in diabetes and chronic kidney disease
Abstract 350 words maximum: (PLEASE TYPE) Peripheral neuropathy is a common and debilitating complication of diabetes and chronic kidney disease (CKD). The pathophysiological mechanisms contributing to peripheral neuropathy in these conditions remain unclear. In this thesis, axonal excitability studies and nerve ultrasonography were utilised to assess peripheral nerve structure and function in human subjects with either diabetes, CKD, or both to investigate the mechanisms underlying peripheral nerve dysfunction in each condition.
To commence, it was essential that an instrument to assess the severity of neuropathy in patients CKD with and without type 2 diabetes (T2DM) was formally validated. Having validated The Total Neuropathy Score in Chapter 1, axon al excitability studies were then utilised to determine the relative contributions of T2DM and CKD underlying nerve dysfunction in diabetic kidney disease in Chapter 2. It was established that CKD, and not diabetes, underlies axonal pathophysiology in patients with diabetic kidney disease.
Studies were then conducted in autoimmune diabetes. In type 1 diabetes (TlDM), despite good glycaemic control as measured by HbA1c, development of peripheral neuropathy frequently occurs. In search of an explanation, the association between acute glucose control and nerve structure and function was explored in Chapter 3. Greater acute glucose variability and longer time spent in hyperglycaemia were associated with worse nerve excitability measures, altered corneal nerve morphology, and a higher number of corneal micro-neuromas. In Chapter 4, the mechanisms underlying axonal dysfunction in a recently recognised form of autoimmune diabetes known as latent autoimmune diabetes in adults (LADA) were then investigated. The basis of nerve dysfunction in LADA was different to TlDM and T2DM, and LADA patients exhibited more severe changes in nerve excitability and ultrasound measures.
Investigations were then undertaken in T2DM. There is a strong association between the metabolic syndrome (Mets) and the development of peripheral neuropathy. To explain this relationship, the effect of the metabolic syndrome (MetS) in T2DM was examined in Chapter 5. A reduction in the function of the Na+/K+ pump was found to explain the more severe changes in nerve structure and function in T2DM patients with Mets compared to patients with T2DM alone. Finally, in exploration of potential neuroprotective options for peripheral neuropathy, the effect of anti-diabetic medication on nerve function was then investigated in Chapter 6. Exenatide treatment was associated with better nerve function in cross-sectional and prospective studies. Prominent abnormalities remained in patients receiving SGLT2 inhibitor or DPP-4 inhibitor therapy.
Declaration relating to disposition of project thesis/dissertation
I hereby grant to lhe University of New South Wales or its agents a non-exclusive licence to archive and lo make available (including to members of the public) my thesis or disserlation in whole or in part in lhe University libraries in all forms of media, now or here after known. I acknowledge that I retain all intellectual property rights which subsist in my thesis or dissertation. such as copyright and patent rights, subject to applicable law. I also retain the right to use all or part of my thesis or dissertation in future works (such as articles or books).
.. , , , ...... ,44, .. (•., ..... ,,,,,, ... ,.... , .... ,,,, ...... §/-?:(:.::��( ...... , .. . Si nature Date The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years can be made when submitting the final copies of your thesis to the UNSW Library. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research. ORIGINALITY STATEMENT
‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’
Signed ……………………………………………......
Date ……………………………………………...... COPYRIGHT STATEMENT
‘I hereby grant the University of New South Wales or its agents a non-exclusive licence to archive and to make available (including to members of the public) my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known. I acknowledge that I retain all intellectual property rights which subsist in my thesis or dissertation, such as copyright and patent rights, subject to applicable law. I also retain the right to use all or part of my thesis or dissertation in future works (such as articles or books).’
‘For any substantial portions of copyright material used in this thesis, written permission for use has been obtained, or the copyright material is removed from the final public version of the thesis.’
Signed ……………………………………………......
Date ……………………………………………......
AUTHENTICITY STATEMENT ‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis.’
Signed ……………………………………………......
Date ……………………………………………...... ( ...... � � INCLUSION OF PUBLICATIONS STATEMENT ; Austral1a·s Global Ut:��yv University
UNSW is supportive of candidates publishing their research results during their candidature as detailed in the UNSW Thesis Examination Procedure.
Publications can be used in their thesis in lieu of a Chapter if: • The candidate contributed greater than 50% of the content in the publication and is the "primary author", ie. the candidate was responsible primarily for the planning, execution and preparation of the work for publication • The candidate has approval to include the publication in their thesis in lieu of a Chapter from their supervisor and Postgraduate Coordinator. • The publication is not subject to any obligations or contractual agreements with a third party that would constrain its inclusion in the thesis
Please indicate whether this thesis contains published material or not:
This thesis contains no publications, either published or submitted for publication □ (if this box is checked, you may delete all the material on page 2)
Some of the work described in this thesis has been published and it has been documented in the relevant Chapters with acknowledgement □ (if this box is checked, you may delete all the material on page 2)
This thesis has publications (either published or submitted for publication) incorporated into it in lieu of a chapter and the details are presented below
CANDIDATE'S DECLARATION I declare that: • I have complied with the UNSW Thesis Examination Procedure • where I have used a publication in lieu of a Chapter, the listed publication(s) below meet(s) the requirements to be included in the thesis. Candidate's Name Date (dd/mm/yy) Tushar lssar S/t/::2.i?zr POSTGRADUATE COORDINATOR'S DECLARATION To only be filled in where publications are used in lieu of Chapters I declare that:
G the information below is accurate G where listed publication(s) have been used in lieu of Chapter(s), their use complies with the UNSW Thesis Examination Procedure G the minimum requirements for the format of the thesis have been met. PGC's Name PGC's Signature Date (dd/mm/yy)
For each publication incorporated into the thesis in lieu of a Chapter, provide all of the requested details and signatures required Details of publication #1: Full title: The utility of the Total Neuropathy Score as an instrument to assess neuropathy severity in chronic kidney disease: A validation study Authors: lssar T, Arnold R, Kwai, NCG, Pussell, BA, Endre, ZH, Poynten, AM, Kiernan, MC, Krishnan, AV Journal or book name: Clinical Neurophysiology Volume/page numbers: Volume 129, Issue 5, Pages 889-894 Date accepted/ published: Accepted 4th February 2018 Status Published X Accepted and In ress The Candidate's Contribution to the Work Tushar lssar was responsible for the study design, recruitment, data collection, data interpretation, and the manuscript composition. Location of the work in the thesis and/or how the work is incorporated in the thesis: Chapter 1. Validates the Total Neuropathy Score as a means to assess peripheral neuropathy in chronic kidney disease, which was required for subsequent studies. PRIMARY SUPERVISOR'S DECLARATION I declare that: • the information above is accurate • this has been discussed with the PGC and it is agreed that this publication can be included in this thesis in lieu of a Chapter • All of the co-authors of the publication have reviewed the above information and have a reed to its veracit b si a 'Co-Author Authorisation' form. Primary Supervisor's name Primary Su isor's signature Date (dd/mm/yy) Arun Krishnan c7)/ OrX Io{J Details of publication #2: Full title: Relative contributions of diabetes and chronic kidney disease to neuropathy development in diabetic nephropathy patients. Authors: lssar T, Arnold R, Kwai NCG, Walker S, Yan A, Borire AA, Poynten AM, Pussell BA, Endre ZH, Kiernan MC, Krishnan AV Journalor book name: Clinical Neurophysiology Volume/page numbers: Volume 130, Issue 11, Pages 2088-2095 Date accepted/ published: Accepted 12th August 2019 Status Published X Accepted and In In progress ress submitted The Candidate's Contribution to the Work Tushar lssar was responsible for the study design, recruitment, data collection, data interpretation, and the manuscript composition. Location of the work in the thesis and/or how the work is incorporated in the thesis: Chapter 2. Investigates the effects of diabetes and chronic kidney disease on peripheral nerve dysfunction in diabetic kidney disease.
PRIMARYI declare that: SUPERVISOR'S DECLARATION • the information above is accurate . this has been discussed with the PGC and it is agreed that this publication can be included in this thesis in lieu of a Chapter . All of the co-authors of the publication have reviewed the above information and have aqreed to its veracity by siqninq a 'Co-Author Authorisation' form.
PrimaryArun Krishnan Supervisor's name Primary Date ( dd/mm/yy) Supervisor's signature 1 1510� / � I
FullDetails title: of Associations publication between #3: acute glucose control and peripheral nerve structure and function in type 1 diabetes Authors: lssar T, Tummanapalli SS, Kwai NCG, Chiang JCB, Arnold R, Poynten AM, Markoulli M, Krishnan AV Journal or book name: Diabetic Medicine Volume/page numbers: Online (ahead of print) Date accepted/ published: Accepted 9th April 2020 Accepted and In In progress Published X \ Status I I press I (submitted) I TusharThe Candidate's lssar was responsibleContribution for to the the study Work design, recruitment, data collection (with exception of in-vivo corneal confocal microscopy), data interpretation, and the manuscript composition. Location of the work in the thesis and/or how the work is incorporated in the thesis: Chapter 3. Explores the association between acute glucose variation and peripheral nerve structure and function in type 1 diabetes.
PRIMARYI declare that: SUPERVISOR'S DECLARATION . the information above is accurate . this has been discussed with the PGC and it is agreed that this publication can be included in this thesis in lieu of a Chapter . All of the co-authors of the publication have reviewed the above information and have aqreed to its veracity by siqninq a 'Co-Author Authorisation' form.
PrimaryArun Krishnan Supervisor's name Primary Su,ltervisor's signature Date (dd/mm/yy) tJ
FullDetails title: of Altered publication peripheral #4: nerve structure and function in latent autoimmune diabetes in adults Authors: lssar T, Yan A, Kwai NCG, Poynten AM, Borire AB, Arnold R, Krishnan AV Journalor book name: Diabetes/Metabolism Research and Reviews Volume/page numbers: Volume 36, Issue 3, e3260 Date accepted/ published: Accepted 27th November 2019 \ Published X \ Accepted and In \ 'n progress Status press I (submitted) I The Candidate's Contribution to the Work Tushar lssar was responsible for the study design, recruitment, data collection, data inter retation, and the manuscri t com osition. Location of the work in the thesis and/or how the work is incorporated in the thesis: Chapter 4. Investigates the pathophysiological changes in peripheral nerve structure and function in Latent Autoimmune Diabetes in Adults in comparison with type 1 and 2 diabetes PRIMARY SUPERVISOR'S DECLARATION I declare that: • the information above is accurate • this has been discussed with the PGC and it is agreed that this publication can be included in this thesis in lieu of a Chapter • All of the co-authors of the publication have reviewed the above information and have a reed to its veracit b si a 'Co-Author Authorisation' form. Primary Supervisor's name Primary Sup rvisor's signature Date (dd/mm/yy) Arun Krishnan (Jf7 0 o2 Id I' Details of publication #5: Full title: Impact of the metabolic syndrome on peripheral nerve structure and function in type 2 diabetes Authors: lssar T, Tummanapalli SS, Borire AB, Kwai NCG, Poynten AM, Arnold R, Markoulli M, Krishnan AV Journal or book name: Volume/page numbers: Date accepted/ published: Status Published Accepted and In In progress X ress submitted The Candidate's Contribution to the Work Tushar lssar was responsible for the study design, recruitment, data collection (with exception of in-vivo corneal confocal microscopy), data interpretation, and the manuscript com osition. Location of the work in the thesis and/or how the work is incorporated in the thesis: Chapter 5. Investigates the pathophysiological changes in peripheral nerve structure and function in type 2 diabetes with the metabolic syndrome in comparison to type 2 diabetes alone. PRIMARY SUPERVISOR'S DECLARATION I declare that: • the information above is accurate • this has been discussed with the PGC and it is agreed that this publication can be included in this thesis in lieu of a Chapter • All of the co-authors of the publication have reviewed the above information and have a reed to its veracit b si a 'Co-Author Authorisation' form. Primary Supervisor's name Primary Su Date (dd/mm/yy) Arun Krishnan ()� 7rz; I¢2 I - Details of publication #6: Full title: Effect of exenatide on peripheral nerve function in type 2 diabetes Authors: lssar T, Kwai NCG, Poynten AM, Arnold R, Milner KL, Krishnan, AV Journal or book name: Volume/page numbers: Date accepted/ published: Status Published Accepted and In In progress X ress submitted The Candidate's Contribution to the Work Tushar lssar was responsible for the study design, recruitment, data collection, data interpretation, and the manuscript composition. Location of the work in the thesis and/or how the work is incorporated in the thesis: Chapter 6. Explores the effect of exenatide, SGL T2 inhibition, or DPP-4 inhibition on peripheral nerve function in type 2 diabetes. PRIMARY SUPERVISOR'S DECLARATION I declare that: • the information above is accurate . this has been discussed with the PGC and it is agreed that this publication can be included in this thesis in lieu of a Chapter . All of the co-authors of the publication have reviewed the above information and have aqreed to its veracity by siqning a 'Co-Author Authorisation' form. Primary Supervisor's name Primary Supervisor's signature Date (dd/mm/yy) Arun Krishnan o� lo;;_/ �I Acknowledgements
This thesis would not have been achievable without the guidance of my supervisors. I am forever grateful to Professor Arun Krishnan for his wisdom, enthusiasm, inspiration, and endless support. I will always remember his tremendous efforts throughout my career, and I owe my achievements to him. I am eternally thankful to have had the support of Dr Ann Poynten. Her invaluable insight and admirable expertise made the studies of this thesis possible. I am especially appreciative for her kindness, patience, and generosity in providing me the facilities to conduct research. I would also like to express my gratitude to Dr Natalie Kwai and Dr Ria Arnold for their valued advice throughout my studies and instrumental efforts in imparting their technical knowledge and skills.
I would like to thank all of my research colleagues at the Institute of
Neurological Sciences for making my time there so enjoyable. Shyam and
Aimy, thank you for your valuable assistance and support. It was a such a pleasure to work with you both and I wish you every success in the future. I would also like to thank Terry, Jeremy, Kimberly, and Vincent for providing endless laughter and positivity throughout the difficult periods that comes with research.
This thesis would not have been possible without the help of several other important people. I would like to express my appreciation for having the privilege to learn from Professor Hugh Bostock. His generosity in taking the
ii time educate me in neurophysiology is an experience I will carry throughout my career and his contributions to Science will forever be an inspiration. I would also like to thank Dr Adeniyi Borire and Dr Maria Markoulli for their expertise and contributions to my manuscripts. I am grateful for the technical support of Mr Tony Yakoubi, who was always willing to lend a helping hand. I would like to thank Mrs Analiza Santiago, Mrs Isabell Ghikas, and Mrs
Collette Brodie for their assistance with recruiting study participants. I would like to express my sincere gratitude for the patients that generously donated hours of their time for these studies.
Finally, this body of work would not have been possible without the support of my loving family. My parents, Arvind and Meena, have made tremendous sacrifices throughout their lives to support me in a countless number of ways.
My brother Aarush and my sister Misha, thank you for your endless support and encouragement. You all mean everything to me.
iii Abstract
Peripheral neuropathy is a common and debilitating complication of diabetes and chronic kidney disease (CKD). The pathophysiological mechanisms contributing to peripheral neuropathy in these conditions remain unclear. In this thesis, axonal excitability studies and nerve ultrasonography were utilised to assess peripheral nerve structure and function in human subjects with either diabetes, CKD, or both to investigate the mechanisms underlying peripheral nerve dysfunction in each condition.
To commence, it was essential that an instrument to assess the severity of neuropathy in patients CKD with and without type 2 diabetes (T2DM) was formally validated. Having validated The Total Neuropathy Score in Chapter
1, axonal excitability studies were then utilised to determine the relative contributions of T2DM and CKD underlying nerve dysfunction in diabetic kidney disease in Chapter 2. It was established that CKD, and not diabetes, underlies axonal pathophysiology in patients with diabetic kidney disease.
Studies were then conducted in autoimmune diabetes. In type 1 diabetes
(T1DM), despite good glycaemic control as measured by HbA1c, development of peripheral neuropathy frequently occurs. In search of an explanation, the association between acute glucose control and nerve structure and function was explored in Chapter 3. Greater acute glucose variability and longer time spent in hyperglycaemia were associated with worse nerve excitability measures, altered corneal nerve morphology, and a higher number of corneal micro-neuromas. In Chapter 4, the mechanisms underlying axonal dysfunction
iv in a recently recognised form of autoimmune diabetes known as latent autoimmune diabetes in adults (LADA) were then investigated. The basis of nerve dysfunction in LADA was different to T1DM and T2DM, and LADA patients exhibited more severe changes in nerve excitability and ultrasound measures.
Investigations were then undertaken in T2DM. There is a strong association between the metabolic syndrome (MetS) and the development of peripheral neuropathy. To explain this relationship, the effect of the metabolic syndrome
(MetS) in T2DM was examined in Chapter 5. A reduction in the function of the Na+/K+ pump was found to explain the more severe changes in nerve structure and function in T2DM patients with MetS compared to patients with
T2DM alone. Finally, in exploration of potential neuroprotective options for peripheral neuropathy, the effect of anti-diabetic medication on nerve function was then investigated in Chapter 6. Exenatide treatment was associated with better nerve function in cross-sectional and prospective studies. Prominent abnormalities remained in patients receiving SGLT2 inhibitor or DPP-4 inhibitor therapy.
v Publications
Publications resulting from studies undertaken in this doctorate
Chapter 1
Issar, T., Arnold, R., Kwai, N. C. G., Pussell, B. A., Endre, Z. H., Poynten, A. M., . . .
Krishnan, A. V. (2018). The utility of the Total Neuropathy Score as an instrument to
assess neuropathy severity in chronic kidney disease: A validation study. Clinical
Neurophysiology, 129(5), 889-894. doi: 10.1016/j.clinph.2018.02.120
Chapter 2
Issar, T., Arnold, R., Kwai, N. C. G., Walker, S., Yan, A., Borire, A. A., . . . Krishnan, A. V.
(2019). Relative contributions of diabetes and chronic kidney disease to neuropathy
development in diabetic nephropathy patients. Clinical Neurophysiology, 130(11),
2088-2095. doi: 10.1016/j.clinph.2019.08.005
Chapter 3
Issar, T., Tummanapalli, S. S., Kwai, N. C. G., Chiang, J. C. B., Arnold, R., Poynten, A. M.,
. . . Krishnan, A. V. (2020). Associations between acute glucose control and
peripheral nerve structure and function in type 1 diabetes. Diabetic Medicine. doi:
10.1111/dme.14306
Chapter 4
Issar, T., Yan, A., Kwai, N. C. G., Poynten, A. M., Borire, A. A., Arnold, R., & Krishnan, A.
V. (2020). Altered peripheral nerve structure and function in latent autoimmune
diabetes in adults. Diabetes/Metabolism Research and Reviews, 36(3), e3260. doi:
10.1002/dmrr.3260
vi Associated publications achieved during candidature
Habeych, M., Trinh, T., Issar, T., Kwai, N. C. G., & Krishnan, A.V. (2020). Motor
unit number estimation of facial muscles using the M Scan-Fit method.
Muscle & Nerve, In press doi: 10.1002/mus.27010
Tummanapalli, S. S., Issar, T., Kwai, N. C. G., Pisarcikova, J., Poynten, A. M., Krishnan, A.
V., . . . Markoulli, M. (2019). A Comparative Study on the Diagnostic Utility of
Corneal Confocal Microscopy and Tear Neuromediator Levels in Diabetic Peripheral
Neuropathy. Current Eye Research, 45(8), 921-903.
doi: 10.1080/02713683.2019.1705984
Tummanapalli, S. S., Issar, T., Yan, A., Kwai, N. C. G., Poynten, A. M., Krishnan, A. V., . . .
Markoulli, M. (2019). Corneal nerve fiber loss in diabetes with chronic kidney
disease. The Ocular Surface, 18(1), 178-185. doi: 10.1016/j.jtos.2019.11.010
Tummanapalli, S. S., Issar, T., Kwai, N. C. G., Poynten, A., Krishnan, A. V., Willcox, M., &
Markoulli, M. (2019). Association of corneal nerve loss with markers of axonal ion
channel dysfunction in type 1 diabetes. Clinical Neurophysiology, 131(1), 145-154.
doi: 10.1016/j.clinph.2019.09.029
Tummanapalli, S. S., Willcox, M. D. P., Issar, T., Yan, A., Pisarcikova, J., Kwai, N. C. G., . .
. Markoulli, M. (2019). Tear film substance P: A potential biomarker for diabetic
peripheral neuropathy. The Ocular Surface, 17(4), 690-698. doi:
10.1016/j.jtos.2019.08.010
vii Tummanapalli, S. S., Willcox, M. D. P., Issar, T., Kwai, N. C. G., Poynten, A. M., Krishnan,
A. V., . . . Markoulli, M. (2019). The Effect of Age, Gender and Body Mass Index on
Tear Film Neuromediators and Corneal Nerves. Current Eye Research, 45(4), 411-
418. doi: 10.1080/02713683.2019.1666998
Borire, A. A., Issar, T., Kwai, N. C. G., Visser, L. H., Simon, N. G. C., Poynten, A. M., . . .
Krishnan, A. V. (2019). Sonographic assessment of nerve blood flow in diabetic
neuropathy. Diabetic Medicine, 37(2), 343-349. doi: 10.1111/dme.14085
Yan, A., Issar, T., Tummanapalli, S. S., Markoulli, M., Kwai, N. C. G., Poynten, A. M., &
Krishnan, A. V. (2019). Relationship between corneal confocal microscopy and
markers of peripheral nerve structure and function in Type 2 diabetes. Diabetic
Medicine, 37(2), 326-334. doi: 10.1111/dme.13952
Borire, A. A., Issar, T., Kwai, N. C. G., Visser, L. H., Simon, N. G., Poynten, A. M., . . .
Krishnan, A. V. (2018). Correlation between markers of peripheral nerve function
and structure in type 1 diabetes. Diabetes/ Metabolism Research Reviews, 34(7),
e3028. doi: 10.1002/dmrr.3028
Arnold, R., Issar, T., Krishnan, A. V., & Pussell, B. A. (2016). Neurological complications in
chronic kidney disease. JRSM Cardiovascular Disease, 5, 2048004016677687. doi:
10.1177/2048004016677687
viii Awards and Presentations
Awards
1. William McIlrath Travel Scholarship 2018
2. Prince of Wales Hospital Equipment Grant 2017, 2018
3. JRSM’s Third Most Downloaded Publication 2017
4. Prince of Wales Hospital Travel Grant 2016, 2017
5. Australian Postgraduate Award Scholarship 2016–2018
Presentations
International Diabetes Federation Congress 2019
Poster presentation: “The effect of hyperglycaemia and glucose variability on peripheral nerve structure and function”
Tow Coast Association Awards 2018
Oral presentation: “Mechanisms of axonal dysfunction in diabetes and chronic kidney disease”
Australian Diabetes Society Annual Scientific Meeting 2016, 2017 o Oral presentation: “GLP-1 agonism alters peripheral nerve function in patients with type 2 diabetes” o Poster presentation: “Effects of GLP-1 agonism on peripheral nerve function in type 2 diabetes”
ix Table of Contents
Acknowledgements...... ii Abstract...... iv Publications...... vi Awards...... ix Presentations...... ix Table of Contents...... x Index of Tables...... xii Index of Figures...... xiv Abbreviations...... xvii Literature Review...... 1 Diabetes Mellitus...... 2 Type 1 Diabetes Mellitus...... 3 Latent Autoimmune Diabetes of Adults...... 4 Type 2 Diabetes Mellitus and the Metabolic Syndrome...... 6 Chronic Kidney Disease...... 8 Diabetic Kidney Disease...... 9 The Human Nervous System...... 11 The Peripheral Axon...... 15 Structure of Peripheral Axons...... 15 Function of Peripheral Axons...... 19 Ion Channels, Pumps, and Exchangers...... 23 Clinical Assessment of Peripheral Nerve Structure and Function...... 38 Axonal Excitability...... 41 Nerve Ultrasonography...... 58 In-vivo Corneal Confocal Microscopy...... 61 Diabetic Neuropathy...... 65 Uraemic Neuropathy...... 75 Methodology...... 80 Recruitment of Patients and Control Subjects...... 81 Equipment and Materials...... 82
x Axonal Excitability Assessment and Mathematical Modelling...... 84 Nerve Ultrasonography...... 92 In-vivo Corneal Confocal Microscopy...... 93 Nerve Conduction Studies...... 95 Clinical Assessment of Neuropathy...... 95 Glucose Variability...... 99 Metabolic Syndrome Components...... 101 Statistical Analyses...... 102 Chapter 1 – Validation of the Total Neuropathy Score as a means to assess peripheral neuropathy in chronic kidney disease...... 103 Chapter 2 – Investigation of the effect of diabetes and chronic kidney disease on axonal pathophysiology...... 120 Chapter 3 – Association between acute glucose control and axonal function and structure in type 1 diabetes...... 143 Chapter 4 – Axonal pathophysiology in Latent Autoimmune Diabetes of Adulthood...... 165 Chapter 5 – Pathophysiological mechanisms underlying altered axonal function and structure in type 2 diabetes with metabolic syndrome...... 187 Chapter 6 – Effect of exenatide on axonal function in type 2 diabetes...... 210 Summary and Future Directions...... 230 References...... 238
xi Index of Tables
Literature Review and Methodology
Table 1. Classification of chronic kidney disease...... 9
Table 2. Nerve excitability testing paradigms...... 53
Chapter 1
Table 1.1. Components of the total neuropathy score...... 107
Table 1.2. Subject demographics...... 112
Table 1.3. Internal consistency of the TNS...... 114
Table 1.4. Comparison of TNS item scores between groups...... 115
Table 1.5. Comparison of TNS item scores between subgroups and controls...... 116
Chapter 2
Table 2.1. Subject demographics...... 131
Table 2.2. Total neuropathy score and subscore comparison...... 132
Table 2.3. Nerve excitability findings...... 134
Table 2.4. Modelled parameters between DKD, CKD, T2DM, and control subjects...... 138
xii Chapter 3
Table 3.1. Participant demographics...... 154
Table 3.2. Partial correlations between acute glycaemic metrics and measures of peripheral nerve structure and function...... 156
Chapter 4
Table 4.1. Participant characteristics...... 176
Table 4.2. Modelled parameters for LADA, type 1 diabetes, type 2 diabetes and control cohorts...... 182
Chapter 5
Table 5.1. Participant demographics...... 199
Table 5.2. Corneal confocal microscopy cohorts...... 202
Table 5.3. Corneal confocal microscopy correlations...... 204
Chapter 6
Table 6.1. Subject demographics and clinical measures...... 220
Table 6.2. Prospective exenatide cohort clinical measures at baseline and at 3-month follow-up...... 224
xiii Index of Figures
Literature Review and Methodology
Figure 1. Schematic of a motor neuron and domains of a myelinated axon...... 14
Figure 2. The action potential...... 21
Figure 3. Distribution of ion channels, pumps, and exchangers in a myelinated axon...... 25
Figure 4. Stimulus-response curve...... 43
Figure 5. Strength-duration relationship...... 45
Figure 6. Threshold electrotonus...... 47
Figure 7. Current-threshold relationship...... 50
Figure 8. Recovery cycle...... 51
Figure 9. Sonograph of healthy and neuropathic median nerve...... 61
Figure 10. Confocal micrograph of healthy and neuropathic cornea...... 64
Figure 11. Pathogenesis of distal symmetric polyneuropathy...... 68
Figure 12. Median motor nerve excitability set-up...... 86
Figure 13. Mathematical model parameters...... 91
Figure 14. Median nerve ultrasonography...... 93
Figure 15. Corneal confocal microscopy...... 94
Figure 16. Total Neuropathy Score...... 96
Figure 17. Analysis of continuous glucose monitoring recordings...... 100
xiv Chapter 1
Figure 1.1. Histogram of TNS results for all patients with CKD compared with controls....115
Chapter 2
Figure 2.1. Schematic of the peripheral nerve membrane highlighting the two compartments of the axon and the parameters that were modelled...... 129
Figure 2.2. Mean excitability data for patients with T2DM, CKD or DKD and healthy controls...... 133
Figure 2.3. Summary of mathematical modelling of nerve excitability data for DKD,
CKD, and T2DM...... 137
Chapter 3
Figure 3.1. Receiver operating characteristic analysis of composite nerve excitability score obtained from participants with diabetes and people without diabetes...... 157
Figure 3.2. Corneal confocal image highlighting a microneuroma of a corneal nerve fibre and representative image of the inferior whorl from a participant with type 1 diabetes...... 159
Chapter 4
Figure 4.1. Median nerve cross-sectional area measurements...... 177
Figure 4.2. Group comparison of nerve excitability parameters...... 179
Figure 4.3. Group comparison of nerve excitability recordings...... 180
xv Chapter 5
Figure 5.1. Median nerve ultrasonography highlighting differences in median nerve cross-sectional area between participants with type 2 diabetes and metabolic syndrome, type 2 diabetes alone, and healthy controls...... 201
Figure 5.2. Representative corneal confocal micrographs comparing differences in nerve structure in the central and corresponding inferior whorl region of the cornea for participants with type 2 diabetes and metabolic syndrome, type 2 diabetes alone, and healthy controls...... 203
Chapter 6
Figure 6.1. Group comparison of nerve excitability measures...... 222
Figure 6.2. Nerve excitability recordings showing threshold electrotonus and the recovery cycle of a 72-year-old male patient with mild neuropathy...... 225
xvi Abbreviations
CCM corneal confocal microscopy
CKD chronic kidney disease
CNS central nervous system
CONGA continuous overall net glycaemic action
CSII continuous subcutaneous insulin infusion
DKD diabetic kidney disease
DPP-4 dipeptidyl peptidase-4
DSPN distal symmetric polyneuropathy
GLP-1 glucagon-like peptide-1
HCN hyperpolarisation-activated cyclic nucleotide-gated
Ih mixed cation conductance through HCN
I/V current-threshold relationship
KIR inwardly rectifying potassium channel
KV voltage-gated potassium channel
LADA latent autoimmune diabetes of adulthood
MDII multiple daily insulin infusion
Nap persistent voltage-gated sodium channel
Nat transient voltage-gated sodium channel
NaV voltage-gated sodium channel
NCS nerve conduction studies
PKC protein kinase C
PNS peripheral nervous system
T1DM type 1 diabetes
T2DM type 2 diabetes
xvii TEd depolarising (conditioning current) threshold electrotonus
TEh hyperpolarising (conditioning current) threshold electrotonus
TNS total neuropathy score
SDTC strength-duration time constant (τSD)
SGLT2 sodium-glucose co-transporter-2
xviii Literature Review
Literature Review
1 Literature Review
Diabetes Mellitus
Diabetes Mellitus (diabetes) defines a state of hyperglycaemia due to reduced, altered or absent insulin secretion and or action. The prevalence of diabetes in adults was estimated to be 463 million in 2019 and is projected to increase to 700 million by
2045 (Saeedi et al., 2019). Diabetes can be classified into the following general categories: type 1 diabetes (due to cellular-mediated autoimmune β-cell destruction, usually leading to absolute insulin deficiency), type 2 diabetes (due to progressive loss β-cell insulin secretion frequently on a background of insulin resistance), gestational diabetes (diabetes diagnosed in second or third trimester of pregnancy that was not overt diabetes prior to gestation) and specific types of diabetes due other causes such as monogenic diabetes syndromes (e.g. maturity-onset diabetes of the young), diseases of the exocrine pancreas (e.g. cystic fibrosis and pancreatitis), and drug- or chemical-induced diabetes (e.g. glucocorticoid use and after organ transplantation) (American Diabetes Association, 2020). This thesis will primarily focus on type 1 and type 2 diabetes. It is important to note that type 1 diabetes and type 2 diabetes are heterogeneous diseases in which clinical presentation and disease progression may vary considerably.
The diagnosis of diabetes may be based on plasma glucose criteria, either fasting plasma glucose (≥7 mmol/L; 126 mg/dL) or 2 hour plasma glucose during a 75g oral glucose tolerance test (≥11.1 mmol/L; 200 mg/dL), or glycated haemoglobin criteria
(≥6.5%; 48 mmol/mol) (American Diabetes Association, 2020). In a patient with classical symptoms of hyperglycaemia or hyperglycaemic crisis, a random plasma glucose ≥ 11.1 mmol/L (200 mg/dL) can be used (American Diabetes Association,
2020). In cases where type 1 diabetes requires confirmation, islet cell autoantibodies
2 Literature Review and autoantibodies to glutamate decarboxylase (GAD65), insulin, the tyrosine phosphatases (IA-2 and IA-2β), and zinc transporter 8 (ZnT8) may be used
(American Diabetes Association, 2020).
Once hyperglycaemia occurs, patients with all forms of diabetes are at risk for developing the same chronic complications, including nephropathy, retinopathy, and neuropathy. Rates of complication progression may differ between diabetes type
(American Diabetes Association, 2020). Diabetes is a leading cause of disability and death globally (International Diabetes Federation, 2015, Vos et al., 2016, Zheng et al.,
2018). In Australia, diabetes and resulting comorbidities account for an annual cost of approximately A$15 billion (Lee et al., 2013). This estimated value is inclusive of the cost for medication and adjuvant therapy directly targeting diabetes, government subsidies, and cost to employers and it is projected to increase (Schofield et al., 2017).
Treatment of diabetes is focused not only on glycaemic control but management of co-morbidities and etiological factors. Clinically, this may be achieved through a combination of drug therapy, insulin supplementation and rigorous lifestyle management involving a number of health care professionals.
Type 1 Diabetes Mellitus
The typical clinical course of type 1 diabetes includes a preclinical phase, presentation of diabetes (at which time patients are usually symptomatic of hyperglycaemia), a partial remission or honeymoon phase, and a continuing requirement for insulin therapy (Haller et al., 2005). Type 1 diabetes is characterised by a specific autoimmune destruction of the Islets of Langerhans in the pancreas, which was first described in the early 1900s (Opie, 1901). Autoimmune destruction of β-cells, the
3 Literature Review insulin secreting cells of the pancreas, provokes the total loss of insulin and co- released C-peptide initiating the requirement for exogenous insulin. In the absence of insulin therapy, patients with type 1 diabetes will eventually progress to metabolic decompensation and life‐threatening diabetic ketoacidosis (Craig et al., 2011). Type 1 diabetes typically arises in childhood or adolescence, but onset can occur in adulthood
(Pozzilli and Pieralice, 2018). Patients with an adult-onset of autoimmune diabetes that do not require insulin-therapy for at least 6 months after diagnosis are distinguished as having latent autoimmune diabetes of adulthood (LADA) (Pozzilli and Pieralice, 2018).
Treatment of Type 1 Diabetes
Type 1 diabetes management involves exogenous synthetic insulin therapy with the goal of mimicking normal physiological insulin secretions patterns (Chiang et al.,
2018). Insulin may be delivered through multiple daily insulin injections (MDII) or continuous subcutaneous insulin infusion (CSII) via pump therapy (Chiang et al.,
2018). MDII involves injecting a slow acting insulin analogue that delivers basal levels of insulin and a rapid acting analogue before meals for carbohydrate clearance with food intake (Chiang et al., 2018). CSII utilises fast acting insulin analogues that automatically perfuse insulin at a basal rate through a cannula placed subcutaneously
(Pozzilli et al., 2016).
Latent Autoimmune Diabetes of Adulthood
Adult-onset autoimmune diabetes encompasses a wide spectrum of heterogeneous genotypes and phenotypes, ranging from classic adult-onset type 1 diabetes to LADA
(Buzzetti et al., 2017). LADA was a term introduced in the 1990s to define a
4 Literature Review subgroup of patients who had non-insulin requiring diabetes that was initially thought to be type 2 diabetes but had autoimmune markers of type 1 diabetes (Tuomi et al.,
1993, Zimmet et al., 1994). In 2005, the Immunology of Diabetes Society proposed three main criteria for the diagnosis of LADA: adult age of onset (>30 years), presence of any islet cell autoantibody, and absence of insulin requirement for at least
6 months after diagnosis (Fourlanos et al., 2005). The major criticism of this criteria is the subjectivity of the clinician’s decision to commence insulin treatment. LADA is more heterogenous than youth-onset autoimmune diabetes and shares genetic, clinical, and metabolic features with both type 1 and type 2 diabetes, which suggests that LADA is an admixture of the two major types of diabetes (Cervin et al., 2008,
Liu et al., 2015, Pozzilli and Pieralice, 2018). Patients with LADA have highly variable rates of β-cell destruction as well as different degrees of insulin resistance and autoimmunity (Pozzilli and Pieralice, 2018). Whether LADA is a distinct disease syndrome or part of an autoimmune continuum is yet to be confirmed (Laugesen et al., 2015).
Treatment of Latent Autoimmune Diabetes of Adulthood
To date, no specific guidelines for the treatment of patients with LADA have been published. Treatment of adults with LADA is currently guided by the clinical intuition and expertise of the physician (Buzzetti et al., 2017). Most patients with LADA are initially treated with therapies intended for type 2 diabetes (Pozzilli and Pieralice,
2018). This approach might result in rapid progression to an insulin-dependent state, especially in patients who have high GAD autoantibody titres (Zampetti et al., 2014).
5 Literature Review
Type 2 Diabetes Mellitus and the Metabolic Syndrome
Type 2 diabetes accounts for 90-95% of all diabetes and encompasses individuals with relative (rather than absolute) insulin deficiency and peripheral insulin resistance
(American Diabetes Association, 2020). Although the specific aetiologies are not known, autoimmune destruction of β-cells does not occur and these individuals may not need insulin treatment to survive (American Diabetes Association, 2020).
However, β-cells dysfunction does occur and results in reduced insulin release, which is insufficient for maintaining normal glucose levels (Zheng et al., 2018). The main drivers of the type 2 diabetes globally are the rise in obesity, a sedentary lifestyle, energy-dense diets and population ageing (Zheng et al., 2018). Type 2 diabetes is often associated with a strong genetic predisposition or family history (Vujkovic et al., 2020). The risk factors for developing type 2 diabetes increases with age, obesity
(especially in abdominal region), lack of physical activity, hypertension and dyslipidaemia (American Diabetes Association, 2020). Elevated fasting glucose, abdominal obesity, hypertension, and dyslipidaemia comprise the ‘metabolic syndrome’ and increases an individual’s risk of diabetes, heart disease, stroke, as well as peripheral neuropathy (Cortez et al., 2014, Nilsson et al., 2019).
Treatment of Type 2 Diabetes
Type 2 diabetes management targets not only hyperglycaemia but also the other factors of the metabolic syndrome. Lifestyle modification, including dietary advice and diabetes education with an emphasis on physical activity, are recommended as first-line therapies from diagnosis or in conjunction for patients requiring glucose- lowering medication or metabolic surgery (Davies et al., 2018). Glucose-lowering medication includes metformin, sodium-glucose co-transporter-2 (SGLT2) inhibitors,
6 Literature Review glucagon-like peptide-1 (GLP-1) receptor agonists, dipeptidyl peptidase-4 (DPP-4) inhibitors, thiazolidinediones, sulfonylureas, and insulin (Davies et al., 2018).
Metformin is a biguanide and remains the first-line medication for type 2 diabetes
(Davies et al., 2018). Metformin is an oral medication that primarily inhibits hepatic glucose production and has other mechanisms of action to reduce plasma glucose but may result in lower serum vitamin B12 concentration (Thulé, 2012). SGLT2 inhibitors are oral medications that reduce plasma glucose by enhancing urinary excretion of glucose and their glucose-lowering efficacy is dependent on renal function (Davies et al., 2018). GLP-1 receptor agonists are incretin mimetics and are delivered via subcutaneous injection. They stimulate insulin secretion, reduce glucagon secretion, improve satiety, and promote weight loss (Thulé, 2012). GLP-1 is degraded by DPP-4 and the action of DPP-4 limits the serum half-life of endogenous GLP-1 to under 2 minutes (Holst, 2007). DPP-4 inhibitors are administered orally and like GLP-1 receptors agonists, increase insulin secretion and reduce glucagon secretion but have inferior glucose-lowering efficacy (Davies et al., 2018). Thiazolidinediones are oral medications that increase insulin sensitivity and high-density lipoprotein (Davies et al., 2018). Sulfonylureas are insulin secretagogues that are administered orally and stimulate insulin secretion from pancreatic β-cells (Thulé, 2012). Finally, insulin therapy may be advised in type 2 diabetes in the event of β-cell loss and resultant demise of appropriate insulin secretion. Typically, patients utilise longer or intermediate acting insulin analogues, however rapid-acting insulin formulations are used in patients not meeting glycaemic targets (Davies et al., 2018).
7 Literature Review
Chronic Kidney Disease
Chronic kidney disease (CKD) is a significant global health concern and disease burden, prevalence, and mortality rates are rising (Bikbov et al., 2020, Coresh, 2017).
The estimated global prevalence of CKD is between 10–15% (Bikbov et al., 2020,
Coresh, 2017, Hill et al., 2016). As well as being an important risk factor cardiovascular disease, CKD has a major effect on global health as a direct cause of morbidity and mortality (Bikbov et al., 2020). In 2017, CKD resulted in 35.8 million disability adjusted life-years.
CKD encompasses a continuum of disease from mild kidney damage to end-stage kidney disease, which requires renal replacement therapy in the form of dialysis or renal transplantation. Disease severity is classified using a five-stage system based on the estimated glomerular filtration rate (eGFR), which is an indirect measure of how well the kidneys filter wastes from the blood (Table 1) (Bowling et al., 2011, Chadban and Ierino, 2005, National Kidney Foundation, 2002). This estimation is calculated from age, sex, and creatinine clearance using the Modification of Diet in Renal
Disease formula (National Kidney Foundation, 2002). The majority of CKD patients worldwide are in stage 3 (Hill et al., 2016, Vos et al., 2016). CKD is diagnosed by the presence of kidney damage, manifested by abnormal albumin excretion or decreased eGFR, that persists for more than three months (Levey et al., 2005). The aetiology of
CKD may be due to a primary renal disorder or as a complication of a multisystem disorder. Diabetes is the leading cause of CKD as well as end-stage kidney disease and accounts for more than half of the global prevalence (Bikbov et al., 2020,
National Kidney Foundation, 2012, Stanton, 2014, Vos et al., 2016). Complications of
8 Literature Review
CKD such as hypertension, anaemia, and metabolic bone disease become apparent with declining eGFR (Holley, 2011, Inker et al., 2011, Thomas et al., 2008).
Table 1. Classification of chronic kidney disease Stage 1 Evidence of kidney damage with normal eGFR >90 mL/min/1.732 Stage 2 Evidence of kidney damage with mild reduction of eGFR 60–89 mL/min/1.732 Stage 3 Moderately reduced eGFR 30–59 mL/min/1.732 Stage 4 Severely reduced eGFR 15–29 mL/min/1.732 Stage 5 Renal failure or dialysis eGFR < 15 mL/min/1.732
Classification as defined by the National Kidney Disease Outcomes and Quality Initiative clinical practice guidelines (National Kidney Foundation, 2002).
Treatment of Chronic Kidney Disease
Only a small proportion of patients with mild or moderate CKD will progress to end- stage kidney disease (Hallan et al., 2006). Management of CKD primarily involves prevention of complications and identifying patients at risk of disease progression
(Fraser and Blakeman, 2016). At end-stage kidney disease, renal replacement therapy in the form of haemodialysis, peritoneal dialysis, or renal transplantation is required to sustain life.
Diabetic Kidney Disease
CKD caused by diabetes is known as diabetic kidney disease (DKD) or diabetic nephropathy. It is estimated that 30% of patients with type 1 diabetes or 40% of patients with type 2 diabetes will develop DKD during their lifetimes (Alicic et al.,
9 Literature Review
2017). For the majority of these patients, DKD will develop within 10 years of diagnosis (Couser et al., 2011). Cardiac autonomic neuropathy and certain plasma biomarkers have been identified as risk factors associated with rapid decline in eGFR in type 2 diabetes (Peters et al., 2017, Tahrani et al., 2014). Genetic risk factors for the development of DKD and rapid decline of eGFR have emerged but these require further validation (Davoudi and Sobrin, 2015, Jiang et al., 2019). Other clinical risk factors include increased albuminuria, hyperglycaemia, hypertension, dyslipidaemia, obesity, and smoking (Hussain et al., 2020). In the management of diabetic kidney disease, SGLT2 inhibitors and GLP-1 receptor antagonists have been shown to slow the progression of kidney disease (Ninčević et al., 2019). Oxidative stress in the kidney from hyperglycaemia and dyslipidaemia is postulated to be the key component of in the development diabetic nephropathy (Eid et al., 2019, Forbes et al., 2008,
Savelieff et al., 2020). Insulin resistance has also been identified as a major determinant of DKD via a number of pathogenic pathways (Karalliedde and Gnudi,
2016).
10 Literature Review
The Human Nervous System
The human nervous system is comprised of billions of neurons responsible for the transmission and processing of electrical signals around the body. These signals may arise internally, to coordinate actions around the body, or externally, to perceive the surrounding environment. Although structurally and functionally linked, the human nervous system can be divided into two anatomically distinct components, the central nervous system (CNS) and the peripheral nervous system (PNS). The CNS is comprised of the brain and spinal cord and functions to integrate information and coordinate activity. The PNS serves to connect the CNS to the body and encompasses the remaining elements of the nervous system, which includes the cranial nerves (III–
XII), spinal nerves, peripheral nerves, sensory receptors, and neuromuscular junctions
(Catala and Kubis, 2013).
At the cellular level, the nervous system is composed of neurons and neuroglia, which differ in structure and function. Neurons are electrically excitable cells which are capable of receiving and sending information. Neurons serve as the primary signalling unit of the nervous system and enable the communication between the CNS and the rest of the body via the PNS. Different types of neurons exist throughout the nervous system and they are typically classified into three types depending on their function.
Sensory neurons respond to stimuli in the periphery and send signals to the spinal cord and brain, where the information is perceived by other sensory neurons. Motor neurons in the brain send commands to other motor neurons in the spinal cord where they exert control over muscles and glands. Interneurons connect neurons in different regions of the brain or spinal cord. Structurally, the typical neuron consists of a cell body (soma) which gives rise to several short processes known as dendrites and a
11 Literature Review relatively longer, single tubular process termed the axon (Figure 1). In the PNS, the soma of motor neurons are located in the ventral horn of the spinal cord, under the protection of the blood-brain barrier, whereas the soma of sensory neurons are situated in the dorsal root ganglion and afforded no protection from the blood-brain barrier. In the context of peripheral nerve injury, this is especially important because it renders sensory neurons especially susceptible to damage. The soma houses the nucleus. The rough endoplasmic reticulum, Golgi bodies, and other organelles necessary for the maintenance of cellular function are present in the soma, dendrites, and axon (Berthold et al., 2005, Catala and Kubis, 2013). The dendrites are the principle structures that receive input from other neurons or sensory receptors and conduct this signal towards the soma. The axon is a highly specialised region that serves as a conduit for impulse conduction away from the soma and towards the dendrites of other neurons, muscles or glands. Depending on the type of neuron, axons may range in length from less than a millimetre to over a metre long and account for at least 95% of the cell mass in long neurons (Bear et al., 2016, Miller,
2014). The neuronal cell plasma membrane, which is approximately 5 nm thick, contains proteins essential for neuronal function and protein composition of the membrane varies between the soma, dendrites, and axon (Bear et al., 2016). In contrast, neuroglia do not directly participate in electrical signalling and information processing but they provide crucial structural and metabolic support for neurons. For example, Schwann cells in the PNS form the myelin seen on axons greater than 1–2
µm in diameter and provides an insulating sheath that permits saltatory conduction and the rapid and reliable propagation of impulses (Catala and Kubis, 2013, Fitzhugh,
1962, Huxley and Stämpeli, 1949, Kuriscák et al., 2002). Schwann cells are also involved in the metabolic support of axons and protecting them against toxic
12 Literature Review molecules and reactive oxygen species (Beirowski et al., 2014, Vincent et al., 2011,
Wang et al., 2012).
13 Literature Review
Paranode Juxtaparanode Internode
Node
Soma Axon hillock
Axon Axoglial Periaxonal Dendrite junction space Axon terminals Initial segment Myelin Node of Action sheath Ranvier potential
Figure 1. Schematic of a motor neuron and domains of a myelinated axon. The axon is
ensheathed in myelin segments which allows action potential propagation from node to node,
known as saltatory conduction.
14 Literature Review
The Peripheral Axon
The primary role of the PNS is to facilitate the transmission and receipt of electrical signals between CNS and the body. This communication is facilitated by electrical impulses, known as action potentials, which travel along the axon to permit the communication between the neurons and muscles, glands, or other neurons. Another important function of the axon is that it provides a conduit for synthesised proteins that have important structural and functional roles to be transported (Wang and He,
2013). The rapid and reliable transmission of action potentials is heavily dependent on the strategic distribution of myelin and proteins such as voltage-gated ions channels, ion pumps, and ion exchangers along specific domains of the axon (Barnett and
Larkman, 2007, Kursula, 2014). These structural and molecular factors are important determinants in the generation and conductance of action potentials, a property known as axonal excitability. Assessment of axonal excitability has provided important insight into nerve function in healthy and diseased states (Kiernan et al., 2020).
Structure of Peripheral Axons
The axon becomes physiologically distinct from the soma at the axon hillock
(Figure 1) (Leterrier, 2018). The axon hillock marks the beginning of the axon and delineates the boundary in the cytoplasm where the rough endoplasmic reticulum and
Golgi apparatus of the soma is not continuous with the axon (Wang and He, 2013).
The axon hillock is followed by the initial segment and is the region where an impulse is first generated and travels toward the axon terminals (Leterrier, 2018, McCormick,
2013). While ribosomes are found in the axon, the absence of the other cellular organelles important for protein synthesis renders the axon greatly dependent on the soma for trophic support, which is important in the context of sensory nerves in which
15 Literature Review the soma is outside the protection of the blood-brain barrier (Miller, 2014). This dependence leaves distal regions of the longest axons, such as the axons in the sciatic nerve, susceptible to toxic and metabolic injury (Berthold et al., 2005).
Myelinated Fibres
In the PNS, axons greater than 1–2 µm in diameter are insulated with consecutive myelin segments over the plasma membrane of the axon (Catala and Kubis, 2013).
The axonal membrane, known as the axolemma, is therefore encased by a myelin sheath, which isolates the fibre from neighbouring axons. These successive myelin segments are Schwann cells wrapped in a multilamellar, spiral fashion and spaced approximately 1 µm apart from adjacent Schwann cells, leaving regions of the axolemma without myelin (Freeman et al., 2016). These myelin-free, segments of axolemma are known as the nodes of Ranvier and are exposed to extracellular environment (Berthold et al., 2005). Despite these regularly spaced intervals in the myelin sheath, approximately 99.9% of the axonal membrane is covered in myelin and is termed the internode (Salzer et al., 2008). Myelin greatly increases the resistance between cytoplasm of the axon and extracellular space and also reduces current loss along the axolemma of the internode (Ramahi and Ruff, 2014). This effect of myelin enables almost all the current produced by one node of Ranvier to excite the adjacent node and therefore increase conduction velocity by what is termed saltatory conduction (Ramahi and Ruff, 2014).
Domains of Myelin
Myelin segments range from 300-2000 µm in length and can be subdivided into distinct domains, each with a specific protein composition and function (Figure 1)
16 Literature Review
(Berthold et al., 2005, McCormick, 2013). The node of Ranvier is of particular importance in myelinated fibres. In addition to being exposed to the extracellular environment, this region is distinguished from other domains as it contains an immense density (approximately 1000/µm2) of voltage-gated sodium (Na+) channels
(Ritchie and Rogart, 1977). This distribution of Na+ channels is similar to the axon initial segment and consequently, the node of Ranvier is the site of generation of the action potential in saltatory conduction (Rasband, 2010a). The nodal region is characterised by increased mitochondrial density due to the metabolically active nature of this site (Zhang et al., 2010). Immediately adjacent to the node, is the paranodal region which is 3–5 µm in length (Berthold et al., 2005, Inouye et al.,
2014). At the paranode, microvillar processes from the outermost regions of Schwann cells protrude into the axolemma (Ichimura and Ellisman, 1991). These microvilli contain various matrix proteins that support node formation, maintenance, and function (Corfas et al., 2004, Ogawa et al., 2006). The axolemma at the paranode contains a complex of cell adhesion molecules that includes caspr and contactin, which permits the formation of an axo-glial junction between Schwann cells and axonal membrane (Einheber et al., 1997, Hivert et al., 2016, Menegoz et al., 1997,
Peles et al., 1997, Rios et al., 2000). Continuing underneath the myelin segment, the paranode is followed by the juxtaparanode which extends 10 µm in length (Arroyo et al., 1999, Inouye et al., 2014). The juxtaparanode is characterised by a high density of fast potassium (K+) channels, which may function to dampen nodal excitability after action potential generation to prevent re-excitation or maintain internodal excitability
(Chiu and Ritchie, 1984, Rasband et al., 1998, Wang et al., 1993, Waxman and
Ritchie, 1993). Absence of a normal axo-glial junction results in an accumulation of
K+ channels in the paranode from the adjacent juxtaparanode and an impairment in
17 Literature Review saltatory conduction (Bhat et al., 2001). Thus an important function of the axo-glial junction is to form a barrier between the Na+ channels in the node and K+ channels located in juxtaparanode (Corfas et al., 2004). Finally, adjacent to the juxtaparanode is the internodal region, which constitutes 95% of the length of myelin segments and spans 150–1200 µm (Lascelles and Thomas, 1966, Salzer et al., 2008). Between the internodal axolemma and the inner membrane of the Schwann cell, there is a uniform separation termed the periaxonal space, which spans 15 nm (Salzer et al., 2008). The internodal region is much less densely populated compared to the other domains, however it contains a variety of ion channels, pumps and exchangers involved in the maintenance of membrane potential, and in particular, internodal excitability
(Krishnan et al., 2009a).
Unmyelinated Fibres
In the PNS, there are more thin unmyelinated axons (<1 µm in diameter) than myelinated axons (Feldman et al., 2017, Malik et al., 2011). These axons are known as C fibres and function to carry information for the autonomic nervous system as well as afferent impulses in response to pain (Catala and Kubis, 2013). C fibres are enveloped and grouped together by non-myelinating Schwann cells to form Remak bundles (Feldman et al., 2017). A consequence of the lack of myelin in C fibres is slow impulse conduction (Miller, 2014). Unlike myelinated fibres, the Na+ and K+ channels taking part in action potential generation are distributed uniformly along the axon and instead of saltatory conduction, impulse propagation occurs through local excitation of the neighbouring patch of membrane to generate an action potential
(McCormick, 2013). Injury to these fibres may occur in the early phase of diseases of
18 Literature Review the nervous system and may cause symptoms of burning pain (Lauria et al., 2014,
Malik, 2014). C fibres will be discussed in more detail in the further sections.
Function of Peripheral Axons
The function of peripheral axons is to provide a means of communication for neurons through an action potential. Specifically, the action potential is series of discreet changes in the voltage of the axon produced by the flow of Na+ and K+ across the axolemma (Hodgkin and Huxley, 1952d). Pioneering studies by Hodgkin, Huxley, and Katz on giant squid axons provided the physiological basis of the mathematical model that describes an action potential generation, which was contingent on the permeability of axolemma to Na+ and K+ ions (Hodgkin and Huxley, 1952a, 1952b,
1952c, Hodgkin and Katz, 1949).
Membrane Potential
The sequence of events that underlie the action potential require an actively maintained chemical and electrical concentration gradient across the inner and external surface of the axolemma and is known as the membrane potential. This gradient is fundamental for the action potential and is enabled by the phospholipid bilayer of the membrane that serves as barrier between the intracellular and extracellular environment and is maintained by energy-dependent ion pumps, which are affected in metabolic neuropathies (Cannon, 2014, McCormick, 2013). By convention, the membrane potential is relative to the inside of the axon and can be represented by the equation:
� = ������� ������ �ℎ� ���� (� ) − ������� ������� �ℎ� ���� (� )
19 Literature Review
At rest, axons possess a relative negative charge on the inner surface of the axolemma due the unequal distribution of Na+, K+ and chloride (Cl–) ions across the membrane
(McCormick, 2013). In human axons, the resting membrane potential is –75 to –80 mV and determined by the equilibrium potential for K+ and to a lesser extent, by the equilibrium potential for Na+ and the Na+/K+ pump (Uncini and Kuwabara, 2015). Cl– appears to contribute considerably less to the determination of resting potential of mammalian neurons. If the membrane potential becomes more positive than the resting value, this is termed depolarisation and if the membrane potential becomes more negative than the resting value, this is referred to as hyperpolarisation. While the axolemma is largely impermeable to ions, it contains specialised transmembrane proteins known as ion channels (Cannon, 2014). These ion channels open to form selective permeation pathways that facilitate the passage of specific ions across the phospholipid bilayer through an aqueous pore (Cannon, 2014). Opening and closing of ion channels is said to be voltage-gated as their permeability is regulated by membrane voltage. It is the rapid redistribution of charge through these ions channels that forms the molecular basis of the action potential.
The Action Potential
An action potential is a transient reversal (<1 ms) of the resting membrane potential such that the inside of axolemma becomes positively charged with respect to the outside momentarily (Barnett and Larkman, 2007). The action potential consists of three phases and each phase represents a characteristic change in membrane potential which is driven by the flow of a specific ion across the membrane (Figure 2).
20 Literature Review
+40
Upstroke Downstroke
Threshold Membrane Membrane voltage (mV) depolarisation –75 hyperpolarisation
Resting potential After-hyperpolarisation Time Figure 2. The action potential
Initiation of an action potential may be caused by a chemical or physical stimulus producing a ‘local depolarising response’ or ‘generator potential’ (Barnett and
Larkman, 2007). In motor neurons, which are focus of this thesis, this local depolarisation leads to the opening of a relatively small number of voltage-gated Na+ channels in the axon hillock or node of Ranvier leading to an inward current of Na+ into the axon from the extracellular space (Barnett and Larkman, 2007, Hodgkin and
Katz, 1949). Should depolarisation reach a sufficient level, known as ‘threshold’, an action potential will be triggered. For most neurons, threshold is around –45 to –55 mV. At this membrane potential, the complement of open Na+ channels is sufficient for the inward Na+ current to exceed the combined outward current of K+ and Cl– ions
21 Literature Review that opposes membrane depolarisation (McCormick, 2013). At threshold, there is rapid opening of many more voltage-gated Na+ channels, resulting in an even greater influx of Na+ (McCormick, 2013). This increase of Na+ influx results in further depolarisation, which in turn triggers the opening of more Na+ ion channels in a positive feedback fashion (McCormick, 2013). In this sense, the process of voltage- gated Na+ channel recruitment is termed ‘regenerative’; once threshold is reached, an action potential will always occur, therefore being an ‘all-or-none’ response. Once an action potential has been triggered, the depolarisation can reverse the membrane potential to as high as +40 mV, but not quite reaching the equilibrium potential of Na+
(Barnett and Larkman, 2007). The equilibrium potential of Na+ is the membrane potential at which there is no net movement of Na+ across the axolemma and is approximately +66 mV (Cannon, 2014). Depolarisation of the membrane is limited due to the fast inactivation kinetics of the transient Na+ ion channels involved the action potential (McCormick, 2013).
The second phase of the action potential involves restoration of the membrane potential towards the resting value by what is termed repolarisation. While voltage- gated K+ channels open in this phase allowing the efflux of K+ ions, repolarisation is primarily the result of the inactivation of transient Na+ channels and current leak into the internode (Ritchie, 1995). Voltage-gated K+ channels were found to be responsible for repolarisation of the axolemma in giant squid axons, however they have a negligible contribution in human axons (Schwarz et al., 1995).
The final phase of the action potential is where the effect of voltage-gated K+ channels is primarily seen in human axons. The slow deactivation kinetics of some
22 Literature Review voltage-gated K+ channels enables the prolonged efflux of K+ ions, causing an after- hyperpolarisation beyond the resting membrane potential (McCormick, 2013).
Resting membrane potential is eventually returned after the closure of these ion channels and activity of energy-dependent pumps.
Saltatory Conduction
Saltatory conduction (from the Latin saltare, “to leap”) refers to the propagation of action potentials along myelinated axons from one node of Ranvier to the next
(Huxley and Stämpeli, 1949). Classically, this is thought to be due to the insulator effect of myelin, which enables almost all the current produced by one node of
Ranvier to travel through the internode (without leak) and excite the adjacent node with an internode delay of only about 20 μs (McCormick, 2013, Ramahi and Ruff,
2014). Importantly, the magnitude of current generated at the node is >5 times greater than what is required to generate an action potential at the neighbouring node and is referred to as the safety factor of transmission (Huxley and Stämpeli, 1949, Stämpfli,
1954, Tasaki, 1953). The safety factor ensures the propagating impulse will not diminish and can be calculated as a ratio of current leaving the node of Ranvier to current required for excitation at the next node (Tasaki, 1953). This ratio must be greater than one for current conduction through a node to be successful and it can be adversely affected in demyelinating pathologies (Bowley and Chad, 2019, Kiernan and Kaji, 2013).
Ion Channels, Pumps, and Exchangers
The molecular basis of the action potential is formed by the presence of ion channels that form selective permeation pathways across the phospholipid bilayer of the axonal
23 Literature Review membrane. While the first electrophysiological recordings from individual ion channels were not made until the 1970s, Hodgkin and Huxley predicted the key properties now known to be essential for the action potential: ion selectivity, voltage sensitivity, and importantly, channel closing, which ensures the action potential moves along the axon in one direction (Barnett and Larkman, 2007, Neher and
Sakmann, 1976). Ion channels are transmembrane proteins that have a three- dimensional structure and an aqueous pore (McCormick, 2013). The net flow of ions through these channels is the source of the electrical current that rapidly changes the membrane potential during an action potential (Cannon, 2014). These rapid transients are possible because of the high density of ion channels, the high throughput of an open channel (107ions/s), and the ability of channels to change conformation from open or closed within 1 ms or less (Cannon, 2014). Many of the ion channels involved in the action potential are voltage-gated, as changes in their conformation to open or close are in response to alterations in membrane potential (McCormick, 2013). The movement of ions in or out of the axon is determined by their concentration gradient across the axolemma, which is established by energy-dependent pumps and exchangers (Barnett and Larkman, 2007). The following section outlines the variety of ion channels, pumps and exchangers that present in specific areas of the axon and are essential to impulse conduction and other electrical properties of the axon (Figure
3).
24 Literature Review
Myelin sheath
+ + + Na Na Na Na+/K+ Ca2+ Nap Nat Nat Kf Ih 3Na+ K+ K+ K+ 3Na+ 2K+
2+ Ks Ca Ks 3Na+ Na+/Ca2+ Na+/K+–ATPase exchanger
Node Paranode Juxtaparanode Internode
Figure 3. Distribution of ion channels, pumps, and exchangers in a myelinated axon. Nat: + + transient voltage-gated Na channel; Nap: persistent voltage-gated Na channel; Kf: fast + + voltage-gated K channel; Ks: slow voltage-gated K channel; Ih: inward rectifier
Sodium (Na+) Channels
Action potentials in mammalian axons can be modelled entirely by voltage-gated Na+
(NaV) channel kinetics, therefore making these channels the most important contributor to neurotransmission (Schwarz et al., 1995). Structurally, NaV channels comprise of a single pore-forming a-subunit with one or two auxiliary b-subunits
(Krishnan et al., 2009a, Noda et al., 1984).
The a-subunit is a large, single polypeptide glycoprotein, with a molecular mass of approximately 270 kDa (McCormick, 2013). It consists of four (I-IV) linked homologous transmembrane protein domains, with each domain comprising of six a- helical segments that span the neuronal membrane (S1-6) (Eijkelkamp et al., 2012).
The a-subunit contains specialised regions, which are necessary for activation,
25 Literature Review inactivation and ion permeation. The S4 segment of each domain acts as the membrane ‘voltage sensor’ and has an amino acid sequence that is conserved among the different voltage-gated ion channels (Catterall, 2012, Stühmer et al., 1989).
Importantly, this segment can respond to alterations in membrane potential to mediate the voltage-dependant conformational that occur during channel activation and inactivation (Krishnan et al., 2009a). An additional pore loop between S5 and S6 contributes to the formation of the pore and acts as a selectivity filter to only permit the passage of Na+ from the extra-cellular environment through the aqueous pore
(Krishnan et al., 2009a).
The b1 and b2-subunits are members of the immunoglobulin domain family of cell- adhesion molecules with molecular masses of 39 and 37 kDa, respectively
(McCormick, 2013, Namadurai et al., 2015). They contain a large extracellular domain, a transmembrane region and a small intracellular domain. To date, four b- subunits (NaVb1-4) have been identified (Namadurai et al., 2015). The b subunits interact with the extracellular matrix, intracellular cytoskeleton, and cell adhesion molecules (Catterall et al., 2005). They are involved in channel localisation, modulating channel gating, as well as increasing the amplitude of the inward Na+ current (Catterall et al., 2005, Gurnett and Campbell, 1996, Isom, 2001).
Correct action potential generation and propagation is dependent on the three distinct operational states of NaV channels: closed, open and inactivated. At the resting membrane potential, NaV channels are closed. Membrane depolarisation causes a conformational change of the S4 helix, which in turn opens the channel to permit the influx of Na+ (McCormick, 2013). Positive charges in the S4 region may act as
26 Literature Review voltage sensor such that an increase in the positivity inside the cell results in the conformational change (McCormick, 2013). Sustained depolarisation causes the channel to become inactivated. The channel becomes inactivated within milliseconds of opening and is caused by folding of the intracellular loop that binds domains III-
IV, thereby blocking the channel (Taddese and Bean, 2002). The mechanism of inactivation is hypothesised to be due to a block of the aqueous pore triggered or facilitated as a secondary consequence of activation (McCormick, 2013).
To date, ten Nav a-subunit isoforms have been cloned and functionally characterised
(NaV1.1–1.9 and Nax) (Catterall, 2012, Hiyama et al., 2002). These a-subunit isoforms are expressed in a tissue-specific fashion and have different electrophysiological characteristics, kinetic properties, and toxin sensitivities (Goldin,
2001). There are three isoforms that are primarily present in the peripheral nervous system, NaV1.7, NaV1.8 and NaV1.9 (Bennett, 2014, Catterall, 2012, McDermott et al., 2019). NaV1.1, NaV1.2, and NaV1.6 have been identified in both the peripheral and central nervous systems (Eijkelkamp et al., 2012). NaV1.6 is of particular importance because it is the predominant isoform present at the node of Ranvier in humans and produces the persistent and transient Na+ currents (Caldwell et al., 2000, Herzog et al., 2003).
During the phase of action potential that corresponds with membrane depolarisation,
+ 98% of the Na current is due to the opening of ‘transient’ NaV (Nat) channels, which exhibit the stereotypical rapid gating kinetics (Burke et al., 2001). The remaining 2%
+ of the total Na current is attributable to ‘persistent’ NaV (Nap) channels, which activate at membrane potentials 10-15 mV more negative than transient NaV channels
27 Literature Review and undergoes incomplete voltage-dependent inactivation (Crill, 1996, Kiss, 2008,
Krishnan et al., 2009a). Thus, Nap channels may be activated at resting membrane potential and permit a Na+ conductance over a wider range of membrane potentials compared to transient NaV channels (Baker and Bostock, 1998). Although Nap channels only contributing a small fraction of the total Na+ current, they have an important role in regulating membrane excitability by amplifying subthreshold depolarising inputs, which contributes to repetitive neuronal firing (Eijkelkamp et al.,
2012, Taddese and Bean, 2002). The activation of Nap channels at resting membrane potential suggests they contribute to ectopic symptom generation such as paraesthesia, pain and cramping (Bennett et al., 2019, Bostock and Rothwell, 1997, Kwai et al.,
2013, Misawa et al., 2009, Mogyoros et al., 1997, Tamura et al., 2006). However, it remains is unclear whether the Nat and Nap conductances arise from different NaV channels or if they arise from the same channel type but with different gating kinetics
(Krishnan et al., 2009a).
Disease States
Mutations in the genes encoding for the various a or b-subunits may result in altered gating kinetics, reduced channel density and improper Na+ currents leading to various neurological disorders. There is a clear association between altered NaV channel function and familial forms of epilepsy and pain disorders (Bennett, 2014, Bennett et al., 2019, Bennett and Woods, 2014, Blesneac et al., 2018, Catterall, 2012, Colloca et al., 2017, Eijkelkamp et al., 2012, Themistocleous et al., 2018). Mutations in these genes have also been implicated in inherited forms of autism, ataxia, and migraine
(Eijkelkamp et al., 2012)
28 Literature Review
Potassium (K+) Channels
In both structure and function, K+ channels represent the most diverse family of ion channels (Humphries and Dart, 2015, Kuang et al., 2015). This section will only focus on K+ channels relevant to the mammalian peripheral axon. Mammalian axons express several types of potassium channels including voltage-gated K+ channels
+ + + (KV), inwardly rectifying K channels (KIR), leak K channels, and K channels sensitive to Ca2+ or Na+ (Coetzee et al., 1999, Krishnan et al., 2009a). K+ channels consist of four a-subunits, which may be associated with auxiliary b-subunits
(Krishnan et al., 2009a). The typical structure of K+ channels is a tetramer of four a- subunits arranged around a central pore, with each subunit contributing one transmembrane domain to form the passage (Kuang et al., 2015). The transmembrane pore has an ion selectivity filter that is highly selective and at least 10,000 times more permeant for K+ than Na+ ions (Kuang et al., 2015). The a-subunit consists of a variable number of transmembrane domains depending on the K+ channel type, which is the initial division of the phylogenetic tree for K+ channels (Humphries and Dart,
+ 2+ 2015). KV channels have six transmembrane domains, K channels sensitive to Ca
+ have seven transmembrane domains, while KIR channels and leak K channels have two and four transmembrane domains, respectively (Humphries and Dart, 2015). In neurons, due to equilibrium potential of K+ (approximately –85 mV), the opening of
K+ channels generally mediates outwards currents which serve to dampen cellular excitability (Humphries and Dart, 2015).
+ Voltage-gated K (KV) Channels
There are 40 known human KV channels, which can be separated into 12 subfamilies on the basis of their a-subunit (KVa1–12) (Gutman et al., 2005). Like Nav channels,
29 Literature Review
the positively charged S4 segment make KV channels electrically sensitive and also mediates the eventual opening of the conduction pathway (Kuang et al., 2015, Long et al., 2007). In addition to their a-subunit, KV channels can be classified further according to their encoding gene, channel kinetics, location and toxin sensitivity
(Krishnan et al., 2009a). While the mammalian myelinated axon has a variety of KV channels, the following section will only focus on the two types of KV most relevant to the in vivo assessment of axonal ion channel function, those with fast or slow kinetics.
Fast KV channels (KV1) were first shown to be concentrated in the juxtaparanode, with lower densities present in the internode (Röper and Schwarz, 1989, Wang et al.,
1993). Further evidence suggests they are also enriched at the axon initial segment
(Rasband, 2010b). In the juxtaparanode of myelinated axons, the Kv1 subunits that have been specifically identified are KV1.1, KV1.2, and KV1.4 (Rasband, 2004,
Tsantoulas and McMahon, 2014). Anchoring molecules on the axolemma and
Schwann cell membrane are crucial for the clustering of juxtaparanodal KV channels
(Hivert et al., 2016, Rash et al., 2016, Uncini and Kuwabara, 2015). Fast KV channels are activated between membrane potentials –50 mV and +50 mV and deactivate within a few milliseconds at membrane potentials of –120 mV and – 60 mV (Reid et al., 1999, Safronov et al., 1993, Vogel and Schwarz, 1995). In non-myelinated axons, blockade of fast KV channels leads to action potential prolongation, which suggests these channels may play a role in membrane repolarisation (Bostock et al., 1981,
Sherratt et al., 1980). However, as mentioned above, in mature myelinated axons repolarisation occurs due to the inactivation of transient Na+ channels and current leak into the internode (Ritchie, 1995). In functional terms, fast KV are thought to be
30 Literature Review responsible for limiting the re-excitation of the node following conduction of an action potential and maintaining a stable nodal resting potential (Barrett and Barrett,
1982, Chiu and Ritchie, 1984, Rasband, 2004, Waxman and Ritchie, 1993).
Slow KV channels are formed by the combination of KV7.2 and KV7.3 subunits
(Gutman et al., 2005). Slow KV are present in highest density at the node and axon initial segment and they have slow activation and deactivation kinetics (Devaux et al.,
2004, Röper and Schwarz, 1989, Schwarz et al., 2006). Slow KV channels do not inactivate and generate a steady outward current that stabilises the membrane in response to depolarising currents (Brown and Passmore, 2009). Slow KV channels are active at resting membrane potential and play an important role in its maintenance
(Reid et al., 1999, Schwarz et al., 2006). Due to their slow activation, slow KV channels are not thought to be responsible for membrane repolarisation following action potential conduction, however they do produce a small hyperpolarising afterpotential to reduce neuronal excitability (Baker et al., 1987, Eng et al., 1988).
Slow KV channels are also activated in response to prolonged depolarisation during high frequency activity to prevent inappropriate after-discharge (Schwarz et al.,
2006).
Disease States
In humans, loss of function mutations of KV1.1 are associated with episodic ataxia type 1, an autosomal dominant condition characterised by myokymia and severe contractions of the head and limbs resulting in loss of coordination and balance
(Browne et al., 1994). An accumulating body of research suggests that some human neuropathic pain syndromes are caused by autoimmune antibodies against KV1
31 Literature Review subunits (Bennett and Vincent, 2012, Tsantoulas and McMahon, 2014). Peripheral nerve demyelination results in the redistribution of the fast KV channels into the nodal, paranodal, and internodal regions of the axon due to the loss of structural molecules that usually anchor these channels in the juxtaparanode and results in impaired impulse conduction (Arroyo et al., 1999, Boyle et al., 2001, Rasband et al.,
1998). Mutations in KV7.2 and KV7.3 lead to neonatal epilepsy and myokymia, providing further evidence these channels play an important role in maintaining resting membrane potential and limiting repetitive firing (Dedek et al., 2001,
Humphries and Dart, 2015). Loss of function mutations in KV7.3 have also been associated with autism spectrum disorders (Gilling et al., 2013).
Inwardly Rectifying K+ Channels
+ In terms of structure, KIR channels are the simplest K channels with each subunit being formed by two transmembrane domains separated by a pore-forming region
(Humphries and Dart, 2015). These subunits form homo- or hetero-tetramers to produce functional KIR channels (Bichet et al., 2003, Hibino et al., 2010). KIR channels can be divided into seven subfamilies based on their modulatory mediators and properties of ion conduction (Kuang et al., 2015). The unique feature of KIR channels is that they possess a characteristic asymmetrical K+ conductance whereby
K+ moves into the cell on hyperpolarisation rather than an outward K+ conductance on depolarisation as seen in other K+ channels (Humphries and Dart, 2015, Kuang et al.,
2015). KIR channels also possess a cytosolic domain that regulates the gating of the channel (Kuang et al., 2015). Inward rectification occurs because these channels are blocked by intracellular substances on depolarisation, whereas these blockers are
+ released on hyperpolarisation to permit K influx (Hibino et al., 2010). KIR channels
32 Literature Review tend to be active around the equilibrium potential for K+, and thus help set and maintain the resting membrane potential, but close in response to depolarisation so as not to oppose membrane excitation (Kuang et al., 2015). There are three relevant types of inward rectifiers currents, a K+-selective inward rectification dependent on extracellular K+ concentration, a current mediated by adenosine triphosphate (ATP),
+ + and a mixed cation conductance (Ih) of Na and K through hyperpolarisation- activated cyclic nucleotide-gated (HCN) channels (Humphries and Dart, 2015, Mayer and Westbrook, 1983, Pape, 1996).
The K+-selective channel opens with hyperpolarisation and also exhibits voltage- dependent gating that is influenced by the extracellular K+ concentration, with activation at more depolarised potentials as the extracellular K+ concentration increases (Krishnan et al., 2009a). The conductance depends on the difference between the membrane potential and the K+ equilibrium potential, rather than on membrane potential alone, with an increase in extracellular K+ leading to an influx of
K+ (Birch et al., 1991).
The ATP-sensitive inward rectifier provides a link between cell metabolism and electrical activity (Krishnan et al., 2009a). In the PNS, they are primarily localised in the dorsal root ganglion, where they are thought to play a minor role in setting the basal excitability (Chi et al., 2007, Du et al., 2011, Kawano et al., 2009). These channels are voltage-insensitive however they are inhibited by intracellular ATP (Du and Gamper, 2013, Jonas et al., 1991). Given their location, opening of these channels reduces excitability and pain induced by a range of painful stimuli (Du et al., 2011,
Kawano et al., 2009, Tsantoulas and McMahon, 2014). In situations of low energy
33 Literature Review supply, these channels are open and cause membrane hyperpolarisation, which may protect the axon from the progressive depolarising effects of reduced Na+/K+–ATPase function (Krishnan et al., 2009a, Sun and Feng, 2013). While these channels are mainly found in sensory neurons, agonists of ATP-sensitive channels have been used in animal models of diabetic neuropathy to counter the reduced function of Na+/K+–
ATPase and have demonstrated improvements in motor nerve conduction velocity
(Greene, 1986b, Hohman et al., 2000).
The Ih current flowing through HCN channels plays an important role in the determination of resting membrane potential (Biel et al., 2009, Pape, 1996). HCN channels are voltage-gated channels that belong to a superfamily known as ‘pore-loop cation’ channels (Biel et al., 2009). In mammals, four subunits (HCN1–4) have been identified and their encoding genes are members of the voltage-gated K+ superfamily
(Krarup and Moldovan, 2012). In neurons, mainly HCN1 and HCN2 are expressed
(Chaplan et al., 2003). In vivo, these subunits form homotetramers arranged around a central pore (Biel et al., 2009). Each subunit consists of six transmembrane a-helical segments (S1-S6), with a voltage sensor at S4 and a selectivity filter between S5 and
S6 (Benarroch, 2013, Biel et al., 2009). In the peripheral nervous system, they are mainly localised in the dorsal root ganglion (Biel et al., 2009). While HCN channels are selective for Na+ and K+, the conductance for K+ is greater and an increase in extracellular K+ concentration strongly increases current amplitude (Benarroch, 2013,
Krarup and Moldovan, 2012). HCN channels are activated by hyperpolarisation and directly by cytosolic cAMP (Biel et al., 2009, Krarup and Moldovan, 2012). This conductance, which is thought to originate from the internode, exhibits a slow and complex time course of activation, with an increase in the rate of activation with
34 Literature Review
increasing hyperpolarisation (Baker et al., 1987, Biel et al., 2009). Ih currents begin to activate at –45 mV to –60 mV and activation generally reaches a maximum at –110 mV and does not inactivate (Biel et al., 2009, Pape, 1996). Blockade of this conductance leads to an increase in the input resistance of the cell, suggesting that Ih plays a role in lowering input resistance particularly in situations of membrane hyperpolarisation that would reduce nerve excitability, such as heightened Na+/K+–
ATPase activity during prolonged nerve stimulation (Applegate and Burke, 1989,
Benarroch, 2013). Furthermore, variations in the modulation and expression of HCN channels in sensory axons compared to motor axons has provided some explanation for the functional differences between these neuron types (Howells et al., 2012).
Disease States
Dysfunctional ATP-sensitive inward rectifiers have been associated with generalised seizures after brief hypoxia (Yamada and Inagaki, 2005). Abnormal regulation of
HCN expression or function has been implicated in epilepsy as well as inflammatory and neuropathic pain (Benarroch, 2013, Dibbens et al., 2010, Jiang et al., 2008).
Sodium-Potassium Pump (Na+/K+–ATPase)
+ + The Na /K pump consists of two subunits, a and b arranged in a tetramer (ab)2
(McCormick, 2013). The Na+/K+ pump is located primarily located in the internodal membrane however clusters of this protein has been observed in the nodal and paranodal regions of the axon (Ariyasu et al., 1985, Bostock et al., 1991, Mata et al.,
1991, Wood et al., 1977, Young et al., 2008). The Na+/K+ pump participates in the maintenance of the resting membrane potential by transporting Na+ and K+ against their concentration gradients, with 10% of this maintenance being ascribed to pump
35 Literature Review activity (Thomas, 1972). The Na+/K+ pump is strongly stimulated by increases in intracellular concentration of Na+ and extrudes 3 Na+ ions out of the cell while transporting 2 K+ ions into the cell, achieving this task through the hydrolysis of ATP
(McCormick, 2013, Thomas, 1972). Pump activity also increases with elevations in extracellular K+ and membrane depolarisation and decreases with membrane hyperpolarisation (Rakowski et al., 1989, Rang and Ritchie, 1968). Because of the unequal transport of ions, the operation of this pump generates a hyperpolarizing electrical potential and is said to be electrogenic (McCormick, 2013). Paralysis of the
Na+/K+ pump leads to an excess of positive charge within the axon causing membrane depolarisation. Paralysis also abolishes the direct contribution of the hyperpolarising pump current to the membrane potential, leading to extracellular accumulation of K+ and further depolarisation (Krishnan et al., 2009a). Reduced axonal Na+/K+ pump function due to metabolic changes occurring as a result of hyperglycaemia has been demonstrated in animal models of diabetic neuropathy (Greene et al., 1988, Sima,
1996, Stevens et al., 1993). The intracellular Na+ accumulation that would be expected from pump dysfunction may contribute to the osmotic and structural changes that have been observed in diabetic nerves (Breiner et al., 2017, Brismar,
1993).
Disease States
Mutations in the gene encoding for an a-subunit of the Na+/K+ pump have been documented in some patients with familial hemiplegic migraine as well as benign familial infantile epilepsy, and may underlie the hyperexcitability characteristic of these conditions (De Fusco et al., 2003, Vanmolkot et al., 2003).
36 Literature Review
Na+/Ca2+ Exchanger
The Na+/Ca2+ exchanger is a membrane transporter that maintains intracellular Ca2+ homeostasis and is likely to be located in the node (Craner et al., 2004). It composed of 10 transmembrane helices and extrudes one Ca2+ ion while importing 3 Na+ ions
(Liao et al., 2012, Liao et al., 2016). This process is driven by the electrochemical gradient for Na+ and does not require ATP (Waxman and Ritchie, 1993).
Furthermore, the net direction of transport is determined by both the membrane potential and transmembrane Na+ and Ca2+ gradients (Zhang and David, 2016).
Reverse mode operation of the exchanger, resulting in Ca2+ influx and Na+ efflux, would be favoured by axonal depolarisation and elevated intracellular Na+ concentration, both resulting from energy depletion (Zhang and David, 2016).
Furthermore, reduced function of the Na+/K+ pump due to ATP depletion results in axoplasmic accumulation of Na+, which in turn reverses the Na+/Ca2+ exchanger to remove excess Na+ (Freeman et al., 2016). Consequently, there is Ca2+ accumulation and this may activate calpain, a protease capable of inducing proteolytic cleave of neurofilaments, mitochondrial damage, and Wallerian degeneration (Uncini and
Kuwabara, 2015).
37 Literature Review
Clinical Assessment of Peripheral Nerve Structure and Function
Over the recent years, several new techniques have been developed to investigate peripheral nerve function and structure in health and disease (Gasparotti et al., 2017).
In addition to widening the spectrum of diagnostic tools, these advancements have provided new insights into disease pathophysiology and an understanding of the relationship between structure and function (Frank et al., 2019).
In the assessment of large myelinated fibres, there are various electrophysiological techniques available to investigate nerve function (Shabeeb et al., 2018). At present, the gold standard for assessing nerve function in the clinical setting are nerve conduction studies (NCS). Methodologically, NCS involve maximal surface stimulation of a peripheral nerve and measuring the peak response and conduction velocity at the innervated muscle belly or sensory nerve by placing an active electrode on the skin over the recording site (Tavee, 2019). Measurements obtained from NCS are used to determine the number of functioning nerve fibres and speed of conduction to discern the underlying pathology as either axon loss or demyelination, respectively
(Tavee, 2019). Despite their diagnostic utility, NCS are not able to provide further information on the underlying causes of altered conduction, such as changes in membrane potential or ion channel dysfunction (Kiernan et al., 2005a). Another emerging neurophysiological technique is axonal excitability (also known as nerve excitability) which examines the properties underlying the excitability of the axon
(Kiernan et al., 2020, Tomlinson et al., 2018). While NCS utilise maximal surface stimuli and measure impulse conduction between stimulating and recording sites, axonal excitability studies utilise submaximal surface stimuli and provide insight into the function of ion channels, pumps, and exchangers embedded in the axonal
38 Literature Review membrane at the point of stimulation (Kiernan et al., 2020). The use of submaximal stimulation in axonal excitability studies is also advantageous because it is less painful than nerve conduction studies, which is important for patient comfort and compliance throughout the testing period. In the assessment of large fibre structure, magnetic resonance neurography and ultrasonography have emerged as useful tools to complement electrophysiological examination. For example, nerve ultrasonography provides an indication of the changes in nerve structure (fascicular pattern, epineurium, perineurium), the amount of perineural-endoneurial fluid, and the precise location of pathological processes (Gasparotti et al., 2017). The large fibre assessments relevant to this thesis are axonal excitability studies and nerve ultrasonography.
NCS, axonal excitability studies, and neuroimaging cannot detect the pathological processes occurring in small unmyelinated axons (Gasparotti et al., 2017, Kiernan et al., 2020). Conduction properties of small nerve fibres may be investigated using evoked potentials (laser-evoked, contact heat-evoked, and pain-evoked) and microneurography, however each of these techniques have their own inherent limitations (Colloca et al., 2017, Gasparotti et al., 2017). Methods to assess small fibre structure include skin biopsy, which is the gold standard of small fibre loss, and in vivo confocal corneal microscopy (Colloca et al., 2017). Skin biopsy enables the assessment and reliable quantification of density of intraepidermal and subepidermal nerve fibres, making this technique a useful clinical research tool (Lauria et al., 2004,
Lauria et al., 2011). Corneal confocal microscopy allows the visualisation and quantification of the small nerve fibres of trigeminal origin (Oliveira-Soto and Efron,
39 Literature Review
2001). Of the small nerve fibre assessments, only corneal confocal microscopy is of relevance to this thesis.
40 Literature Review
Axonal Excitability
While some properties of axons were investigated sporadically in the early 1900s, it was not until 1970 that the concept of axonal excitability was thoroughly explored in humans (Bergmans, 1970). Bergmans demonstrated that changes in the minimal voltage required to elicit a response from a single motor unit (a motor neuron and the skeletal muscle fibres innervated by that motor neuron) reflected physiological alterations of the axon (Bergmans, 1970). This voltage was termed ‘threshold’.
Bergmans discovered that by measuring changes in the threshold of axons, induced by artificial polarisation or impulse activity, considerable information about axon physiology, such as membrane potential, could be deduced (Bergmans, 1970, Bostock et al., 1998). However, despite the importance of these findings, the technical difficulty of his method and the continual manual adjustment it required proved to be a major hurdle for routine use of these techniques. Ultimately, this hindered the expansion of nerve excitability to the clinical setting.
Axonal Excitability Studies
The development of a computer assisted axonal excitability assessment software,
Qtrac, renewed interest in axonal excitability as a clinical technique for the investigation of peripheral nerve function in health and disease (Kiernan et al., 2000).
Since the introduction of Qtrac, nerve excitability studies have been utilised in vivo to investigate the pathophysiological mechanisms underlying metabolic neuropathies, immune-mediated neuropathies, hereditary neuropathies, channelopathies, neurodegeneration, neurotoxicity, ataxia, and trauma (Kiernan et al., 2020, Tomlinson et al., 2018).
41 Literature Review
Current axonal excitability study protocols utilise ‘threshold tracking’ as the preferred technique (Bostock et al., 1998). The basic premise of threshold tracking is to measure the change in the stimulus strength required to produce a compound action potential of a defined size. This required stimulus is termed the threshold. Typically, the target compound action potential corresponds to ~40% of the maximum compound action potential. Further, tracking a compound actional potential of defined size allows the study of axons of similar size (Bostock et al., 1998, Burke et al.,
2001). When the test response is smaller than the tracking target, the intensity of the subsequent stimulus is increased and conversely, when the response is larger than the tracking target, the stimulus is reduced (Bostock et al., 1998). Responses are tracked as a percentage change in threshold for normalisation.
While threshold is a measure of excitability and may be used as a biomarker of membrane potential, its interpretation may be confounded by various factors
(Krishnan et al., 2009a). It is therefore necessary to assess multiple threshold parameters across various testing paradigms (Kiernan et al., 2020). The key elements involved in a standard axonal excitability assessment protocol include the initial stimulus–response curve followed by four distinct testing paradigms: strength- duration properties, threshold electrotonus, current-threshold relationship, and the recovery cycle. With the exception of strength-duration properties, the other testing paradigms utilise conditioning stimuli and are designed to investigate how axons behave when membrane potential is changed. Paradigms are summarised in Table 2
(see page 53)
42 Literature Review
Stimulus-Response Curve
First, stimulus–response curves are generated in a dose response pattern by which the stimulus output is increased until the maximal compound muscle action potential is achieved (Figure 4). The stimulus required to reach ~40% of the maximal response, the threshold, is then determined and remains the target for the succeeding test paradigms.
Figure 4. Stimulus-response curve. Curve depicts an increase in compound action potential size with increasing stimulus intensity. Threshold refers to stimulus intensity required to elicit a target response (in this case, 40% of the maximal compound action potential). Membrane depolarisation causes a left-ward shift while hyperpolarisation causes a right-ward shift.
43 Literature Review
Strength-Duration Properties
The relationship between stimulus strength and duration dictates that as the stimulus duration is increased, the stimulus intensity required to produce a target response is reduced and vice versa (Weiss, 1901). It was originally described as an exponential relationship and from this, rheobase was defined as the minimum current (of infinite duration) required to elicit a response from the nerve and chronaxie was defined as the stimulus duration at double the rheobase current (Figure 5) (Lapicque, 1909,
Weiss, 1901). Weiss proposed a simple linear relationship between stimulus charge and stimulus duration to describe strength-duration behaviour: Q = a + bt (Weiss,
1901). This linear relationship has aided the derivation of strength-duration properties in axonal excitability studies, as the rheobase can be calculated from the gradient and the strength-duration time constant (SDTC (τSD), analogous to the chronaxie in human peripheral axons) can be calculated from the x-intercept from only two stimulus durations (Figure 5) (Bostock, 1983, Mogyoros et al., 1996). τSD illustrates the rate at which threshold current increases as stimulation duration is reduced, which increases with depolarisation and decreases with hyperpolarisation (Krishnan et al., 2009a,
Mogyoros et al., 1996). τSD reflects the passive properties of the nodal membrane and is a surrogate marker of Nap function (Bostock, 1983, Bostock and Rothwell, 1997,
Mogyoros et al., 1996). As mentioned previously, Nap channels have an important role in repetitive neuronal firing and ectopic symptom generation, making them a clinically relevant axonal excitability measure (Bostock and Rothwell, 1997).
44 Literature Review
A
Rheobase Stimulus StrengthStimulus (mA)
Stimulus Duration (ms) Chronaxie (τSD) B ) mA.ms ( Slope = Rheobase Charge
τSD Threshold
0 Stimulus Duration (ms)
Figure 5. Strength-duration relationship. (A) Plot of stimulus strength versus stimulus duration, illustrating rheobase as the threshold current required for a stimulus of infinite duration and chronaxie as the stimulus duration at which the threshold current is twice the rheobase. (B) Plot of threshold charge versus stimulus width, illustrating the determination of strength-duration time constant using Weiss’ law. Strength–duration time constant is determined as the negative intercept on the x-axis of the line measured using two stimulus widths.
45 Literature Review
Threshold Electrotonus
Threshold electrotonus (TE) examines properties of the internodal membrane
(Kiernan et al., 2020). Assessment of TE is achieved by measuring the change in threshold at multiple time points during and after prolonged (100 ms) subthreshold depolarising (TEd; +40% of threshold) and hyperpolarising (TEh; –40% of threshold) conditioning currents (Figure 6). These conditioning currents are not sufficient to produce an action potential, however they enable a slow spread of current under the myelin sheath into the internode, consequently altering the membrane potential and activating a number of accommodative internodal conductances (Baker et al., 1987,
Kiernan et al., 2020). Threshold electrotonus is the only clinical method of assessing internodal conductances in vivo. As the internodal region accounts for approximately
99.9% of the axonal membrane, this technique provides valuable information regarding the complex changes in excitability that occur with membrane polarisation
(Salzer et al., 2008). The pattern of threshold change to depolarising and hyperpolarising currents can be broken down into distinct characteristic phases.
46 Literature Review
Figure 6. Threshold electrotonus. Curve illustrates change in threshold during and after 100 ms subthreshold polarising currents (depolarising direction plotted upwards in red and hyperpolarisation direction plotted downward in blue). Black arrows indicate changes in threshold electrotonus with membrane depolarisation (‘fanning-in’) while white arrows depict changes with membrane hyperpolarisation (‘fanning-out’).
In response to subthreshold depolarising currents, an initial depolarising fast phase (F phase) develops that is proportional to the applied conditioning current and reflects depolarisation of the node (Krishnan et al., 2009a). The F phase is followed by a slow depolarisation phase (S1 phase) due to the gradual spread of current into the internode
47 Literature Review
and activation of fast KV channels, which limits responses to depolarising currents
(Baker et al., 1987, Bostock et al., 1998, Kiernan et al., 2020). Subsequently, nodal slow KV channels are activated and produce an accommodative response to depolarisation and slow decay toward control threshold (S2 phase) (Baker et al., 1987,
Bostock and Baker, 1988). When the depolarising condition current ends at 100 ms, threshold rapidly returns to baseline and overshoots controls levels before slowly returning to baseline. This undershoot is attributed to the slow deactivation of slow
KV channels and repolarisation of the nerve (Bostock and Rothwell, 1997).
In response to subthreshold hyperpolarising currents, there is an initial hyperpolarising fast phase (F phase) proportional to the applied conditioning current and reflects nodal hyperpolarisation (Kiernan et al., 2020). This is followed by a slow hyperpolarising phase (S1 phase), which is more pronounced than the S1 phase seen in TEd as internodal KV channels are deactivated by hyperpolarisation and thus do not limit the extent of polarisation (Bostock et al., 1998). Finally, extended hyperpolarisation is eventually tempered by the activation of inwardly rectifying cation current (Ih) (Pape, 1996). With cessation of the hyperpolarising stimuli, there is a slow return of threshold to baseline followed by an overshoot as a result of the slow
+ deactivation of Ih and reactivation of persistent Na currents (Kiernan et al., 2020).
Compared to other nerve excitability measures, there is marked variability in TEh, with the between-subject variability being much greater than the within-subject variability (Howells et al., 2013, Tomlinson et al., 2010). This variation is due to differences in the voltage for half-activation of Ih between individuals (Howells et al.,
2012, Jankelowitz et al., 2007a).
48 Literature Review
Importantly, threshold electrotonus is also a sensitive measure of membrane potential
(Baker and Bostock, 1989, Kiernan and Bostock, 2000). Depolarisation of the membrane potential causes an activation of fast and slow KV channels so that conductance is increased and the threshold electrotonus waveform becomes flatter, resulting in a ‘fanning-in’ appearance (Bostock et al., 1998, Kiernan and Bostock,
2000). Hyperpolarisation of the membrane potential reduces the conductance of the internodal membrane, causing a ‘fanning-out’ of the waveform (Bostock et al., 1998,
Kiernan and Bostock, 2000).
Current-Threshold Relationship
The current-threshold relationship (I/V) enables the assessment of the rectifying properties of the axon by using longer lasting conditioning currents of variable strengths. This is achieved by tracking the change in threshold of a 1 ms test impulse following injection of 200 ms conditioning currents stepped from –100%
(hyperpolarising) to +50% (depolarising) of threshold. Strong depolarising currents result in outward rectification which is achieved by the activation of fast and slow KV channels, while strong hyperpolarising currents result in activation of the inwardly rectifying cation conductance, Ih (Kiernan et al., 2000). The current-threshold relationship results in a characteristic plot analogous to a conventional current-voltage plot and enables quantification of the resting input conductance as well as inwardly and outwardly rectifying conductances (Figure 7) (Kiernan et al., 2020). Resting input conductance is observed from the slope between –10% to +10% current injection and will be affected by channels open at resting membrane potential (Kiernan et al.,
2020). The steepness of the curve in the depolarising and hyperpolarising direction reflects outward and inward rectification respectively; the smaller the change in
49 Literature Review threshold for an injected current, the greater the accommodation (Kiernan et al.,
2020). For example, a steepening in response to hyperpolarising currents indicates greater Ih.
50
Outward rectification
0 Resting I/V slope Current (% threshold)
-50 Inward rectification
-100 -500 0 Threshold Reduction (%)
Figure 7. Current-threshold relationship. Polarising currents are 200 ms in duration but vary in strength. Curve is analogous to a current–voltage plot, with response to depolarising current depicted in the upper right quadrant in red and response to hyperpolarizing current in the lower left quadrant in blue.
Recovery Cycle
The recovery cycle utilises paired pulses with varying interstimulus intervals
(between 2 and 200 ms) to assess the recovery of excitability following impulse conduction. Following impulse conduction myelinated axons undergo a well-
50 Literature Review recognised sequence of changes in excitability (Figure 8).
Figure 8. The recovery cycle. Plot illustrates changes in threshold at various time points (2– 200 ms) following a supramaximal conditioning impulse.
Immediately following an action potential, inactivation of transient Na+ channels produces an absolute refractory period, during which the axon is completely inexcitable and no further action potentials can be generated (Hodgkin and Huxley,
1952d). It is impossible to measure the duration of absolute refractoriness because there is a limitation on the stimulus that can be delivered to a patient (Burke et al.,
2001). As Na+ channels recover from inactivation, there is a relative refractory period, which is characterised by decreased excitability. During the relative refractory period, an action potential may be generated with a greater threshold current than normal. In
51 Literature Review human axons, this period typically lasts for 3 ms (Hess et al., 1979, Maurer et al.,
1977). Relative refractoriness varies with membrane potential due to the effect on Na+ channels. Membrane depolarisation increases the relative refractory period due to the greater extent of inactivated of Na+ channels (Burke et al., 2001). Conversely, hyperpolarising shifts in membrane potential decrease the number of inactivated Na+ channels and decreases relative refractoriness (Burke et al., 2001). Unlike most nerve excitability measures, the relative refractory period is extremely temperature sensitive
(Kiernan et al., 2001). While refractoriness is mainly determined by Na+ channels, paranodal demyelination or structural deficiencies may allow fast KV channels to participate (Garg et al., 2018, Howells et al., 2018, Jankelowitz and Burke, 2013).
Following the refractory period, the axon becomes more easily excitable, termed superexcitability. The basis of superexcitability reflects the size of the depolarising afterpotential produced by a passive capacitive charged stored on the internodal membrane (Barrett and Barrett, 1982, Kiernan et al., 1996). This produces a re- excitation of the node by the back-flow of current from the internodal membrane through low resistance pathways under and through the myelin sheath (Barrett and
Barrett, 1982). Superexcitability peaks between 5–7 ms following impulse conduction.
Superexcitability is followed by another period of reduced excitability, termed subexcitability. Subexcitability is due to axonal hyperpolarisation and subsides with the gradual closure of slow KV channels opened during the conditioning discharge
(Lin et al., 2000, Taylor et al., 1992). Axons begin to become subexcitable at 15-20
52 Literature Review
ms and peak subexcitability occurs at approximately 35-40 ms. Resting level of
excitability returns approximately 100 ms following impulse generation.
Table 2. Nerve excitability testing paradigms Testing Channels Outcome Measures Significance and Pathological States Paradigm Assessed
Strength- • Nap • Rheobase • τSD reflects passive properties of nodal duration • SDTC (τSD) membrane properties • Ectopic symptom generation (↑ τSD)
Threshold • Kf • Various time • Examines internodal conductances electronus • Ks points • Membrane depolarisation (‘fanning-in) vs.
• Ih • Peak threshold membrane hyperpolarisation (fanning out’) reduction • S2 accommodation
Current- • Kf • Resting I/V slope • Assesses outward (Kf and Ks activation) threshold • Ks • Minimum I/V and inward (Ih activation) rectifying relationship • Ih slope properties of axon
Recovery • Nat • RRP • Membrane depolarisation (increase in cycle • Kf • Superexcitability RRP) vs. membrane hyperpolarisation
• Ks • Subexcitability (decrease in RRP) • Paranodal demyelination or structural deficiencies
+ + Nat: transient voltage-gated Na channel; Nap: persistent voltage-gated Na channel; Kf: fast
+ + voltage-gated K channel; Ks: slow voltage-gated K channel; Ih: inward rectifier; SDTC
(τSD): strength-duration time constant; RRP: relative refractory period
Sources of Variability in Axonal Excitability Studies
Investigators have examined the influence of particular demographic variables in
studies of axonal excitability variability in healthy controls. There are small but
significant changes with age, conflicting results regarding sex, and a minimal effect
with BMI (Casanova et al., 2014, Jankelowitz et al., 2007b, McHugh et al., 2011).
53 Literature Review
Studies have not explicitly assessed ethnicity as an independent variable on nerve excitability measures. Other sources of variability are site of stimulation, temperature, serum K+, and certain drugs.
Site of Stimulation
Nerve excitability studies examine properties of the axonal membrane at the point of stimulation. Studies have compared nerve excitability recordings of the median nerve
(to abductor pollicis brevis) and ulnar nerve (to abductor digiti minimi) to assess their interchangeability within healthy subjects (Murray and Jankelowitz, 2011). While the current-threshold relationship and recovery cycle were similar, differences were found in the strength-duration time constant and measurements from threshold electrotonus, which questioned the interchangeability of these sites. Of particular relevance to the development of peripheral neuropathy, studies have demonstrated length-dependent differences in nerve excitability between the upper and lower limbs.
Slow KV conductances are more prominent for median axons than for peroneal axons, suggesting that axons innervating the lower limbs have less protection from depolarising stress and could develop ectopic activity more readily (Kuwabara et al.,
2001). This is supported by further evidence that compared to proximal axons in the lower limb, distal axons are more susceptible to ischaemia (Krishnan et al., 2005a).
Effect of Temperature
While the effect of temperature has effects on nerve conduction velocity and amplitude, its effect on axonal excitability is less pronounced in the range of temperatures typically encountered in a clinical setting (Kiernan et al., 2000, Kiernan et al., 2001). Nevertheless, in axonal excitability studies skin temperature should be
54 Literature Review measured at the site of stimulation, be greater than 32 °C, and kept stable throughout the test (Kiernan et al., 2020). The insensitivity of most nerve excitability measures to temperature change is explicable by their strong dependence on membrane potential which is only minimally affected by temperature (Kiernan et al., 2001). However, it must be noted that the relative refractory period is very sensitive to temperature due to alterations in the kinetics of Na+ channel gating (Kiernan et al., 2000). With respect to the range of temperatures that may be encountered clinically (29–35 °C), Na+ channel inactivation is accelerated with warmer temperatures and the relative refractory period is shortened, while with cooler temperatures the relative refractory period is prolonged (Kiernan et al., 2001). At more extreme temperatures, the effects on axonal excitability are more pronounced. With extreme cooling (20 °C) and hyperthermia
(~40 °C), axons show changes in nerve excitability consistent with alterations in membrane potential and changes in accommodation (Howells et al., 2013, Kovalchuk et al., 2018).
Effect of Serum K+
Both intra- and extracellular K+ concentration are well established determinants of impulse generation and conduction (Hodgkin and Huxley, 1952d). The evidence suggesting that serum K+ has an effect on nerve excitability was observed in end- stage renal disease, in which there was various changes in nerve excitability measures consistent with membrane depolarisation that were attributed to the hyperkalaemia
(Krishnan et al., 2005b, Z'Graggen et al., 2006). Further studies confirmed that it was in fact serum K+ responsible for these changes and not other serum solutes. Arnold and colleagues (2014) applied a serum K+ clamp during haemodialysis while all other solutes were removed and motor excitability studies were performed before, during,
55 Literature Review and after the session (Arnold et al., 2014). During the dialysis session, membrane depolarisation was sustained and only normalised after serum K+ had been dialysed to within normal limits. Changes in axonal excitability due to alterations in serum K+ concentration have since been extended to healthy populations and it was found that the relative refractory period and superexcitability were sensitive to serum K+
(Kuwabara et al., 2007). Boërio and colleagues (2014) demonstrated that nerve excitability measures were affected by serum K+ even within the normal physiological ranges of 3.5–4.5 mmol/L (Boërio et al., 2014). Significant differences were noted between low-normal (3.5–3.9 mmol/L) and high-normal (4.3–4.5 mmol/L) values and that the direction of change in axonal excitability recordings was consistent with patterns reported for abnormally high serum K+ in end-stage renal disease.
Effects of Drugs
A number of drugs have purported effects on ion channel function and may therefore influence axonal excitability recordings. Riluzole, mexiletine, lidocaine, and flupirtine have been demonstrated to affect peripheral nerve excitability (Fleckenstein et al.,
2013, Kuwabara et al., 2005, Moldovan et al., 2014, Vucic et al., 2013). However, the exact effects of this drugs remain unclear.
Limitations of Axonal Excitability
While nerve excitability studies are useful tools in providing information regarding axonal ion channel function, there are limitations which must be considered. A significant limitation is that their application is specific to the assessment of large myelinated fibres. Small, unmyelinated fibres may be damaged in various metabolic, infectious, genetic, immune-mediated, and drug-induced conditions (Themistocleous
56 Literature Review et al., 2014). Axonal excitability studies cannot assess small fibre function in these conditions. However small fibre structure and function may be investigated using thermal thresholds, microneurography, evoked potentials, sudomotor function, laser
Doppler flare, skin biopsy, and in vivo corneal confocal microscopy (Malik, 2020).
Another potential limitation of axonal excitability occurs when there is significant axonal degeneration. As mentioned above, nerve excitability studies typically use
40% of the maximal compound muscle action potential as the threshold for all testing paradigms and this allows the study of axons of similar size (Bostock et al., 1998).
However, the nerve excitability measures obtained at this tracking level are only reflective of the axons recruited at that stimulus intensity. This becomes problematic when attempting to compare diseased axons with healthy axons if electrical recruitment of the diseased axons is unpredictable (Shibuta et al., 2010, Shibuta et al.,
2013). Analyses of nerve excitability recordings made at a single tracking level therefore limits the scope of axonal function that can be determined.
Mathematical Modelling of Axonal Excitability
Nerve excitability properties depend on a complex interaction between nodal and internodal ion channels and ion pumps, passive membrane components, and the intracellular and extracellular constituents of the peripheral nerve (Kiernan et al.,
2020). Because of this complexity, deducing the biophysical basis of abnormal nerve excitability recording or predicting the effects of a pathological change is very difficult. This issue has been partially circumvented through the development and refinement of computer-assisted mathematical modelling of nerve excitability recordings to aid interpretation (Boërio et al., 2014, Bostock et al., 1991, Howells et
57 Literature Review al., 2012, Kiernan et al., 2005b). Since its development, modelling has been implemented in nerve excitability studies of metabolic, immune-mediated, toxic, inflammatory, degenerative, and hereditary neuropathies (Garg et al., 2018, Howells et al., 2018, Jankelowitz and Burke, 2013, Kwai et al., 2016a, Liang et al., 2014, Lin et al., 2008).
Nerve Ultrasonography
Ultrasound is defined as a sound wave (oscillating mechanical pressure wave) with frequency above 20 kHz, which is the upper limit of the acoustic range of normal human hearing (Vlassakov and Sala-Blanch, 2015). Sound waves are generated by transducers that consist of piezoelectric crystals. Transducers function both as emitters (transforming an electrical signal into mechanical vibration, known as the reverse piezoelectric effect) and receivers (converting the returning mechanical waves into an electrical signal, known as the piezoelectric effect) of ultrasound (Vlassakov and Sala-Blanch, 2015).
Ultrasonographic assessment of peripheral nerves (the recurrent laryngeal nerve and nerves of the upper and lower limb) in healthy subjects, patients, and cadavers were reported as early as the 1980s (Fornage, 1988, Solbiati et al., 1985). A few years later,
Buchberger and colleagues (1991) demonstrated the utility of ultrasound for the diagnosis of carpal tunnel syndrome (Buchberger et al., 1991). A decade after the commencement of peripheral nerve ultrasound, improvements in image resolution allowed the correlation of the appearance of sonographs with histological findings as well as accurate morphometric studies (Heinemeyer and Reimers, 1999, Silvestri et al., 1995). Diagnostically, peripheral nerve ultrasonography complements
58 Literature Review electrophysiology studies and clinical examination by providing real-time visualisation of normal and abnormal morphology, mobility, and vascularity of nerves and surrounding tissues (Vlassakov and Sala-Blanch, 2015). Imaging in transverse
(cross-sectional) and longitudinal planes, dynamic examination, and colour doppler are available in aiding accurate diagnosis of nerve pathology (Suk et al., 2013). Major limitations of nerve ultrasonography as a technique is that it is heavily operator- dependant and standardised protocols are evolving (Vlassakov and Sala-Blanch,
2015).
Sonographic Features of a Normal Peripheral Nerve
Axons are enveloped by a thin supporting layer known as the endoneurium. A group of axons and their endoneuriums form a fascicle (Figure 9). The nerve fascicle constitutes a distinct structural and functional unit and is surrounded and protected by a thick cellular layer known as the perineurium (Reina et al., 2015). The nerve fascicles in a peripheral nerve are further enveloped and protected by the epineurium, which composed of fibroadipose tissue (Reina et al., 2015). Fascicles can be isolated or clustered in fascicular groups, with connective tissue between them. The epineurium can be divided into a thicker epineural membrane, which surrounds the entire peripheral nerve, and an interfascicular epineurium between fascicles or fascicular groups. The interfascicular epineurium contains a greater amount of fat, intrinsic blood vessels, and other supportive tissues (Reina et al., 2015).
Most peripheral nerves of the upper and lower limbs, including cervical roots and the brachial plexus, can be easily imaged at high-quality using ultrasound provided the nerve is less than 5–6 cm in depth from the skin (Vlassakov and Sala-Blanch, 2015).
59 Literature Review
On cross-sectional scans, distal peripheral nerves have a typical ‘honeycomb’ appearance, with ovoid hypoechoic nerve fascicles dispersed within a milieu of hyperechoic perineurial connective tissue (Vlassakov and Sala-Blanch, 2015).
Longitudinally, the fascicles appear as linear hypoechoic bands. The epineurium usually appears hyperechoic due to its composition of dense connective tissue with high acoustic impedance. Its distinctive appearance helps to delineate nerves from their surrounding structures.
A validated quantitative method of assessing peripheral nerve sonographs is measuring the nerve cross-sectional area in the transverse plane by tracing the inner margin of the epineurium (Figure 9) (Cartwright et al., 2008, Won et al., 2013).
Peripheral nerve cross-sectional measurements have good intraobserver, interobserver, and side-to-side reliability (Cartwright et al., 2013a, Tagliafico and
Martinoli, 2013). The relationship between cross-sectional area with age, weight, height, and sex has varied across studies (Cartwright et al., 2013b). Cross-sectional area varies with site and is usually greater distally and closer to entrapment sites in healthy individuals (Zaidman et al., 2009).
Sonographic Features of an Injured Peripheral Nerve
The principle ultrasound signs of neuropathology are an increase in cross-sectional area and changes in fascicular architecture and echogenicity (Figure 9). Nerve ultrasound may be used in the study of systemic neuropathies in which the peripheral nervous system is widely affected, and focal neuropathies involving a specific nerve at a specific site which are often due to compression. The most common systemic polyneuropathy sonographic sign is the increased cross-sectional nerve area
60 Literature Review
(Vlassakov and Sala-Blanch, 2015). In situations of nerve compression, the most significant ultrasound signs are proximal thickening of the nerve, an abrupt thinning in the area of compression, distal oedema, loss of nerve ultrastructure (Vlassakov and
Sala-Blanch, 2015).
A B
Figure 9. Sonograph of healthy (A) and neuropathic median nerve (B). Image demonstrates increase in median nerve cross-sectional area (yellow dashed lined) in patient with diabetic neuropathy.
In Vivo Corneal Confocal Microscopy
The cornea is the most densely innervated structure in the human body (Shaheen et al., 2014). 50–450 sensory trigeminal neurons transmit nerve fibres via the ophthalmic division of the trigeminal nerve to enter the cornea as stromal nerves, which form the dense sub-basal plexus, and then converge to an inferior-central location termed the inferior whorl (Al-Aqaba et al., 2019, Müller et al., 2003). Free nerve endings originate from these sub-basal nerves (Müller et al., 2003). Density of corneal nerves is greater centrally than peripherally and the entire cornea is estimated to contain 7000 nerve terminals per square millimetre (Marfurt et al., 2010, Müller et al., 2003). Corneal nerves are responsible for the sensations of touch, pain, and temperature and play an important role in the blink reflex, wound healing, and tear
61 Literature Review production and secretion (Shaheen et al., 2014). Approximately 20% of corneal nerves are thinly myelinated Aδ mechanoreceptors that generate acute pain (Shaheen et al., 2014). Another 70% are polymodal, which fire in response to wide range of noxious stimuli (Al-Aqaba et al., 2019). The remaining 10% are C fibre cold receptors (Shaheen et al., 2014). As in other peripheral innervation, non-myelinating
Schwann cells wrap several C fibres together to form a Remak bundle, which is essential for maintenance and function of unmyelinated axons and nociceptors
(Müller et al., 2003). Corneal nerve fibres are highly vulnerable to degeneration following minimal perturbation due to metabolic, toxic, immune, or inflammatory injury but are also capable of regeneration (Azmi et al., 2019, Jia et al., 2018, Lewis et al., 2017, Petropoulos et al., 2020, Sacchetti and Lambiase, 2017).
Corneal nerves can be visualised and quantified using in vivo corneal confocal microscopy (CCM) which has emerged as a rapid, non-invasive, and reliable clinical tool over the last 20 years (Oliveira-Soto and Efron, 2001). CCM requires the examiner to capture multiple images of the sub-basal nerve plexus and analyse them using a manual or automated software (Dehghani et al., 2014, Petropoulos et al.,
2014). Measures of interest include nerve density, length, branch density, tortuosity, and fractal dimension (geometric complexity) in the central and inferior whorl regions of the cornea to assess small fibre damage (Figure 10) (Ferdousi et al., 2020b). CCM has been shown to be a highly reproducible method between testing occasions, examiners, and instruments and normative values have been established (Hertz et al.,
2011, Kalteniece et al., 2017, Petropoulos et al., 2013, Tavakoli et al., 2015). CCM has shown comparable diagnostic efficiency with intraepidermal nerve fibre density via skin biopsy, which is the gold standard to assess small nerve fibre neuropathy but
62 Literature Review is invasive, requires laboratory processing, and may cause infection (Abhishek and
Khunger, 2015, Alam et al., 2017, Chen et al., 2015). CCM has predominantly been used to assess patients with diabetes mellitus, but the technique has also been used to demonstrate small-fibre damage in chemotherapy-induced neuropathy, human immunodeficiency virus neuropathy, Fabry’s disease, Charcot-Marie-Tooth disease type 1A, inflammatory neuropathies, as well as Parkinson’s disease, multiple sclerosis, and amyotrophic lateral sclerosis (Markoulli et al., 2018, Petropoulos et al.,
2020). While CCM is emerging as a useful tool in the investigation and diagnosis of diabetic neuropathy, specific studies have observed no differences in measures between patients with and without diabetic neuropathy (Andersen et al., 2018a,
Petropoulos et al., 2015). As the cornea is avascular and the nerve fibres lack myelination, corneal nerves are especially susceptible to the metabolic perturbations that occur in diabetes (Petropoulos et al., 2020). Other evidence also points to impaired blood flow to the trigeminal ganglion in animal models of diabetes as a cause of degeneration (Davidson et al., 2012). A major limitation of corneal confocal microscopy as a technique is that only one-thousandth of the total cornea surface area can be imaged at a time (typically 0.16 mm2) (Allgeier et al., 2018). This small field of view is insufficient to reliably characterise corneal nerve morphology, especially in longitudinal studies. To address this, several mosaicking techniques have been developed (Allgeier et al., 2018). Disadvantages of corneal confocal microscopy include the operator-dependant nature of the technique, patient discomfort, and necessity of a topical anaesthetic.
63 Literature Review
A B
C D
Figure 10: Confocal micrograph of healthy (A, B) and neuropathic (C, D) cornea. Figure demonstrates the normal corneal nerve architecture in the central (A) and inferior whorl (B) regions. In contrast, a loss of corneal nerves in the central (C) and inferior whorl (D) regions is observed in a patient with diabetic neuropathy.
64 Literature Review
Diabetic Neuropathy
Diabetic neuropathy is the most prevalent chronic complication of both type 1 and type 2 diabetes and occurs in more than half of patients with diabetes (Pop-Busui et al., 2017, Selvarajah et al., 2019). Diabetic neuropathy may affect different parts of the nervous system and thus present with diverse clinical manifestations (Feldman et al., 2019). The most common type of neuropathy is distal symmetric polyneuropathy
(DSPN) and accounts for approximately 75% of diabetic neuropathies (Albers and
Pop-Busui, 2014, Dyck et al., 2011). Estimates of the incidence of DSPN after the diagnosis of diabetes vary. Evidence from large observational cohort studies, the
Diabetes Control and Complications Trial (DCCT), and Epidemiology of Diabetes
Interventions and Complications (EDIC) Study suggests at least 20% of type 1 diabetes patients develop DSPN after 20 years and 50% of type 2 diabetes patients develop DSPN after 10 years (Pop-Busui et al., 2017). DSPN places a substantial burden on healthcare systems, society, and the affected individual. In the USA, the total annual cost of managing symptomatic DSPN and its complications was estimated to be between US$4.6 billion and $13.7 billion (Feldman et al., 2019,
Selvarajah et al., 2019). DSPN is also the key initiating factor for the development of foot ulceration and most common cause of lower-limb amputation in high-income countries (Armstrong et al., 2017, Boulton et al., 2018). For the individual, DSPN is associated with increased morbidity, distressing neuropathic pain, and significant impairments in quality of life (Ziegler et al., 2014). The major predictors of DSPN are duration of diabetes and poor glycaemic control (Grisold et al., 2017, Tesfaye et al.,
2005). Other risk factors for the development of DSPN in both type 1 and type 2 diabetes are older age, increased height, smoking, alcohol abuse, hyperlipidaemia, hypertension, and presence of other diabetic complications (Callaghan et al., 2015,
65 Literature Review
Ziegler et al., 2014). In type 2 diabetes, obesity and insulin resistance have also been associated with DSPN (Andersen et al., 2018b, Callaghan et al., 2020, Christensen et al., 2020, Han et al., 2015, O'Brien et al., 2017, Ylitalo et al., 2011, Zhou et al., 2020).
Furthermore, there is strong association between the number of metabolic syndrome components and DSPN (Bonadonna et al., 2006, Callaghan et al., 2016a, Callaghan et al., 2016b, Costa et al., 2004, Hanewinckel et al., 2016a). With respect to prevention, rigorous glucose control can decrease the incidence of DSPN in type 1 diabetes but has little effect in type 2 diabetes, which suggests different mechanisms underlie
DSPN in each condition (Callaghan Brian C. et al., 2012, Callaghan B. C. et al., 2012,
Gaede et al., 2008). In type 2 cohorts, insulin-sensitising therapies have been shown to reduce the incidence of neuropathy (Pop-Busui et al., 2013).
Clinical Presentation of DSPN
DSPN manifests with a ‘stocking and glove’ distribution whereby the hands and lower limbs are usually affected (Feldman et al., 2019). Symptoms occur in a length- dependent fashion such that they start distally at the toes, spread proximally, and then to the upper limb digits when the symptoms reach the knees (Feldman et al., 2019).
The most common early symptoms involve small fibres which include pain and dysesthesias (Malik et al., 2011, Pop-Busui et al., 2017). Pain is present in up to 25% of patients with DSPN and may be described as burning, lancinating, tingling, or shooting and is typically worse at night (Pop-Busui et al., 2017). Pain may also be accompanied with paraesthesia, allodynia, and hyperalgesia (Pop-Busui et al., 2017).
Reductions in intra-epidermal and corneal nerve fibres at the inferior whorl, which is the most distal extension of corneal nerve fibres, are among the earliest changes
(Edwards et al., 2017, Ferdousi et al., 2020b, Malik et al., 2011, Petropoulos et al.,
66 Literature Review
2015). Involvement of large fibres, which usually occurs later in the disease course, may cause numbness, tingling without pain, loss of protective sensation (which is particularly indicative of DSPN), abnormal ankle reflexes (which occurs early), reduced vibration sensation, impaired joint position, and diminished touch-pressure sensations (Albers and Pop-Busui, 2014, Pop-Busui et al., 2017). Clinical findings of
DSPN are a loss of sensation to pinprick and cold temperature, which are both mediated by small fibres, and loss of vibration and proprioception, which are both mediated by large fibres, in a ‘stocking and glove’ distribution (Feldman et al., 2019).
With advanced DSPN, weakness of the small muscles of the foot and dorsiflexors is observed and foot drop may also develop (Albers and Pop-Busui, 2014, Feldman et al., 2019).
Nerve conduction studies are the current gold standard for the diagnosis of DSPN
(Selvarajah et al., 2019). Findings in DSPN include reduced amplitudes, decreased conduction velocities, and prolonged F responses (Weisman et al., 2013). Typically, reductions in sensory fibre amplitudes are observed before decreases in the amplitude of motor fibres (Feldman et al., 2019). Abnormalities in nerve conduction (which may be subclinical) appears to be first objective and quantitative indication of DSPN and is necessary in the research setting to confirm the diagnosis (Dyck et al., 2011, Tesfaye et al., 2010). However, it should be noted that nerve conduction studies are not required to meet the definition of DSPN in routine clinical practice, and as such, the diagnosis of DSPN could be held to more rigorous standards. Polyneuropathy due to vitamin deficiencies (e.g. B12), other endocrine and metabolic disorders (e.g. hypothyroidism), or with an infectious, inflammatory, toxin, inherited, vascular, and neoplastic basis must be excluded (Zochodne, 2014). In addition to conventional
67 Literature Review
nerve conduction studies, other electrophysiological measures may have some utility
in the diagnosis of DSPN (Shabeeb et al., 2018).
Pathogenesis and Pathophysiology of DSPN
Research into the field of diabetic neuropathy has largely focused on the metabolic
and/or redox state of the dorsal root ganglion and Schwann cells (Feldman et al.,
2017). Various pathways have been implicated in the pathogenesis of DSPN (Figure
11). Ultimately, these pathways lead to an excess formation of mitochondrial and
cytosolic reactive oxygen species, inflammation, and Na+/K+ pump dysfunction
(Feldman et al., 2017). Indeed, axonal excitability studies have added to our
understanding of DSPN, however these techniques are limited to the study of large
fibres. Nevertheless, investigations using axonal excitability are discussed where
relevant.
Insulin Hypergylcaemia Metabolic Syndrome Resistance/Deficiency
Increased glucose Polyol PKC Hexosamine Dyslipidaemia variability pathway pathway pathway
Increased Increased Increased Mitochondrial Excessive free sorbitol diacylglycerol F-6-P dysfunction fatty acids Neuronal PKC NADPH depletion activation
Na+/K+ pump Oxidative stress Inflammation dysfunction
Impaired Axoglial Axonal Nerve regeneration dysjunction degeneration dysfunction
Neuropathy
Figure 11. Pathogenesis of distal symmetric polyneuropathy. PKC: protein kinase C; F-6-P:
fructose-6-phosphate; NADPH: nicotinamide adenine dinucleotide phosphate.
68 Literature Review
Polyol, Hexosamine, and Protein Kinase C Pathways
Excess glucose in hyperglycaemia is converted to sorbitol by aldose reductase, which results in osmotic imbalance in the cell and compensatory efflux of myoinositol and taurine (Greene et al., 1988). Loss of myoinositol impairs normal nerve physiology as it is an essential component for the Na+/K+ pump (Greene, 1986a, 1986b, Greene and
Lattimer, 1984). Consequently, this would result in an accumulation of intracellular sodium and hence, reduction in the trans-axonal Na+ gradient. Further injury to neurons occurs as aldose reductase activity depletes cellular stores of nicotinamide adenine dinucleotide phosphate (NADPH) which is required for the regeneration of glutathione, an essential antioxidant (Feldman et al., 2017). This leads to the generation of cytoplasmic reactive oxygen species which cause cellular dysfunction
(Oates, 2008). Furthermore, sorbitol is metabolised to fructose which can create further osmotic swelling (Oates, 2008).
Evidence of a reduction in the trans-axonal Na+ gradient was observed in early axonal excitability studies of DSPN. In these studies, hyperglycaemia was associated with excitability changes indicative of reduced nodal Na+ currents, namely reductions in both the SDTC and relative refractory period and an increase in the rheobase (Kitano et al., 2004, Misawa et al., 2006a, Misawa et al., 2005b, Misawa et al., 2004).
Treatment of hyperglycaemia was observed to reverse these observations, suggesting an increase of Na+ currents presumably through restoration of the trans-axonal Na+ gradient (Kitano et al., 2004, Kuwabara and Misawa, 2008). It is important to note changes in the relative refractory period (increase or decrease) have varied between studies, however it tends to increase with the severity of DSPN (see below). Clinical
69 Literature Review trial findings demonstrated that administration of epalrestat, an aldose reductase inhibitor, was found to rapidly increase nodal Na+ currents, as measured by SDTC, and improve nerve conduction in patients with DSPN (Misawa et al., 2006b). Clinical trials of aldose reductase inhibitors have largely failed to treat DSPN but improvements in nerve conduction parameters have been observed in some cases
(Feldman et al., 2017, Sekiguchi et al., 2019). Finally, while a reduction in Na+ currents seems to be implicated in DSPN, it should be noted that an abnormal increase in nodal Na+ currents is associated with neuropathic pain and worsening quality of life measures in patients with diabetes (Kwai et al., 2013, Misawa et al.,
2006a, Misawa et al., 2009).
The hexosamine and protein kinase C (PKC) pathways are also involved in peripheral nerve injury. Increased glycolysis in response to hyperglycaemia results in a glycolysis intermediate, fructose-6-phosphate, entering the hexosamine pathway, which undergoes a series of reactions that ultimately results in damage to peripheral nerves (Feldman et al., 2017). In addition to the polyol pathway, a disruption to
Na+/K+ pump function may occur via the PKC pathway. Increased glycolysis results in the accumulation of diacylglycerol which activates neuronal PKC, and in turn, disrupts function of the Na+/K+ pump and perhaps impairs nerve blood flow (Geraldes and King, 2010).
The possibility that Na+/K+ pump dysfunction may play a role in the development of
DSPN in diabetes has also been supported by axonal excitability studies. Excitability recordings undertaken in patients with DSPN by Krishan and colleagues (2005) demonstrated a ‘fanning-in’ appearance of the threshold electrotonus, consistent with
70 Literature Review axonal depolarisation potentially from impaired pump activity (Krishnan and Kiernan,
2005). Other findings from this study were reductions in the relative refractory period and an increase in superexcitability, consistent with a reduction in nodal Na+ conductance suggesting that a reduction in Na+/K+ pump activity is coupled with decreased nodal Na+ currents (Krishnan and Kiernan, 2005). Impaired Na+/K+ pump function in diabetes is further supported by excitability studies which implemented dynamic manoeuvres such as limb ischaemia and maximal voluntary contraction to demonstrate decreased pump activity in DSPN (Krishnan et al., 2008, Kuwabara et al., 2002). Other axonal excitability studies have demonstrated changes indicative of reductions in nodal Na+ conductances and Na+/K+ pump function in diabetes patients without DSPN or with subclinical DSPN, which worsen with increasing severity of
DSPN. (Bae et al., 2011, Sung et al., 2012). With neuropathy progression there is
‘fanning- in’ of threshold electrotonus, a decrease in subexcitability, and increases in the rheobase, relative refractory period, and superexcitability (Sung et al., 2012).
In animal models of DSPN, intracellular Na+ accumulation and decreased Na+/K+ pump function have shown to disrupt the nodal and paranodal regions through axoglial dysjunction and myelin retraction (Sima and Brismar, 1985, Sima et al.,
1986). It would therefore be expected that the function of nodal and paranodal channels would be affected in patients with diabetes. Indeed, early axonal excitability studies in patients with DSPN demonstrated that hyperglycaemia caused a reduction in nodal and paranodal K+ conductances (Misawa et al., 2005a). These observations were further substantiated with axonal excitability studies undertaken in type 1 diabetes patients without neuropathy demonstrating similar results, suggesting these reductions occur early in the disease course (Arnold et al., 2013a, Kwai et al., 2016a).
71 Literature Review
Mathematical modelling of these recordings indicated there were reductions in a range of Na+ and K+ conductances in the nodal and paranodal regions of the axon
(Kwai et al., 2016a).
Large scale alterations in peripheral nerve structure are also observed in DSPN with ultrasonography (Lee and Dauphinée, 2005). Increase in sorbitol is thought to cause oedema which increases peripheral nerve cross-sectional area (Vlassakov and Sala-
Blanch, 2015). Significant enlargement of cross-sectional area of the median and tibial nerves has been validated as a good parameter in the diagnosis of DSPN (Riazi et al., 2012, Watanabe et al., 2009, Watanabe et al., 2010). Association studies using nerve ultrasound and nerve excitability have demonstrated that while cross-sectional area increases with DSPN in type 1 and type 2 diabetes, cross-sectional area correlates with worsening nerve excitability measures in type 1 diabetes only (Borire et al., 2018b). This differential relationship lends further credence to the notion that different pathophysiological mechanisms may underlie DSPN in type 1 and type 2 diabetes. Further to this notion, the pathophysiological mechanisms of peripheral nerve dysfunction in LADA, which was described as an admixture of type 1 and type
2 diabetes, are yet to be to be characterised. This will be examined in Chapter 4.
Insulin, C-peptide, and Glycaemic Variability
Although insulin does not directly control glucose transport into the peripheral nervous system, it is recognised as a neurotrophic factor and insulin deficiency is associated with the development of diabetic neuropathy (Feldman et al., 2017,
Zaharia et al., 2019). Insulin receptors are expressed on peripheral nerves, intraneural mitochondria, and on Schwann cells which is of significance given the interplay
72 Literature Review between Schwann cells and axons that has recently emerged as a key mediator in the pathogenesis of DSPN (Feldman et al., 2019, Goncalves et al., 2017). In type 1 diabetes, insulin deficiency is accompanied with decreases in C-peptide. C-peptide has been observed to be neuroprotective in experimental studies, possibly through an increase in Na+/K+ pump function, but only marginal benefits were seen clinical supplementation in type 1 diabetes patients with DSPN (Wahren et al., 2016, Wahren and Larsson, 2015). In type 2 diabetes, C-peptide is inversely associated with DSPN, independent of confounding factors (Qiao et al., 2017).
In type 1 diabetes, the mode of insulin delivery has been observed to have an effect on ion channel function. Kwai and colleagues (2015) found that patients receiving CSII had axonal excitability parameters preserved within normal limits at follow-up in contrast to patients receiving MDII, who displayed prominent abnormalities (Kwai et al., 2015). The possibility that the apparent normalisation of nerve function with CSII treatment was due to a greater stability in acute glucose variability was considered.
Subsequent studies demonstrated that greater acute glucose variation, as measured by mean amplitude of glycaemic excursion (which is postulated to cause oxidative stress), was positively associated with worsening axonal excitability measures (Kwai et al., 2016b). The effect of other acute measures of glucose control, namely continuous overall net glycaemic action and time in range, on peripheral nerve structure and function are yet to be explored. These associations will be investigated in Chapter 3.
In type 2 diabetes, insulin-sensitisation treatments have been shown to decrease the incidence of neuropathy compared to insulin-providing treatments (Pop-Busui et al.,
73 Literature Review
2013). Insulin resistance has been observed to develop in experimental models of type
2 diabetes and could play a role in the pathogenesis of DSPN (Grote and Wright,
2016, Kim and Feldman, 2012). Evidence of this concept is lacking in human studies however. While treatment of DSPN primarily focuses on symptomatic management, animal studies have observed neuroprotective effects of commonly used medication used to treat type 2 diabetes (El Mouhayyar et al., 2020, Pop-Busui et al., 2017). The effect of these medications on axonal excitability will be investigated in Chapter 6.
The Metabolic Syndrome
In type 2 diabetes, the components of the metabolic syndrome promote the onset and progression of diabetic neuropathy independent of hyperglycaemia (Callaghan et al.,
2016a, Feldman et al., 2017). Animal models of metabolic syndrome demonstrate there is neuronal dysfunction and oxidative stress, even in normoglycaemia (Cortez et al., 2014). Dyslipidaemia increases the formation of acylcarnitine in peripheral nerves and an excess leads to mitochondrial dysfunction and apoptotic fission (Stino et al.,
2020). Dyslipidaemia can also lead to an increase in oxidized low-density lipoprotein
(LDL) and advanced glycation end-LDL, and both can activate various pro- inflammatory signalling pathways that disrupt mitochondrial function (Eid et al.,
2019, Stino et al., 2020). The excessive free fatty acids catabolised by b-oxidation in response to hyperlipidaemia are harmful to the peripheral nervous system, especially
Schwann cells (Feldman et al., 2019). As mentioned above, there is a strong association between the number of metabolic syndrome components and DSPN, however the underlying mechanisms are yet to be examined. These pathophysiological mechanisms will be studied in Chapter 5.
74 Literature Review
Uraemic Neuropathy
Neurological complications are a major cause of morbidity in CKD and affect both the CNS and PNS causing encephalopathy, stroke, cognitive impairment, restless leg syndrome, autonomic disturbances, and neuromuscular disorders such as peripheral neuropathy (known as uraemic neuropathy) and myopathy (Arnold et al., 2016a,
Baumgaertel et al., 2014, Chillon et al., 2016, Vellanki and Bansal, 2015). Uraemic neuropathy is a distal sensorimotor polyneuropathy caused by uremic toxins. Despite improvements in dialysis care, uraemic neuropathy remains the most common neurological disorder in end-stage kidney disease and occurs in 60-100% of dialysis patients as well as approximately 70% of pre-dialysis patients (Aggarwal et al., 2013,
Karunaratne et al., 2018, Krishnan et al., 2005b, Krishnan et al., 2009b, Laaksonen et al., 2002, Tilki et al., 2009).
Clinical Presentation
Uraemic neuropathy typically presents with sensory symptoms such as paraesthesia, burning, and numbness which progresses slowly in a length-dependent fashion
(Basilio Vagner, 2011, Krishnan and Kiernan, 2009, Said, 2013). Given the length- dependent nature of uraemic neuropathy, there is preferential involvement of distal nerves in the lower limbs than upper limbs (Krishnan and Kiernan, 2007, Said, 2013).
Clinical examination in early stages reveals symptoms and signs confined to the lower limbs, including distal sensory loss to vibration and reduced ankle deep tendon reflexes (Krishnan and Kiernan, 2009, Said, 2013). With more severe disease, upper limb involvement may occur in a ‘stocking and glove’ distribution (Krishnan et al.,
2009b, Said, 2013). In advanced cases, motor nerve involvement occurs leading to muscle atrophy and weakness, which is most prominent distally (Krishnan and
75 Literature Review
Kiernan, 2009). Pain, which is usually absent in the early stages, may become prominent with advanced neuropathy (Said, 2013). Patients presenting with both
CKD and diabetes develop length-dependent neuropathy of greater severity (Krishnan and Kiernan, 2009). Currently, no instrument has been formally validated to assess the severity of peripheral neuropathy in patients with CKD. This will be addressed in
Chapter 1 of this thesis, in which the Total Neuropathy Score will be validated for
CKD patients with and without diabetes. Furthermore, in patients with both CKD and diabetes, the relative contributions of each of these conditions to neuropathy development is yet to be determined. This question will be addressed in Chapter 2.
Nerve conduction studies are the gold standard for the diagnosis of uraemic neuropathy (Krishnan et al., 2009b). Nerve conduction studies in CKD patients with neuropathy demonstrate features of generalised neuropathy of the axonal type with reductions in sensory amplitudes and, to a lesser extent, reduced motor amplitudes with relative preservation of conduction velocities (Krishnan and Kiernan, 2009).
Reduction in the sural sensory amplitude is the most sensitive nerve conduction parameter in the diagnosis of uraemic neuropathy (Laaksonen et al., 2002).
Pathogenesis and Pathophysiology
Observational studies in the 1970s indicated there were lower rates of neuropathy in patients treated with peritoneal dialysis compared with those on haemodialysis (Babb et al., 1971). This led to the theory that accumulation of ‘‘middle molecules,’’ substances with a molecular weight of 300–12000 Da such as b2-microglobulin and parathyroid hormone, were the cause of uraemic neuropathy and the better clearance of these molecules by the peritoneal membrane compared to standard haemodialysis
76 Literature Review membranes explained the reduced neuropathy prevalence with peritoneal dialysis
(Babb et al., 1981, Dhondt et al., 2000). However, the middle molecule hypothesis has been disputed on a number of bases (Kjellstrand et al., 1972). The main criticism was the lack of evidence that any of middle molecules are neurotoxic, with the exception of parathyroid hormone (Kjellstrand, 1981, Vanholder et al., 1994).
However, human studies examining the relationship between parathyroid hormone and nerve conduction velocity yielded conflicting results (Avram et al., 1978, Di
Giulio et al., 1978, Schaefer et al., 1980).
Despite this setback, the hypothesis of a dialysable toxin remained prevalent but the mechanism remained unclear. Many substances have been investigated as a potential uraemic neurotoxin including urea, creatinine, guanidine, methylguanidine, guanidinosuccinic acid, uric acid, oxalic acid, phenols, aromatic hydroxyacids, indican, amines, myo-inositol, amino acids and neurotransmitters, though none of these have yielded compelling results or findings were restricted to in vitro studies
(Anand et al., 2019, Bostock et al., 2004).
The hypothesis that neurotoxic effect of a substance may be caused by an alteration in membrane excitability was proposed by Nielsen based on in vitro studies of muscle and red blood cells obtained from dialysis patients (Bittar, 1967, Nielsen, 1973, Welt et al., 1964). Nielsen hypothesised that one or more of the toxins known to accumulate in uraemia may cause neuropathy by inhibiting the activity of the axonal
Na+/K+ pump (Nielsen, 1973). As the Na+/K+ pump is electrogenic with 3 Na+ ions being extruded for every 2 K+ ions pumped into the axon, inhibition leads to an excess of positive charge on the inner aspect of the axonal membrane and an accumulation of
77 Literature Review extracellular K+, which causes further depolarisation (Kaji and Sumner, 1989).
Disruption of these ion gradients created by the Na+/K+ pump may cause reverse operation of the Na+/Ca2+ exchanger, leading to increased levels of intracellular Ca2+ and axonal loss (Craner et al., 2004).
Kiernan and colleagues (2002) were the first to demonstrate evidence of axonal membrane depolarisation, as assessed by axonal excitability, primarily due to hyperkalaemia in end-stage kidney disease (Kiernan et al., 2002). While clinical features of uraemic neuropathy are predominantly lower limb, these excitability studies were conducted on the median nerve and yet demonstrated significant alterations in excitability prior to dialysis (Kiernan et al., 2002, Krishnan et al.,
2006b). Nerve excitability studies were extended to the lower limb in dialysis patients and findings again indicated membrane depolarisation was due to serum K+ and not other substances such as parathyroid hormone, b2-microglobulin, or urea (Krishnan et al., 2005b). Krishnan and colleagues (2005) noted a correlation between the severity of symptoms and nerve excitability abnormalities, suggesting that the altered membrane potential may be directly related to neuropathic symptoms (Krishnan et al.,
2005b).
Despite the strong evidence of K+ as the cause of membrane depolarisation in end- stage kidney disease, the possibility this was due to impaired Na+/K+ pump function remained. Na+/K+ pump function was subsequently studied in dialysis patients using manoeuvres such as limb tourniquet and forced maximal voluntary contraction
(Krishnan et al., 2006a, Krishnan et al., 2006c). These findings argued against Na+/K+ pump impairment and instead suggested heightened pump activity as a compensatory
78 Literature Review mechanism for membrane potential changes in dialysis patients. Hyperkalaemia as the causative agent of membrane depolarisation in dialysis patients was subsequently confirmed using serum K+ clamp studies (Arnold et al., 2014). In addition to causing alterations in nerve excitability, elevations in serum K+, but not parathyroid hormone
or b2-microglobulin, were also found to correspond with nerve enlargement, which reduced after dialysis (Borire et al., 2017).
While these studies implicated hyperkalaemia in the pathophysiology of uraemic neuropathy in end-stage kidney disease, recent clinical trial evidence has highlighted the role of serum K+ in uraemic neuropathy in the pre-dialysis populations as well
(Arnold et al., 2017). Restriction of dietary potassium has been shown to prevent nerve excitability abnormalities indicative of membrane depolarisation at follow-up in stage 3–4 CKD (Arnold et al., 2017). Furthermore, potassium restriction prevented the progression of neuropathy and resulted in an improvement in gait speed (Arnold et al., 2017).
79 Methodology
Methodology
80 Methodology
Recruitment of Patients and Control Subjects
Studies were approved by the South East Sydney Area Health Service Human
Research Ethics Committee (Northern Section) and the Human Research Ethics
Committee of the University of New South Wales. All participants enrolled provided written informed consent to the procedures in accordance with the Declaration of
Helsinki.
Recruitment of Patients
All patients were recruited from the Diabetes Mellitus Centre and the Kidney Care
Centre at the Prince of Wales Hospital in Sydney, Australia. The entire patient cohort consisted of patients with a diagnosis of either type 1 diabetes, type 2 diabetes, latent autoimmune diabetes of adults, chronic kidney disease, and diabetic kidney disease.
All patients enrolled had been diagnosed with their respective conditions for a minimum of one year. Patients were excluded from participating in studies if they had any of the following: renal transplant, a history of neurotoxic/neuromodulatory treatment, carpal tunnel syndrome, peripheral oedema, an additional condition known to cause neuropathy (such as vitamin B12 deficiency), or a neuromuscular, movement, psychiatric, or developmental disorder. In studies utilising in-vivo corneal confocal microscopy, exclusion criteria included current eye infections, corneal abrasions, history of refractive surgery, trauma to anterior segment, or contact lens wear.
Recruitment of Controls
Control subjects were recruited through Prince of Wales Hospital and the University of New South Wales. All control subjects underwent clinical assessment to exclude neuropathy and underwent testing under the same conditions as patients. In addition
81 Methodology to the exclusion criteria above, control participants were excluded from studies if they had a history of taking medication to treat blood pressure, elevated glucose, or lipids.
Equipment and Materials
Axonal Excitability
• High Performance AC Amplifier (ICP511 AC amplifier, Grass Technologies, West Warwick, US) for amplifying nerve excitability recordings
• Data acquisition device (DAQ PCI-6221); Shielded connector block (BNC- 2110); Cable (SHC-68-68-EPM); (National Instruments, Austin, USA) used to convert recordings to a digital signal
• Isolated linear bipolar constant current stimulator (DS5, Digitimer, Welwyn Garden City, UK)
• Hum Bug 50/60 Hz Noise Eliminator (Quest Scientific Instruments, North Vancouver, Canada) used to cancel background electrical noise
• Automated threshold tracking and analysis software - QTRAC (Institute of Neurology, Queen Square, London) with TROND axonal excitability protocol
• Conventional non-polarizable ECG electrodes (Unilect 7831Q; Unomedical, Stonehouse, Great Britain) to provide surface stimulation/reference for surface recordings
• Electrosurgical neutral electrode (Unilect 2406M, Unomedical, Stonehouse, Great Britain) used as ground electrodes
• Red Dot Trace Prep (2236, 3M Canada) for abrading skin prior to electrode placement
82 Methodology
• Thermistor thermometer (5831-A, Omega Engineering, Manchester, UK) used to measure skin temperature
Nerve Ultrasonography
• MyLabOne system with a 10–18 MHz linear probe (Esaote, Genoa, Italy) used for ultrasound scans • Aquasonic Clear ultrasound gel (Parker Laboratories, New Jersey, USA) to provide a conductive medium
In-vivo Corneal Confocal Microscopy
• Heidelberg Retinal Tomograph III Rostock Cornea Module (Heidelberg Engineering GmbH, Heidelberg, Germany) for corneal confocal microscopy
• ACCMetrics Corneal Nerve Fibre Analyser (University of Manchester, Manchester, UK) for confocal image analysis
Nerve Conduction Studies
• Medelec Synergy EMG system (Oxford Instruments, Old Woking, UK) used for nerve conduction studies
• Surface EMG recording electrodes (Kendall Soft-E, H69P, Tyco Healthcare, Gosport, UK).
Neuropathy Assessment
• Neurotip disposable needles (Owen Mumford Ltd., Oxford, UK) for assessing pinprick sensation.
• 128 Hz tuning fork used to measure vibration sense
• Clinical reflex hammer to test tendon reflexes
83 Methodology
Glucose Variability
• iPro CGM System (Medtronic, CA, USA) for blinded continuous glucose recording
• EnliteTM sensor (Medtronic, CA, USA) for measurement of interstitial glucose
• ACCU-CHEK Performa Glucometer (Roche Diagnostics, Mannheim, Germany) for measuring on the spot blood glucose levels obtained by finger- stick.
• Gylcemic Variability Analyzer Program (Óbuda University, Budapest, Hungary)
Axonal Excitability Assessment and Mathematical Modelling
Axonal excitability studies provide important insights into the biophysical properties of human axons in health and disease. Excitability studies enable the rapid acquisition of multiple excitability variables by measuring the change in stimulus required to elicit a target response (~40% of maximal compound action potential) following various patterns of conditioning stimuli. The key elements involved in a standard axonal excitability assessment protocol include the initial stimulus–response curve followed by four distinct testing paradigms: strength-duration properties, threshold electrotonus, current-threshold relationship, and the recovery cycle. With the exception of strength-duration properties, the other testing paradigms utilise conditioning stimuli and are designed to investigate how axons behave when membrane potential is changed. The information gained from these paradigms provides indirect measures of axonal membrane potential and the activity of ion
84 Methodology channels, energy dependent pumps and ion exchange processes involved in impulse conduction (see Axonal Excitability in Literature Review, page 41).
All nerve excitability studies in this thesis were undertaken on the median motor nerve. To reduce skin impedance 3M Red Dot Trace Prep (2236, 3M Canada) was used to prepare both the stimulating and recording sites. Non-polarisable electrodes
(Unilect 7831Q; Unomedical, Stonehouse, Great Britain) were utilised to stimulate the median nerve. The cathode was placed over the optimal site of median nerve stimulation (site of least resistance) and the anode placed 10 cm proximal over the radial bone. The recording electrode was positioned over the motor point of the APB muscle, the reference electrode was placed ~4 cm distally on the proximal phalanx and the earth plate (Unilect 2406M, Unomedical, Stonehouse, Great Britain) was placed on the palm. Temperature was monitored close to the site of stimulation and maintained at >32 °C throughout the study (Figure 12).
85 Methodology
Computer QTRAC
Stimulator
Recording electrodes
Earth Stimulating electrodes
Figure 12. Median motor nerve excitability set-up. Image demonstrates electrode placement and equipment for the computerised QTRAC system with stimulation delivered through an isolated bipolar constant current stimulator.
Stimulation was controlled by a computerised threshold tracking system, QTRAC
(Institute of Neurology, Queen Square, UK), converted to current, and delivered by an isolated linear bipolar constant current stimulator (DS5, Digitimer, Welwyn Garden
City, UK). Compound motor action potentials were recorded by the QTRAC software following amplification (ICP511 AC amplifier, Grass Technologies, West Warwick,
US) and digitisation (DAQ PCI-6221; Shielded connector block BNC-2110; Cable
SHC-68-68-EPM; National Instruments, Austin, USA). A Hum Bug 50/60 Hz Noise
Eliminator (Quest Scientific Instruments, North Vancouver, Canada) was used to remove excess electronic noise. QTRAC threshold tracking software utilised a proportional tracking system that altered the stimulus intensity according to the
86 Methodology difference between actual and target responses, i.e. stimulus intensity was automatically increased or decreased to achieve an action potential of a defined size.
Responses were tracked as a percentage change in threshold for normalisation.
Stimulus-Response Curve
To begin, stimulus–response curves were generated by an incremental (1 mA) increase in the stimulus intensity of a 1 ms impulse until a maximal response was obtained. Reversing this process, the QTRAC software automatically decreases the stimulus until the response returns to zero. At each level of current decrease, the average of three responses are taken. A target response of the steepest point on the stimulus-response curve (~40% of the maximal compound action potential, see Figure
4, page 43) was automatically chosen. The stimulus required to elicit this target response (or change there-of), termed the threshold, was the quantitative value measured for the remainder of the protocol. Test pulses for threshold were 1 ms for all paradigms.
Strength-Duration Properties
Strength-duration properties were assessed by measuring the change in stimulus required to elicit threshold utilising four different stimulus widths: 0.2 ms, 0.4 ms, 0.8 ms and 1 ms. The value reflecting the strength-duration time constant was calculated from the x-axis intercept of the linear relationship between stimulus intensity and stimulus duration derived using Weiss’ formula (see Figure 5, page 45) (Bostock,
1983, Mogyoros et al., 1996).
87 Methodology
Threshold Electrotonus
Threshold electrotonus (TE) was determined by plotting the percentage of threshold change when test pulses were applied during and after 100 ms of subthreshold depolarising (TEd; +40% of threshold) and hyperpolarising (TEh; –40% of threshold) conditioning currents. A total of 26 time points were assessed and more than 10 measures are acquired from this paradigm. By convention TE is plotted as threshold reduction on the y-axis and time on the x-axis (see Figure 6, page 47).
In response to TEd and TEh, measures are determined from the average percentage threshold change between 10–20 ms, 20–40 ms, and 90–100 ms. After cessation of conditioning currents, TEd(undershoot) and TEh(overshoot) are calculated as the mean value of maximal threshold reduction in the 20 ms following cessation of the conditioning stimulus. Additional parameters calculated from responses to TEd include peak threshold reduction and S2 accommodation, which is defined as the difference between peak threshold reduction and TEd(90–100 ms).
Current-Threshold Relationship
The current-threshold (I/V) relationship also utilised subthreshold conditioning currents. However, unlike TE, these currents are longer in duration (200 ms) and their intensity is altered in 10% increments from –100% (hyperpolarising) to +50%
(depolarising) of threshold. The change in excitability induced by these condition currents was assessed at a single time-point following the 200 ms polarising current.
Given that the slope of the I/V relationship is a threshold analogue of input conductance, all I/V variables are expressed as the slope of this relationship between various polarising stimulus intensities (see Figure 7, page 50). There are three
88 Methodology variables obtained from the I/V protocol: resting I/V slope (calculated from the polarising currents –10% to +10%), minimum I/V slope (calculated as the minimum of the curve obtained by fitting a straight line to each three adjacent points in turn), and hyperpolarising I/V slope (calculated from the three most hyperpolarised intervals).
Recovery Cycle
The recovery cycle assessed the restoration of axonal excitability following a supramaximal stimulus. The change in threshold was measured at 18 conditioning-test intervals from 2 to 200 ms after a supramaximal stimulus was delivered. To avoid contamination during the short conditioning-test intervals, the response from a single supramaximal conditioning stimulus was subtracted from the conditioning + test response.
The recovery cycle of excitability has a well-defined and reproducible series of events from which three variables are measured (see Figure 8, page 51). First, the relative refractory period, measured in milliseconds, is the time taken to return to baseline threshold (or to cross the x–axis). Following this, a period of enhanced excitability, termed superexcitability, was calculated as the maximal percentage threshold reduction, averaged from 3 adjacent points. The final phase, subexcitability, was calculated as the maximal percentage threshold increase after 10 ms, averaged from 3 adjacent points.
89 Methodology
Mathematical Modelling
Mathematical Modelling of axonal excitability recordings were undertaken in select studies to investigate the pathological basis of axonal dysfunction. Axonal excitability recordings were analysed using the using the Bostock model of axonal excitability, which is a validated model of the human axon based on a single node and internode connected by paranodal pathways through and under the myelin sheath (Bostock et al., 1991, Jankelowitz et al., 2007a, Kiernan et al., 2005b). The model assists in the interpretation of excitability findings between control and disease data by providing an indication of the underlying changes in and around the axonal membrane in the disease state. This includes changes in the maximal conductance and permeabilities of different types of Na+ and K+ ion channels, alterations in pump currents, biophysical properties, and surrounding ionic concentrations (Figure 13). An approximation of the actual distribution of voltage-gated K+ channels in myelinated axons is achieved by the inclusion of slow and fast K+ channels in both compartments, with slow K+ channels predominantly located at the node, and fast K+ channels predominantly at the internode. Transient and persistent voltage-gated Na+ channels were modelled only at the nodal membrane and the hyperpolarisation-activated cation current was restricted to the internodal axolemma. Leak conductances, Na+/K+–ATPase pump currents and axolemmal capacitances were modelled in both axonal compartments.
The Barrett-Barrett conductance, which represents current flow through and underneath the myelin sheath between the node and internode, was also investigated.
90 Methodology
RestingResting membraneMembrane Pump potentialpotential current
Permeability/ Conductance
Capacitance
Ion concentrations
Figure 13. Mathematical model parameters. Image depicts that changes in maximal conductance and permeabilities of different types of Na+ and K+ ion channels, alterations in pump currents, various biophysical properties, and surrounding ionic concentrations may be investigated.
The model was first adjusted to fit the mean nerve excitability data obtained from the
control group before fitting the mean data of the disease group. Modelling analyses
involved changes in a single or a combination of parameters in a hypothesis-driven
and iterative fashion to objectively fit simulated excitability data with the mean
recorded data as closely as possible using a least squares approach. The overall
discrepancy was assessed and minimised using the weighted sum of the squares of the
error terms between the control and disease group data of the four excitability
91 Methodology paradigms: strength-duration behaviour, threshold electrotonus, current-threshold relationship, and recovery cycle. Weighting factors of these paradigms were 0.5, 1, 1, and 2, respectively and were kept constant for each analysis. Minimum interstimulus interval for the recovery cycle was set at 3 ms. Analyses were run in unclamped mode to permit secondary changes in resting membrane potential caused by changes in conductances or pump currents.
Nerve Ultrasonography
Median nerve ultrasound was performed prior to electrophysiological testing. Imaging was performed with a MyLabOne system with an adjustable 10–18 MHz linear array transducer (Esaote, Italy) using the ‘Musculoskeletal’ factory preset (acoustic power
100%, line density set at medium, dynamic range set at 14, persistence set at 1) and other settings (focus, gain, and depth) were kept constant for each examination. Nerve ultrasound was conducted while participants were sitting comfortably with their forearm fully supinated and fingers semi-extended. The forearm was supported on an armrest to ensure the elbow was flexed at 90°. The median nerve was first identified in the transverse plane at the carpal tunnel inlet at the level of the pisiform bone and then tracked proximally between the superficial (flexor digitorum superficialis) and deep (flexor pollicis longus and flexor digitorum profundus) muscles until the junction of the middle and distal third of the forearm (Figure 14). The median nerve cross-sectional area was then measured on the screen (in mm2) with a stylus using the continuous trace method by outlining the inner margin of the epineurium, which is a validated measure in nerve ultrasound (Cartwright et al., 2008, Won et al., 2013).
92 Methodology
A B
Figure 14. Median nerve ultrasonography. (A) Image depicts median nerve was tracked proximally from the carpal tunnel to the junction of the middle and distal third of the forearm. (B) Image demonstrates measurement of median nerve cross-sectional area.
In-vivo Corneal Confocal Microscopy
Participants were scanned bilaterally with a corneal confocal microscope to visualise corneal nerve morphology (Heidelberg Engineering GmbH, Heidelberg, Germany)
(Figure 15). Prior to image acquisition, topical anaesthesia was applied to the cornea of both eyes and the instrument was positioned onto the centre of the cornea. During the examination, the participant was instructed to fixate on a roving black target on a computer screen placed 35 cm perpendicular to their contralateral eye in order to image nerves in the central cornea and nerves at the inferior whorl. The microscope was set on ‘sequence mode’ and 100 contiguous images of the sub-basal nerve plexus were captured over 40 seconds. For capturing images of the inferior whorl, the roving target was shifted higher to an estimated superonasal area of participant’s contralateral eye. The inferior whorl was subsequently located inferior and slightly nasal to the corneal apex. 100 contiguous images were captured over 40 seconds.
Eight central and three to four inferior whorl images, not overlapping by more than
20%, from both eyes of each participant were selected for quantification. For the central cornea, images with bright and vertically arranged fibres were selected, while
93 Methodology for the inferior whorl, images that captured the entire region were chosen
(Petropoulos et al., 2015, Vagenas et al., 2012). Images were analysed using a validated and fully automated nerve analysis software (Corneal Nerve Fiber Analyzer
V.2, ACCMetrics, University of Manchester, Manchester, United Kingdom) to quantify corneal nerve fibre length (total length of main nerves and nerve branches per square millimeter), density (number of main nerves per square millimeter), branch density (total number of main nerve branches per square millimeter), and fractal dimension (a measure of structural complexity with a lower value indicating less corneal nerve intricacy) in the central cornea (Dehghani et al., 2014, Petropoulos et al., 2014). Corneal nerve fibre length and fractal dimension were also quantified in the inferior whorl. Corneal nerve measures are presented as an average of both eyes.
Microneuromas were defined as nerve abnormalities that present as irregularly shaped, terminal enlargements of nerve endings with variable hyper-reflectivity and poorly defined margins, which were singular or in clusters of two or three.
Microneuromas were counted manually and the same microneuroma imaged in more than one frame was considered as a count of one.
Figure 15. Corneal confocal microscopy
94 Methodology
Nerve Conduction Studies
Nerve conduction studies were performed as part of the clinical assessments of neuropathy in all patients. The tibial and sural nerves were assessed as per standard neurophysiology protocols to determine the compound muscle action potential and sensory nerve action potential upon maximal stimulation, respectively (Liveson and
Ma, 1992). Conduction velocities were also assessed. All studies were performed using a Medelec Synergy EMG system (Oxford Instruments, Old Woking, UK) and surface EMG recording electrodes (Kendall Soft-E, H69P, Tyco Healthcare, Gosport,
UK). The tibial nerve was assessed by orthodromic stimulation between the medial malleolus and calcaneal tendon and recording from abductor hallucis. The sural nerve was assessed by antidromic stimulation approximately at the junction of the middle and lower thirds of the leg (10–16 cm proximal to the lateral malleolus), just lateral to the midline, and recording between the lateral malleolus and calcaneal tendon
(Liveson and Ma, 1992).
Clinical Assessment of Neuropathy
Total Neuropathy Score
Neuropathy presence and severity was determined using the Total Neuropathy Score
(TNS) (Figure 16). The TNS is comprised of eight items which are each scored from
0–4. The scores from the eight items of the TNS were summed to give a total score from 0–32, with a higher score indicating more severe neuropathy and zero indicating an absence of neuropathy. The TNS was administered in a scripted fashion.
95 Methodology
Total Neuropathy Score (TNS)
Figure 16. Total Neuropathy Score. Scoring system for the eight items of the Total Neuropathy Score are shown.
Participants were first asked if they had experienced any abnormal sensations in their
limbs such as feelings of numbness, pins and needles or prickling. If the subject
responded positively, a score ranging from 1–4 was assigned based on how far these
sensations extended proximally: in the fingers or toes (1), up to the ankles or wrists
(2), up to knees or elbows (3) or above (4). Patients were then asked if they felt weak
in their arms or legs. If the subject acknowledged a deficit, this weakness was graded
from 1–4 based on the amount difficulty experienced by the participant: slight (1),
moderate (2), requiring assistance (3) or paralysis (4).
96 Methodology
Pinprick sensibility was assessed using a Neurotip (Owen Mumford, U.K.) The sharp and blunt ends of the Neurotip were first pressed against the participants’ upper limb to allow them to distinguish between the two ends. After the subject had closed their eyes, the sharp and the blunt ends of the Neurotip were then pressed on the most distal aspect of each toe and finger, gradually more proximally over the surface of the foot and hand, and eventually to the knee and elbow until the participant reported the object to be sharp. A reduction in pinprick sensibility in the fingers or toes was scored as (1), up to the wrist or ankle was scored as (2), up to the elbow or knee was scored as (3) or even more proximally was scored as (4). Vibration sensibility was assessed using a 128-Hz tuning fork. The vibratory stimulus was first introduced on the upper limb and subjects were asked if they could feel the vibration or “buzz”. After the subject had closed their eyes, the tuning fork was placed over the most distal aspect of each toe and finger, gradually more proximally over the medial/lateral malleoli and medial/lateral wrist, and eventually at the medial/lateral knee and elbow, until the subject reported they felt the vibration. Vibration sensibility was graded in a similar manner to pinprick sensibility.
Strength was graded using manual muscle tests of ankle dorsiflexion as outlined in
Medical Research Council (MRC) Guidelines (Medical Research Council of the
United Kingdom, 1976). Mild weakness (MRC grade 4) was assigned a score of (1), moderate weakness (MRC grade 3) was assigned a score of (2), severe weakness
(MRC grade 2) was assigned a score of (3) and paralysis (MRC grade 1–0) was assigned a score of (4). Assessments of deep tendon reflexes were first examined at the ankle and then at the knee. If reinforcement was required to induce the ankle reflex this was scored as (1) and if the ankle reflex was completely absent, this was
97 Methodology scored as (2). If reinforcement was required to induce the patellar reflex this was scored as (3) or (4) if this was completely absent too.
Nerve conduction studies, as described above, were then undertaken to assess the compound muscle action potential and sensory nerve action potential upon maximal stimulation of the tibial and sural nerve, respectively. A score graded from 1–4 was attributed depending on how far the amplitude fell below the lower limit of the normal range, determined from an internal normative data set.
Modified Toronto Clinical Neuropathy Score
In select studies where sensory symptoms required heavy weighting, neuropathy severity and presence were assessed using the modified Toronto Clinical Neuropathy
Score. The modified Toronto Clinical Neuropathy Score is validated for the assessment of diabetic distal symmetric polyneuropathy and is comprised of 11 items
(Bril et al., 2009). Symptoms (foot pain, numbness, tingling, weakness, ataxia, and upper limb involvement) were graded from 0–3 depending on their presence and interference with the participant’s activities of daily living. When present, symptoms without interference with sense of well‐being or activities of daily living are graded as
(1), those which interfere with sense of well‐being, but not with activities of daily living as (2), or those which interfere with both as (3). Sensory tests (pinprick, light touch, temperature, vibration, and position sense) were graded from 0–3 depending on the degree of deficit distally. If abnormal, each test was scored as (1) if the deficit was at toes only, (2) if the deficit was between toes and ankle, or (3) if the deficit was above the ankle. Pinprick was assessed using a Neurotip (Owen Mumford, U.K.) and temperature was assessed using a cold tuning fork. Nerve conduction studies, as
98 Methodology outlined above, were also undertaken in the tibial and sural nerves. Presence of neuropathy was defined using the Toronto Diabetic Neuropathy Expert Group definition of confirmed diabetic neuropathy: the presence of nerve conduction abnormality and a sign or symptom of neuropathy (Tesfaye et al., 2010).
Glucose Variability
In select studies where glucose variability was of interest, participants with type 1 diabetes underwent six days of blinded continuous glucose monitoring (iPro,
Medtronic, CA, USA; sensor: EnliteTM, Medtronic, CA, USA). Glucose recordings were analysed using Glycemic Variability Analyzer Program, which has been validated for the assessment of glucose variability (Figure 17) (Marics et al., 2015).
Variables calculated were time in and above range, expressed as a percentage of total monitoring time, and continuous overall net glycaemic action (CONGA), which can assess intra-day and inter-day glucose variability. Time in range was considered when glucose was between 3.9–10.0 mmol/L (70–180 mg/dL) and time above range was defined when glucose values exceeded 10.0 mmol/L (180 mg/dL) (Battelino et al.,
2019). CONGA is defined as the standard deviation of the summed differences between the current glucose observation and an observation n hours prior, with a higher CONGA indicating greater glycaemic variation (McDonnell et al., 2005). The interval, n, was 1 hour. CONGA was calculated for each day of monitoring and averaged. Mathematically, CONGA is calculated using the formula: