Motor Unit Integrity in Pathophysiological States and the Assessment of Potential Neuroprotective Therapeutics

Home , Alpha motor neuron, Motor unit, Perineurium

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in the Graduate School of the Ohio State University

Christopher G. Wier

Neuroscience Graduate Program

The Ohio State University

2018

Dissertation Committee:

Dr. Stephen Kolb, Advisor

Dr. Tony Brown

Dr. Dana McTigue

Dr. Chen Gu

Christopher G. Wier

2018

ABSTRACT

The motor unit (MU), comprised of a single lower motor neuron and all the muscle fibers it innervates, is the final common pathway of the neuromuscular system. Functional MUs are essential in executing the motor demands of the upper motor neurons. Several neuropathological stresses—including: amyotrophic lateral sclerosis (ALS), peripheral nerve injury (PNI), spinal muscular atrophy or aging—can converge at the MU, negatively affecting MU integrity. While various methods exist that can assess the different properties of the MU—connectivity and contractility—the relationship between these properties during neuropathological stress is still poorly understood.

Here, we longitudinally measured MU integrity in models of ALS and PNI using a combination of MU connectivity, contractility and behavioral assessments. We found that loss of muscle contractility is an early defect of

SOD1(G93A) mice—one model for ALS—and precedes MU connectivity decline, suggesting that muscle may also be an important early ALS therapeutic target site. We utilized longitudinal assessments of MU connectivity and behavior in

SOD1(G93A) mice to determine the therapeutic efficacy of overexpressing small

ii heat shock protein B1 (HSPB1), demonstrated to be neuroprotective in several neurodegenerative models, via AAV9-HSPB1 delivery. Our results indicated no significant neuroprotective effect in SOD1(G93A) mice receiving AAV9-HSPB1.

We also longitudinally mapped MU regeneration following a PNI and demonstrated early contractility recovery that aligned with behavioral recovery, and preceding MU connectivity recovery. These results suggest a latent recovery phase of MU reconnectivity that follows early muscle strength recovery.

Following a PNI, mice were treated with either AAV9-HSPB1 or AAV9-SMN to overexpress HSPB1 or survival motor neuron protein (SMN)—another potential neuroprotective protein. Our data did not demonstrate significant accelerated MU recovery following AAV9-HSPB1 or AAV9-SMN delivery.

This body of work illustrates the strengths of longitudinally measuring MU connectivity and contractility. We propose novel pathological mechanisms underlying neuropathological insult to the MU which can potentially be targets for future therapeutics. Furthermore, we were able to more thoroughly assess the efficacy of potential therapeutics towards neuromuscular integrity.

iii

DEDICATION

For my parents, Larry and Mary; my siblings: Nick, Lauren, Sam and Alex; my

aunt, Mary Jo; and lastly my fiancée, Hillary—your love and support is the

foundation of everything I have accomplished.

ACKNOWLEDGEMENTS

There have been countless individuals whom I want to acknowledge and thank for this dissertation. Chief amongst them is my advisor, Dr. Stephen Kolb;

Steve, your patience and pushing me to improve have made me the person I am today—not only as a researcher, but as a person as well. Words simply cannot describe how lucky and thankful I am to have you as a mentor. I would also like to acknowledge my un-official second mentor, Dr. W. David Arnold; Dave, you took me under your wing and taught me the skills put to use in this dissertation and were always there for me for paper edits, or a much needed happy hour.

To my committee members—Dr. Dana McTigue, Dr. Tony Brown, and Dr.

Chen Gu—your feedback and support have helped make this dissertation possible. I’d like to thank you all for your time and effort as well as being exemplary critical minds and pushing me to improve as a researcher.

I’d also like to thank Dr. Kevin Foust, whom designed and contributed the viral vectors used in this dissertation. Also, thank you Dr. Patrick Heilman, specifically for having designed the HSPB1 viral vector before I joined the Kolb lab. These have helped make this dissertation possible.

Current and past members of the Kolb and Arnold Labs have contributed to this body of work in innumerable ways. Samantha Renusch and Patrick

Heilman, you two were the first lab-mates I worked alongside in the Kolb lab. I could not have asked for two better people to spend my time with. You both kept me grounded during our time together. Also, thank you Anthony Reynolds, Kajri

Sheth and Alex Crum—you three have been phenomenal undergraduates and you guys helping to perform animal assessments have been very helpful. Lastly, thanks to Deepti Chugh and Chitra Iyer—whether it was sharing tea with me, discussing papers, counting NMJs or some good natured ribbing, you both helped this unofficial Arnold lab member feel at home.

Members of the Kaspar/Meyer lab have also been instrumental resources, particularly Dr. Kathrin Meyer and Shibi Likhite. Thank you both for your time, resources and assistance with delivering virus to new born mouse pups and helping me perform mouse behavioral measurements.

I would also like to thank the Center for Biostatistics, especially Dr. David

Kline and Marilly Palettas for your guidance and assistance in developing models of our assessments. I would like to thank the Neuroscience Imaging Core at Ohio

State University and Paula Monsma for their technical expertise in acquiring NMJ images.

To my amazing and supportive parents, Larry and Mary—thank you both so much. The love and care you two provided me is more than what I could have hoped for. I have to also thank my Aunt Mary Jo, who for as long as I

can remember has always supported and encouraged my passion for all things science. I also could not have finished this work without the support of the greatest siblings—Nick, Lauren, Sam, and Alex. We have had our fair share of good times and bad times, but you all were always there to keep my ego in check. You all will always be my oldest and best friends. To my unofficial emotional support dog, Jackson—you can’t read this, but who’s a good boy? You are—you’re a good boy!

Lastly, but certainly not least, I want to thank my better half and fiancée—

Hillary. You met me as I was finishing my first year in graduate school and for some reason you decided to stick around. You have been my unbreakable rock of support these last five years. I couldn’t have done this without you, I love you.

vii

VITA

June 2012………………………………………………B.S. Biology, University of Wisconsin, Madison

2012-Present…...... Ph.D. Candidate in Neuroscience, The Ohio State University

PUBLICATIONS

Wier, C.G., Crum, A.E., Reynolds, A.B., Iyer, C.C., Chugh, D., Palettas, M.S., Heilman, P.L., Kline, D.M., Arnold, W.D., and Kolb, S.J. Muscle dysfunction precedes loss of motor unit connectivity in SOD1(G93A) mice. Muscle and Nerve. 2018 (Accepted)

Sheth, K.A., Iyer, C.C, Wier, C.G., Crum, A.E., Bratasz, A., Kolb, S.J., Clark, B.C., Burghes, A.H.M., and Arnold, W.D. Muscle strength and size are associated with motor unit connectivity in aged mice. Neurobiology of Aging 2018; 67: 128-136

Heilman, P.L., Song, S.W., Miranda, C., Meyer, K., Knapp, A.R., Wier, C.G., Kaspar, B.K., and Kolb, S.J. Hereditary neuropathy-associated HSPB1 mutations disrupt non-cell autonomous protection of motor neurons. Experimental Neurology 2017; 297:101-109

viii

PUBLICATIONS, CONTINUED

Arnold, W.D., Sheth, K.A., Wier, C.G., Kissel, J.T., Burghes, A.H.M., and Kolb, S.J. Electrophysiological Motor Unit Number Estimation (MUNE) Measuring Compound Muscle Action Potential (CMAP) in Mouse Hindlimb Muscles. Journal of Visualized Experiments 2015;103

FIELDS OF STUDY

Major field: Neuroscience

TABLE OF CONTENTS

Abstract ...... ii Dedication ...... iv Acknowledgements ...... v Vita ...... viii Fields of Study ...... x Table of Contents ...... xi List of Tables ...... xvii List of Figures ...... xviii Chapter 1: Overview of the neuromuscular system and the motor unit ...... 1 1.1 Input to lower motor neurons ...... 1 1.1.1 Upper motor neurons ...... 1 1.1.2 Sensory neurons (Direct—Type Ia) ...... 3 1.1.3 Interneurons ...... 4 1.2 Lower motor neuron anatomy, physiology, and organization with target muscle fibers ...... 5 1.2.1 Motor neuron anatomy and physiology ...... 6 1.2.1.1 Cell body overview ...... 6 1.2.1.2 Axon overview ...... 7 1.2.1.3 Axon terminal (Neuromuscular junction) overview ...... 9 1.3 Excitation-contraction coupling: conversion of electrical signals to force output ...10 1.3.1 Excitation-contraction coupling ...... 10 1.3.2 Excitation of muscle to influx of Ca2+ in cytosol ...... 10 1.3.3 Organization of sarcomeres in skeletal muscle fibers ...... 11 1.3.4 Sarcomere force transmission ...... 13

1.4 Motor unit organization and function ...... 14 1.4.1 Motor unit organization ...... 15 1.4.2 Motor unit regulation of force generation...... 17 1.5 Neuropathological insults converging at the motor unit ...... 19 1.5.1 Amyotrophic lateral sclerosis ...... 20 1.5.1.1 SOD1(G93A) transgenic mouse model of ALS ...... 20 1.5.1.2 Function of healthy SOD1 ...... 21 1.5.1.3 SOD1 mutations and the generation of an ALS mouse model ...... 21 1.5.1.4 Disease progression in SOD1(G93A) mice ...... 22 1.5.1.5 Effect of SOD1 mutations ...... 25 1.5.2 Peripheral nerve injury ...... 27 1.5.2.1 Peripheral axon anatomy ...... 27 1.5.2.2 Classification of peripheral nerve injuries ...... 28 1.5.2.3 Response to PNI – injury site and distal axon ...... 29 1.5.2.4 Response to PNI – neuromuscular junction/muscle ...... 30 1.5.2.5 Peripheral nerve recovery ...... 31 1.6 Overview of assessing motor unit ...... 33 1.6.1 MU Connectivity outcome measurements ...... 33 1.6.1.1 Compound muscle action potential ...... 33 1.6.1.2 Motor unit number estimate ...... 35 1.6.2 Muscle contractility measurements ...... 38 1.6.1.3 Muscle contraction ...... 38 1.6.3 Pathological assessment of the mu ...... 39 1.6.3.1 Neuromuscular junction reinnervation ...... 39 1.6.4 Behavioral recovery ...... 40 1.6.4.1 Grip strength ...... 40 1.6.4.2 Rotarod ...... 40 1.7 Application of MU connectivity and contractility in stressed states ...... 41 Chapter 2: Muscle dysfunction precedes loss of motor unit connectivity in SOD1(G93A) mice ...... 43 2.1 Statement of Contribution ...... 43 2.2 Introduction ...... 43

xii

2.3 Materials and methods ...... 47 2.3.1 Animals...... 47 2.3.2 Behavioral assessments ...... 48 2.3.3 Mouse anesthesia and animal preparation ...... 48 2.3.4 Electrophysiology ...... 49 2.3.5 Muscle contractility ...... 49 2.3.6 NMJ imaging and quantification ...... 50 2.3.7 Intrarater testing ...... 51 2.3.8 Assessment of longitudinal SOD1(G93A) disease progression ...... 51 2.3.9 Statistics ...... 52 2.4 Results ...... 53 2.4.1 Intrarater reliability of muscle contractility and MU connectivity measurements ...... 53 2.4.2 Decline of muscle contractility and loss of motor unit connectivity are early features in SOD1(G93A) mice ...... 54 2.4.3 Muscle atrophy occurs after loss of muscle contractility ...... 61 2.4.4 NMJ innervation is reduced at P70 ...... 62 2.4.5 Correlations between muscle contractility, MU connectivity, and behavioral assessments ...... 65 2.5 Discussion ...... 66 2.5.1 Loss of muscle contractility in vivo aligns with prior in situ studies ...... 67 2.5.2 Loss of muscle contractility preceded loss of MU connectivity ...... 68 2.5.3 Loss of muscle contractility and MU connectivity are sexually dimorphic in SOD1(G93A) mice ...... 70 2.5.4 In vivo muscle contractility is a potential pre-clinical physiological biomarker .72 2.6 Conclusion ...... 72 Chapter 3: AAV9 mediated overexpression of HSPB1 does not influence disease progression in the SOD1(G93A) mouse model of Amyotrophic Lateral Sclerosis ...... 74 3.1 Statement of Contribution ...... 74 3.2 Introduction ...... 74 3.3 Materials and methods ...... 78 3.3.1 Vectors ...... 78 3.3.2 Mice ...... 78

xiii

3.3.3 Viral injections ...... 78 3.3.4 Behavioral and survival assessments ...... 79 3.3.5 Electrophysiological recordings ...... 79 3.3.6 Perfusion and tissue processing ...... 80 3.3.7 Histology ...... 80 3.3.7.1 Immunohistochemistry ...... 80 3.3.7.2 Cresyl violet staining ...... 81 3.3.8 Western blot ...... 81 3.3.9 Motor neuron quantification ...... 82 3.3.10 Statistics ...... 82 3.4 Results ...... 83 3.4.1 Human HSPB1 overexpressed following AAV9-HSPB1 injection in SOD1(G93A) mice ...... 83 3.4.2 HSPB1 overexpression does not rescue motor performance or improve disease progression ...... 84 3.4.3 Viral overexpression of HSPB1 has no effect on electrophysiological measures of motor unit function in SOD1(G93A) mice ...... 89 3.4.4 Endstage motor neuron counts are not significantly affected following HSPB1 overexpression ...... 91 3.5 Discussion ...... 92 3.5.1 Low dose aav9-hspb1 does not exhibit neuroprotection ...... 92 3.5.2 Possible efficacy of hspb1 overexpression in non-neuronal cell types ...... 93 3.6 Conclusions ...... 94 Chapter 4: Time course of Motor Unit recovery in the triceps surae following sciatic nerve injury ...... 95 4.1 Statement of Contribution ...... 95 4.2 Introduction ...... 95 4.3 Methods ...... 100 4.3.1 Animals...... 100 4.3.2 Surgery ...... 100 4.3.3 Recovery timeline ...... 101 4.3.4 Behavior - hindlimb grip strength ...... 102 4.3.5 Electrophysiology ...... 102

xiv

4.3.6 Muscle physiology ...... 105 4.3.7 Perfusion and tissue processing ...... 105 4.3.8 Immunofluorescence ...... 105 4.3.9 Statistics ...... 106 4.4 Results ...... 107 4.4.1 Muscle contractility recovery precedes MU connectivity regeneration ...... 107 4.4.2 Muscle contractility recovery aligns with grip strength recovery ...... 110 4.4.3 Soleus muscle partially innervated early in regeneration ...... 111 4.4.4 Correlations between muscle contractility, MU connectivity and behavior during regeneration ...... 113 4.5 Discussion ...... 114 4.5.1 Discrepancy between mu connectivity recovery and muscle contractility suggests early and latent recovery phases ...... 115 4.5.2 Behavioral function and underlying muscle contractility ...... 118 Chapter 5: Assessing the regenerative efficacy of AAV9-SMN and AAV9-HSPB1 following a peripheral nerve injury ...... 120 5.1 Statement of Contribution ...... 120 5.2 Introduction ...... 120 5.3 Materials and Methods ...... 125 5.3.1 Animals...... 125 5.3.2 Viral vectors ...... 126 5.3.2.1 AAV9-GFP ...... 126 5.3.2.2 AAV9-SMN ...... 126 5.3.2.3 AAV9-HSPB1 ...... 126 5.3.3 Surgical procedures ...... 127 5.3.3.1 Sciatic nerve crush ...... 127 5.3.3.2 Cisterna magna injection ...... 127 5.3.4 Recovery timeline ...... 129 5.3.5 Behavior ...... 130 5.3.6 Electrophysiology ...... 130 5.3.7 Perfusion and tissue processing ...... 131 5.3.8 Immunofluorescence ...... 131 5.3.8.1 AAV9 transduction ...... 131

5.3.8.2 NMJ quantification ...... 132 5.3.9 Western blot ...... 132 5.3.10 Statistics ...... 133 5.4 Results ...... 133 5.4.1 Nervous tissue transfected following CM injection ...... 133 5.4.2 AAV9 Overexpression of SMN or HSPB1 does not accelerate MU connectivity following PNI ...... 134 5.4.3 AAV9-SMN treated mice demonstrated mild hindlimb grip strength increase late in recovery Timeline ...... 136 5.4.4 NMJ reinnervation recovered by 60dpi ...... 138 5.5 Discussion ...... 141 5.5.1 AAV9-SMN and AAV9-HSPB1 do not accelerate MU recovery following PNI ...... 142 5.5.2 Neuroprotective effects may be masked by naturally robust recovery in mice ...... 142 Chapter 6: Conclusion ...... 145 6.1 Conclusion ...... 145 SOD1(G93A) utility as a model for ALS ...... 150 Peripheral nerve injury and potential therapeutics ...... 154 Utility of AAV9 to deliver potential therapeutics ...... 159 Human candidate genes in animal models ...... 161 Concluding remarks...... 163 References ...... 165 Appendix A: Figures ...... 192 Appendix B: Tables ...... 196

xvi

LIST OF TABLES

Table 2.1 Intrarater reliability correlation coefficients...... 54 Table 2.2 Onset of muscle contractility and MU connectivity decline ...... 60 Table 3.1 AAV9-HSPB1 has no effect on ALS disease symptoms in SOD1(G93A) mice ...... 86 Table 4.1 Longitudinal recovery of MU connectivity, muscle contractility behavior following PNI...... 108 Table 4.2 Pearson correlations during nerve recovery between muscle contractility, MU connectivity and behavioral measurements ...... 114 Table 5.1 Ipsilateral hindlimb grip strength recovery normalized to body weight ...... 138 Table B.1 Longitudinal outcome measurements in SOD1(G93A) male and wildtype male mice...... 196 Table B.2 Longitudinal outcome measurements in SOD1(G93A) female and wildtype female mice...... 198 Table B.3 Electrophysiological, contractile and behavioral correlations ...... 201 Table B.4 Counts of motor neurons transduced ...... 202

xvii

LIST OF FIGURES

Figure 1.1 Descending pathway of the corticospinal tract ...... 3 Figure 1.2 Structural organization of the sarcomere ...... 13 Figure 1.3 Representative schematic of the motor unit ...... 17 Figure 1.4 Representative force-frequency waveform ...... 19 Figure 1.5 Overview of peripheral axon anatomy...... 28 Figure 2.1 SOD1(G93A) males demonstrate earlier muscle contractility decline than SOD1(G93A) females ...... 57 Figure 2.2 SOD1(G93A) males demonstrate earlier MU connectivity decline than SOD1(G93A) females ...... 58 Figure 2.3 SOD1(G93A) male NMJ transmission is not affected at P35 ...... 59 Figure 2.4 Normalized triceps surae wet weights at P70 and P119 ...... 62 Figure 2.5 Reduced NMJ innervation in SOD1(G93A) male mice at P70 ...... 64 Figure 2.6 Correlations of muscle contractility with MUNE...... 66 Figure 3.1 HSPB1 is overexpressed in SOD1(G93A) neuronal tissue ...... 84 Figure 3.2 Similar behavioral outcomes in AAV9-HSPB1 injected mice compared to sham mice ...... 87 Figure 3.3 Similar behavioral outcomes in AAV9-HSPB1 injected males and females compared to sham males and females ...... 88 Figure 3.4 HSPB1 overexpression does not delay motor unit decline ...... 90 Figure 3.5 HSPB1 overexpression did not protect against motor neuron death in SOD1(G93A) mice at disease endstage ...... 91 Figure 4.1 Position of sciatic nerve crush site ...... 101 Figure 4.2 Timeline for assessing motor unit reconnectivity ...... 102 Figure 4.3 Representative motor unit electrophysiological waveforms ...... 104 Figure 4.4 Muscle contractility recovery precedes MU connectivity recovery following sciatic nerve crush ...... 109 Figure 4.5 Muscle contractility recovery closely aligns with behavioral recovery following sciatic nerve crush ...... 111 Figure 4.6 Partial NMJ innervation of soleus muscle at 14dpi ...... 112

xviii

Figure 4.7 Temporal dispersion can mask MU connectivity recovery ...... 117 Figure 5.1 Exposed cisterna magna ...... 128 Figure 5.2 Longitudinal recovery timeline following PNI and CM injection ...... 129 Figure 5.3 Spinal cord transduction following AAV9 cisterna magna injection . 134 Figure 5.4 No accelerated MU connectivity recovery following delivery of AAV9 vectors ...... 135 Figure 5.5 Post-crush delivery of AAV9-HSPB1 or AAV9-SMN did not accelerate grip strength recovery to pre-injury level ...... 137 Figure 5.6 Reinnervation is complete by 60dpi ...... 140 Figure A.1 Muscle Physiology Rig Set-up ...... 192 Figure A.2 AAV9-GFP vector map ...... 193 Figure A.3 AAV9-SMN vector map ...... 194 Figure A.4 AAV9-HSPB1 vector map ...... 195

xix

CHAPTER 1: OVERVIEW OF THE NEUROMUSCULAR SYSTEM

AND THE MOTOR UNIT

α-Motor neurons and the muscle fibers they innervate constitute the motor unit (MU) and represent the final common pathway of the neuromuscular system.

α-Motor neurons are a sub-type of lower motor neurons (LMNs), with their cell bodies originating in the spinal cord and axonal projection to target muscles, distinguishing them from upper motor neurons (UMNs)—motor neurons originating in the motor cortex and brainstem. While these LMNs are ultimately responsible for the execution of motor movement, their activity can be subject to several input sources. The following section will provide an overview of the inputs to α-motor neurons and how these inputs influence MU activity.

1.1 INPUT TO LOWER MOTOR NEURONS

1.1.1 UPPER MOTOR NEURONS

Cell bodies of UMNs are located within the motor cortex, with descending axonal tracts traveling through the spinal cord to innervate their target α-motor neurons. There are several descending tracts, including: reticulospinal, rubrospinal, vestibulospinal, colliculospinal, corticobulbar and corticospinal (Bear

2007). Each of these tracts are involved with specific motor functions, but the focus here will be on the corticospinal tract (CST), in part due to it serving as the principal pathway for voluntary movement. Furthermore, downstream α-motor neurons receiving CST input can be affected in various neurodegenerative disorders, such as spinal muscular atrophy and amyotrophic lateral sclerosis, resulting in distal limb weakness (Lemon and Griffiths 2005, Welniarz, Dusart et al. 2017). Descending axons of the CST initiate from the primary motor cortex and the supplementary motor area (Maier, Armand et al. 2002). Bundles of CST axons travel through the cerebral cortex, passing via the internal capsule, towards the medulla (Figure 1.1). Before entering the spinal cord, CST axons cross at the midline, forming the pyramidal decussation, thus UMN cell bodies in the right cortical hemisphere innervate contralateral α-motor neurons (Armand

1982, Welniarz, Dusart et al. 2017). The lateral CST is one division of the CST, eventually innervating α-motor neurons responsible for controlling movement of distal limbs and digits (Kuypers 1964, de Noordhout, Rapisarda et al. 1999, Bear

2007, Welniarz, Dusart et al. 2017). The firing rate of these α-motor neurons is regulated by glutamate release from these lateral CST axon terminals (Conway,

Halliday et al. 1995, Stifani 2014, Welniarz, Dusart et al. 2017). Glutamate released from lateral CST axon terminals increases the initiation of action potentials, thereby increasing the firing rate of the neuron which, as discussed below, can impact the type of tasks the MU can perform.

Figure 1.1 Descending pathway of the corticospinal tract

Upper motor neurons originating in the motor cortex extend axons from the cortex, and pass through the internal capsule towards the medulla. Between 75%-90% of the axon bundles cross at the midline upon reaching the caudal medulla forming the pyramidal decussation (Schreyer and Jones 1982, Joosten, Schuitman et al. 1992). This results in right hemisphere upper motor neurons innervating contralateral lower motor neurons in the spinal cord. The corticospinal tract has two major divisions, the lateral corticospinal tract and the anterior corticospinal tract. The lateral corticospinal tract, shown above, innervates motor neurons that control movement of the limbs and digits. Image modified from Bear 2017.

1.1.2 SENSORY NEURONS (DIRECT—TYPE IA)

In addition to the lateral CST, α-motor neurons receive input from sensory neurons. The primary sensory neurons that innervate α-motor neurons are type

Ia sensory neurons—bipolar neurons with projections to muscles and α-motor neurons and a cell body located in the dorsal root ganglia (DRG) (Bear 2007).

Type Ia sensory neurons are large diameter neurons that provide proprioceptive feedback regarding muscle fiber length to α-motor neurons (Mears and Frank

1997). Ia fibers innervate intrafusal muscle fibers (modified skeletal muscle fibers located within muscle spindles/stretch receptors) where they are excited following fiber stretching and relay the fiber length change back to the α-motor neurons that innervate the same muscle fiber (Eccles, Eccles et al. 1957, Bewick and Banks 2015). Thus, when a muscle is lengthened Ia sensory fibers undergo increased discharge resulting in α-motor neuron depolarization and muscle contraction (shortening). Interestingly, previous studies suggest that humans and mice with amyotrophic lateral sclerosis (Hammad, Silva et al. 2007, Pugdahl,

Fuglsang-Frederiksen et al. 2007, Sabado, Casanovas et al. 2014) or peripheral nerve injury (Alvarez, Bullinger et al. 2010) can experience a re-organization and diminished maintenance of this monosynaptic connection, limiting proprioception.

1.1.3 INTERNEURONS

Interneurons are another neuronal cell type that can innervate α-motor neurons. Input from interneurons completes a polysynaptic connection—with the interneuron acting as an intermediary between the motor neuron and another neuron (Bear 2007). The role of interneurons in regulating α-motor neuron activity can be demonstrated via their role in indirectly affecting muscle tension.

Type Ib sensory fibers, like Ia fibers, have cell bodies in the DRG and send a projection to muscle fibers and another projection to the spinal cord. The muscle

4 projection innervates the Golgi tendon organ—the boundary between muscle and tendon—which acts as a gauge for muscle tension (Bear 2007). The projection to the spinal cord innervates interneurons localized in the dorsal horn, which can either be inhibitory or excitatory towards α-motor neurons (Downes, Ashby et al.

1995). In the reverse myotatic reflex, type Ib fibers sense muscle tension and innervate inhibitory interneurons, which innervate and inhibit activity of the α- motor neurons innervating the same muscle as the type Ib fibers. The reduced α- motor neuron activity, along with the diminished muscle contractions, helps regulate muscle tension generation, aiding in fine motor skills (Bear 2007).

As mentioned above, LMNs, particularly α-motor neurons, are the final common pathway of neuronal control of movement. Their anatomy and organization with muscle fibers, discussed in the next section, is one of their defining characteristics and illustrates how susceptible they are to neurodegenerative stress.

1.2 LOWER MOTOR NEURON ANATOMY, PHYSIOLOGY, AND

ORGANIZATION WITH TARGET MUSCLE FIBERS

There are several classifications of LMNs in the spinal cord, including: branchial, visceral and somatic motor neurons (Stifani 2014). α-Motor neurons, as discussed in the previous sections, are a sub-type of somatic motor neurons—neurons responsible for voluntary control of movement (Rexed 1954).

The following sections will provide an overview of the general anatomy and physiology of α-motor neurons (or, “motor neurons”) from cell body to axon

5 terminal at the neuromuscular junction and how they are organized with muscle fibers, constituting the MU.

1.2.1 MOTOR NEURON ANATOMY AND PHYSIOLOGY

1.2.1.1 CELL BODY OVERVIEW

Motor neuron cell bodies are located in the ventral horn of the spinal cord

(lamina IX) (Rexed 1954). Early staining demonstrated a range of cell body sizes, about 60µm to 135µm (Ferrucci, Lazzeri et al. 2018); the size of the cell body is one component that can determine the type of functional output the motor neuron will have, discussed further below (Stifani 2014). Dendritic processes extended from motor neurons receive input from various sources, regulating the activity of the motor neuron—discussed above. These dendritic processes express various neurotransmitter receptors, particularly for glutamatergic and GABAergic transmission (Van Den Bosch, Van Damme et al. 2006, Ben-Ari, Gaiarsa et al.

2007, Carunchio, Mollinari et al. 2008). Depending on the pre-synaptic neurotransmitter released, the motor neuron can either be stimulated/depolarized to increase firing rate (glutamate transmission) or inhibited/polarized to depress firing activity (GABA transmission). These receptors are also thought to possibly play a role in neurodegenerative diseases, such as amyotrophic lateral sclerosis

(ALS). One proposed mechanism for ALS is glutamate excitotoxicity, where glutamate is unable to be cleared from the synapse causing the motor neuron to excessively fire, eventually producing toxic levels of intracellular Ca2+ and cellular damage (Rothstein 1995, Howland, Liu et al. 2002, Van Den Bosch, Van Damme et al. 2006, Taylor, Brown et al. 2016).The possible role of GABAergic toxicity in

6 this disease is less clear. A previous study demonstrated that GABA-receptors are increased and exhibit increased sensitivity in the motor neurons of an ALS mouse model (Carunchio, Mollinari et al. 2008). It is possible that toxic levels of

GABA at the cell body leads to increased Cl- influx following prolonged AMPA receptor activation, elevating intracellular Ca2+ and driving motor neuron cell death, which has been demonstrated previously in cultured motor neurons (Van

Damme, Callewaert et al. 2003). The cell body is also pivotal for recovery following a peripheral nerve injury. Altered gene expression—increased expression of regenerative associated genes (RAGs)—is an instrumental early regenerative response to nerve injury (Fu and Gordon 1997, Mason, Wardrope et al. 2002, Stam, MacGillavry et al. 2007, Christie and Zochodne 2013). Lastly, the cell body, specifically the boundary with the axon (“axon hillock”), is the initial start point for action potential generation, discussed in more detail below.

1.2.1.2 AXON OVERVIEW

One defining characteristic of the motor neuron is its ability to communicate with target muscle fibers as far as one meter away from its cell body. The anatomy and physiology of the motor axon is critical for this to occur.

The cytoskeleton of the motor axon is comprised of microtubules (Kevenaar and

Hoogenraad 2015), actin (Spillane, Ketschek et al. 2011, Jones, Korobova et al.

2014, Kevenaar and Hoogenraad 2015) and neurofilament (Kevenaar and

Hoogenraad 2015). In addition to structural support, the cytoskeleton of the axon provides the tracks for retrograde and anterograde axonal transport (Brown

2003). Products from the cell body—such as cytoskeletal components, Golgi

7 vesicles or mRNA—can be transported distally (anterograde transport) and signals from the distal axon—such as NGF, IL-6, and CNTF following nerve injury—can be transported towards the cell body (retrograde transport), which maintains overall health of the motor neuron (Brown 2003, Christie and

Zochodne 2013, Alami, Smith et al. 2014, Kevenaar and Hoogenraad 2015).

Furthermore, the motor axon is optimally structured to transmit nerve impulses, or action potentials, to the nerve terminal at the target muscle fiber. Their increased axonal diameter, between 6µm to 8µm, increases conduction velocity by creating less resistance for ionic current flow (Gutmann and Sanders 1943,

Rushton 1951, Cullheim 1978, Waxman and Foster 1980, Fabricius, Berthold et al. 1993). Additionally, axonal segments are myelinated by Schwann cells, which accelerates signal propagation via saltatory transduction and providing current insulation (Rasminsky and Sears 1972). Early experiments from Hodgkin and colleagues detailed the steps of an action potential in axons (Hodgkin and Huxley

1952, Hodgkin and Huxley 1952, Hodgkin, Huxley et al. 1952). An initial depolarization event overcomes the membrane voltage threshold, resulting in the opening of membrane bound voltage-gated sodium channels, leading to an influx of Na+ into the cell and further depolarization of the membrane. The voltage- gated sodium channels eventually close while voltage-gated potassium channels open, causing K+ to flow out of the neuron and repolarization of the membrane, eventually returning to its resting potential. The action potential is propagated along the motor axon, between myelin sheaths (saltatory transduction) towards the nerve terminal at the neuromuscular junction.

There are numerous neurodegenerative disorders that may arise due to disorganized or dysfunctional axons. Axonal transport dysfunction is another proposed mechanism for ALS pathogenesis, with ALS animal models exhibiting slow retrograde and anterograde transport of tubulin and neurofilament prior to neurodegeneration (Williamson and Cleveland 1999, Perlson, Jeong et al. 2009).

Peripheral nerve injury can also arguably be thought of as dysfunction of the motor axon. An event that severs the connection of the motor neuron with the target muscle fiber preventing neuronal communication can result in muscle atrophy, especially if left untreated (Fu and Gordon 1995, Fu and Gordon 1995,

Palispis and Gupta 2017, Romeo-Guitart, Fores et al. 2017). The final site of motor neuron anatomy and physiology to be discussed is the axon terminal.

1.2.1.3 AXON TERMINAL (NEUROMUSCULAR JUNCTION) OVERVIEW

Muscle fibers are the target innervation site for α-motor neurons and the neuronal-muscular interface constitutes the NMJ. The function of the NMJ is to facilitate neuron to muscle communication, culminating in muscle excitation.

Voltage-gated Ca2+ channels in the active zone of an axon terminal open as an action potential approaches from the motor axon (Uchitel, Scornik et al. 1992).

The influx of Ca2+ triggers the exocytosis of vesicles containing acetylcholine

(ACh) via Ca2+ sensing proteins and SNARE complex proteins—proteins that regulate vesicle binding at the active zone (Sudhof 2013, Plomp, Huijbers et al.

2018). The number of ACh vesicles released per action potential (quantal content) varies depending on species, typically 20 in humans and 50 in mice

(Plomp, Van Kempen et al. 1995, van der Pijl, van Putten et al. 2016, Plomp,

Huijbers et al. 2018); once released, the ACh content travels through the synaptic cleft and binds to nicotinic acetylcholine receptors (AChRs) located on the muscle membrane (Albuquerque, Pereira et al. 2009). ACh binding with

AChRs on the muscle results in depolarization of the muscle membrane, or an endplate potential (EPP), triggering a muscle contraction.

1.3 EXCITATION-CONTRACTION COUPLING: CONVERSION OF

ELECTRICAL SIGNALS TO FORCE OUTPUT

1.3.1 EXCITATION-CONTRACTION COUPLING

Following signal propagation along α-motor neurons and the eventual ACh release from the nerve terminal, a series of functional events occur at the NMJ and innervated muscle which are necessary to convert an electrical stimulus to a force output, a process termed “excitation-contraction (EC) coupling”.

1.3.2 EXCITATION OF MUSCLE TO INFLUX OF CA2+ IN CYTOSOL

EC coupling begins when ACh, released from motor neurons, binds with

AChRs on the motor endplate of the muscle membrane, resulting in muscle membrane depolarization (Horowicz and Schneider 1981). The muscle action potential is propagated along, as well as within, muscle fibers via transverse tubules (t-tubules)—invaginations of the muscle membrane that directs the flow of the action potential and are positioned close to cellular calcium releasing units

(CRU) (Franzini-Armstrong and Porter 1964, Edwards, Cully et al. 2012). The

CRU is comprised of the dihydropyridine receptor (DHPR), ryanodine receptor

(RyR) and the sarcoplasmic reticulum (SR) (Schneider and Chandler 1973, Rios

10 and Brum 1987, Block, Imagawa et al. 1988, Lai, Erickson et al. 1988, Smith,

Imagawa et al. 1988, Takeshima, Nishimura et al. 1989, Franzini-Armstrong and

Jorgensen 1994). As an action potential travels through t-tubules, DHPR, a voltage sensing Ca2+ channel found in the membrane of a t-tubule, is activated by the depolarization leading to an influx of Ca2+. Influx of Ca2+ is sensed by RyR, a Ca2+ channel on the SR and closely associated with DHPR. Once RyR opens, the SR, the main repository of Ca2+ responsible for muscle fiber contraction, releases Ca2+ at a flow rate of about 200μmoles/ms (Hollingworth and Baylor

2013). The importance of Ca2 demand in muscle fibers is exemplified via the change in cytoplasmic Ca2+ concentration—going from a resting level of about

100nM to as high as 20μM (Hollingworth and Baylor 2013, Calderon, Bolanos et al. 2014). Cytosolic Ca2+ concentration returns to resting levels through a combination of protein buffering (Heizmann, Berchtold et al. 1982, Fuchtbauer,

Rowlerson et al. 1991), mitochondrial uptake (Rizzuto, Simpson et al. 1992,

Rogers, Picaud et al. 2007), and pumping back into the SR via the SR Ca2+ adenosine triphosphatase (SERCA) pump (MacLennan, Brandl et al. 1985,

Toyoshima and Inesi 2004). An overview of the underlying structural organization of sarcomeres is necessary to understand how Ca2+ is ultimately required for sarcomere force transmission.

1.3.3 ORGANIZATION OF SARCOMERES IN SKELETAL MUSCLE FIBERS

The organizational unit of skeletal muscle fibers is the sarcomere, which are bundled together to give the fiber a striated appearance (hence “striated muscle”). The major components of the sarcomere include the: A-band, I-band,

M-line and Z-line (Figure 1.2). The A-band, or “thick filament” contains the myosin motors that produce the necessary power for muscle contraction (Scott,

Stevens et al. 2001). There are several isoforms of myosin including: I, IIa and

IIx/IId/IIb (Staron 1997, Pette, Peuker et al. 1999). Myosin isoforms IIx/IId/IIb are analogous with each other, although humans will express IIx/IId and small mammals, such as mice, express IIb (Hilber, Galler et al. 1999). These isoforms are generally linked to the muscle function and MU classification of the fiber— slow-twitch (type I), fast-twitch fatigue resistant (type IIa), and fast-twitch fatigable

(type IIx/IId/IIb) (Burke 1967, Herbison, Jaweed et al. 1982, Sieck and Prakash

1997). The A-band converges in the center of the sarcomere at the M-line, which anchors and establishes a network of thick filaments. Thin actin filaments (“thin filament”) align closely with the thick filaments and act as the tracks for myosin motors to “walk” along (Huxley and Niedergerke 1954, Huxley and Hanson 1954,

Sweeney and Hammers 2018). Tropomyosin and the troponin complex (troponin

T, troponin I and troponin C), proteins which regulate sarcomere contraction, are interspersed among the actin filaments (Xu, Craig et al. 1999, Sweeney and

Hammers 2018). The region where actin filaments do not align with the thick filaments is the I-band, which eventually connects with the Z-disc. The Z-disc, in addition to providing an anchor for a network of actin filaments, also bundles neighboring sarcomeres together. Two titin proteins, each linking half of the sarcomere, connect the Z-disc with the M-line and provides additional sarcomere stabilization (Bang, Centner et al. 2001).

Figure 1.2 Structural organization of the sarcomere The center of the sarcomere is the M-line, which connects a network of thick filaments (myosin motors). The region of all thick filaments is the A-band. Actin filaments (“thin filaments”) provide tracks for myosin motors to “walk” across, resulting in sarcomere contraction/shortening. The Z-line (“Z-disc”) anchors and stabilizes the thin filaments (I-band) as well as acts as a connector of multiple sarcomeres. Image via England, 2013 (England and Loughna 2013).

1.3.4 SARCOMERE FORCE TRANSMISSION

Contraction of sarcomeres is the underlying process driving force production in muscles. During the resting state of the sarcomere, myosin binding to actin filament is regulated by the troponin complex and tropomyosin (Xu, Craig et al. 1999, Gomes, Potter et al. 2002, Sweeney and Hammers 2018). Strands of tropomyosin covering seven actin subunits form a continuous network around the actin filament by binding within the curved loops of the filament. A troponin complex, comprised of troponin T (TnT), troponin I (TnI) and troponin C (TnC), binds to each strand of tropomyosin (Staprans, Takahashi et al. 1972, Gomes,

Potter et al. 2002). The troponin complex binds to tropomyosin via TnT and TnI,

13 the latter of which also binds with an actin subunit, with TnC functioning as a

Ca2+ sensor (Xu, Craig et al. 1999). Initiation of sarcomere contraction occurs when cytosolic Ca2+, from the SR, binds with TnC. The regulatory complex unbinds from actin filaments, permitting myosin-actin binding, or cross bridges.

Actin binding results in a conformational change in myosin, which releases phosphate and adenosine diphosphate generating a power-stroke—myosin bending towards the center of the sarcomere, shortening the I-band (Huxley and

Niedergerke 1954, Huxley and Hanson 1954, Hynes, Block et al. 1987). The process is complete when a new ATP molecule binds with myosin, inducing a conformational change which breaks the cross bridge connection and myosin resets to perform the cycle again (Lorand 1953). Contraction of sarcomeres following a single stimulation results in a muscle twitch. Following a twitch, cytosolic Ca2+ is reduced and the sarcomere eventually returns to a resting state

(Sweeney and Hammers 2018). If successive stimulations from motor neurons close enough in succession—before the relaxation of muscle fibers—twitch responses can summate, producing a tetanic response. In this manner, muscle contractile force is modulated by the firing rates of motor neurons.

Motor neurons and the muscle fibers they innervate form a larger whole— the MU. The following section will detail MU organization and function.

1.4 MOTOR UNIT ORGANIZATION AND FUNCTION

Motor neuron firing activity and muscle contractility are characteristics of the MU (Figure 1.3). It is the final pathway responsible for carrying out motor

14 tasks. The following section will detail MU organization and function, illustrating its importance in the neuromuscular system.

1.4.1 MOTOR UNIT ORGANIZATION

The underlying components of the MU, a motor neuron and muscle fibers, have been discussed in the previous sections. However, there are further types of the MU, particularly with respect to classification and activity. Historically motor units (and by extension muscle fibers) can be classified as either slow (S) or fast

(F) twitch—with fast-twitch muscle fibers further sub-dividing to fast-fatigable (FF) and fast-fatigue resistant (FR) (Burke 1967, Clamann 1993). This classification corresponds to the firing rate of motor neurons (Clamann 1993). Generally, S type motor units have low, slow contraction intensity but can be active for extended periods before they fatigue—most motor units that are required for posture are S type (Burke, Levine et al. 1973, Clamann 1993, Stifani 2014). FR motor units can be thought of as an intermediate type; FR units have a high and fast contraction intensity but are moderately sensitive to fatigue (typically fatiguing within minutes of activation) (Burke 1967). FR MU activity is typically involved in normal, low-impact voluntary movements (Stifani 2014). FF motor units typically produce the largest and fastest muscle contraction intensity, but as a result are prone to rapid fatigue (Burke 1967). Typically, FF MU activity occurs to meet the challenges of a very strenuous, intense activity (Stifani 2014).

It is worth noting that MU type is not necessarily static—MU type conversion is a well-established phenomenon (Buller, Eccles et al. 1960,

Salmons and Vrbova 1969, Bagust, Lewis et al. 1981, Gordon, Stein et al. 1986).

In a classic study by Buller and colleagues, sectioning the nerves innervating cat gastrocnemius and soleus muscles then switching their innervation targets (i.e., cross reinnervating the muscles) resulted in the normally fast-twitch gastrocnemius muscle becoming slow-twitch and the slow-twitch soleus muscle becoming fast-twitch (Buller, Eccles et al. 1960). The underlying molecular events that drive this conversion are still unclear, but previous work has demonstrated that the firing activity of motor neurons is a possible driving factor

(Salmons and Vrbova 1969). Rabbit peroneal nerves, supplying the tibialis anterior and extensor digitorum longus muscles (fast types), were chronically stimulated at 10Hz (an artificially slow frequency) and by the end of the study had become slower. This conversion is also frequently observed following collateral sprouting—an event that occurs following nerve denervation, where a motor axon terminal branches to innervate a denervated muscle fiber—can result in this conversion (Kraft 2007).

Figure 1.3 Representative schematic of the motor unit

α-Motor neurons, originating in the spinal cord, project axons out of the ventral root and extend towards target muscle fibers. A single α-motor neuron can innervate multiple muscle fibers via terminal axonal sprouting. Two MUs are represented in the above figure.

1.4.2 MOTOR UNIT REGULATION OF FORCE GENERATION

Motor units are essential for carrying out the motor demands of the central nervous system. Controlling the level of force transduction is a critical function of the MU—it ensures a motor activity, strenuous or not, is met with an appropriate amount of force. MU recruitment and firing activity (or, rate coding) are two ways force control is regulated (Enoka and Duchateau 2017). MU recruitment during a voluntary motor activity is incremental—initially small units (small motor neuron cell body, small muscle fiber innervation and generally small force production) with larger units recruited as needed (i.e., as activity becomes more strenuous)

(Henneman, Somjen et al. 1965, Zajac and Faden 1985). Using the above MU classification, recruitment order is as follows: S, FR, then FF. MU recruitment in an orderly fashion helps regulate muscle force via proportional control—the force

17 generated by a newly recruited MU is a fixed percent of current force (Clamann

1993).

The firing activity of a recruited MU corresponds with motor neuron activity

(“rate coding”) and is another process that helps regulate the force generation following muscle contraction, demonstrated by previous studies utilizing force- frequency tests (Bigland and Lippold 1954, Macefield, Fuglevand et al. 1996,

Fuglevand, Macefield et al. 1999). In a force-frequency procedure, electrical stimuli are applied to the MU in a series of incremental speeds (frequency) while measuring the corresponding contractile output (force). As MU firing rates increase the resulting muscle contractions, or twitches, are able to summate with one another since the muscle is not able to return to a relaxed state. The resulting wave is generally sigmoidal, illustrated with a representative image in

Figure 1.4. Most voluntary motor actions require a firing rate between 10 to 30 Hz

(Heckman and Enoka 2012, Enoka and Duchateau 2017); thus, the representative waveform illustrates that a minimal increase in MU firing rate can generate a larger force (Macefield, Fuglevand et al. 1996, Fuglevand, Lester et al. 2015, Enoka and Duchateau 2017).

The organization of the MU and its physiological function are susceptible to various neuropathological insults. The next section will focus on two main neuropathological insults studied in this body of work, ALS and peripheral nerve injury.

Figure 1.4 Representative force-frequency waveform

Force production in a healthy contracting muscle has a sigmoidal wave shape in a force-frequency test. A series of incremental frequencies (MU firing rates) is applied to a muscle of interest, with the resulting force (contractility) recorded. Normal voluntary functions typically require a firing frequency of about 30Hz (Enoka and Duchateau 2017), which falls in the low end of the rising phase (arrow). As a result, performing more strenuous activities, which may require more force generation, can be accomplished with only a minimal increase in MU firing rate. Image modified from Enoka, 2017.

1.5 NEUROPATHOLOGICAL INSULTS CONVERGING AT THE MOTOR UNIT

Neuropathological insults disrupt the integrity of the MU—affecting either the neuronal input (“connectivity”), the muscular output (“contractility”), or both.

Amyotrophic lateral sclerosis and peripheral nerve injury are two such insults that were utilized in this work.

1.5.1 AMYOTROPHIC LATERAL SCLEROSIS

Amyotrophic lateral sclerosis (ALS) is a lethal motor neuron disease that affects approximately 5 in 100,000 individuals (Al-Chalabi and Hardiman 2013,

Talbot 2016). Progressive death of both upper and lower motor neurons, in addition to the accompanying muscle denervation and atrophy, is a defining characteristic of ALS. These result in the more common clinical symptoms of

ALS: muscle weakness, twitching, cramping, dyspnea (difficulty breathing) and dysphagia (difficulty swallowing) (Polkey, Lyall et al. 1999, Goetz 2000,

Wijesekera and Leigh 2009, Zarei, Carr et al. 2015). ALS patients will typically succumb to the disease within 3-5 years of diagnosis, commonly from respiratory problems (Talbot 2016). Between 90%-95% of ALS cases arise sporadically

(sALS) while the remaining 5%-10% arise via familial inheritance (fALS)

(Hayashi, Homma et al. 2016, Zou, Zhou et al. 2017). Mutations in the superoxide dismutase-1 gene (SOD1) comprise approximately 20% of fALS patients (Zou, Zhou et al. 2017). The first animal model of ALS overexpresses a mutant form of human SOD1, SOD1(G93A), and is a cornerstone of ALS research and continues to be used to further our understanding of ALS and MU integrity.

1.5.1.1 SOD1(G93A) TRANSGENIC MOUSE MODEL OF ALS

Several transgenic murine models of ALS have been developed over the decades to assess and develop potential pre-clinical therapeutics. Mutations in

TDP-43, c9orf72, FUS and SOD1 have all been utilized to develop transgenic mouse lines (Van Damme, Robberecht et al. 2017). Of these, the glycine to

20 alanine mutation at site 93 in SOD1—SOD1(G93A)—is the oldest and frequently used murine model. The following sections will discuss the generation and disease progression in SOD1(G93A) mice as well as its limitations.

1.5.1.2 FUNCTION OF HEALTHY SOD1

Superoxide dismutase-1 (SOD1) is a ubiquitous protein, expressed by the

SOD1 gene on chromosome locus 21q22.1 (Rosen, Siddique et al. 1993, Perry,

Shin et al. 2010, Milani, Gagliardi et al. 2011, Kaur, McKeown et al. 2016).

Identified in 1938 by Mann and Keilin, SOD1 was originally known as hemocuprein, reflecting the copper binding function of the protein (Perry, Shin et al. 2010). SOD1 functions as an antioxidant found mostly in the cytosol and mitochondria of cells, with a primary role in converting reactive oxygen species

•-- (ROS, O2 ) to oxygen (O2) and hydrogen peroxide (H2O2) (McCord and

Fridovich 1969, Valentine, Doucette et al. 2005). This conversion reaction by

SOD1 assists other antioxidant enzymes in cells to reduce the harm of ROS, such as oxidative damage to proteins, lipids and DNA (Barber, Mead et al. 2006).

1.5.1.3 SOD1 MUTATIONS AND THE GENERATION OF AN ALS MOUSE

MODEL

Mutations of the SOD1 gene and the link with ALS patients were first characterized by Rosen and colleagues in 1993. SOD1 specific primers were designed to screen for alterations in the SOD1 gene of 18 fALS families using single-strand conformational polymorphism analysis. Thirteen fALS families demonstrated anomalies in SOD1, which when sequenced found 11 distinct base-pair change mutations corresponding to 11 amino-acid substitution

21 mutations, including a glycine to alanine substitution at site 93 (G93A) (Rosen,

Siddique et al. 1993). Since these first mutations were characterized, over 180 additional mutations in SOD1 have been linked with fALS patients (Hayashi,

Homma et al. 2016). Mutations in the SOD1 gene account for approximately 20% of fALS cases (Katsuno, Tanaka et al. 2012, Hayashi, Homma et al. 2016).

Despite the wide-ranging mutations characterized in SOD1 and their links to

ALS, it is the SOD1(G93A) mutation that has been the primary tool in ALS research—owing in part to this being the earliest ALS transgenic mouse model.

In 1994 transgenic mice ubiquitously overexpressing approximately 25 copies of human SOD1(G93A) under the human SOD1 promoter were developed and first characterized (Gurney 1994). These mice appeared normal up until postnatal days 90 to 120, wherein hindlimb tremors first presented. Disease would then progress rapidly, with hindlimb paralysis and death between 4-5 months; significant motor neuron death was also reported at disease endstage.

1.5.1.4 DISEASE PROGRESSION IN SOD1(G93A) MICE

As behavioral, physiological, and pathological assays for neurodegeneration have been developed and refined, so too has our understanding of ALS disease progression in SOD1(G93A) mice since its first characterization by Gurney and colleagues in 1994. Gurney and colleagues first demonstrated the onset of behavioral deficits using presence/absence of hindlimb tremors (when mice were suspended by their tails) as well as stride length (with disease-presenting transgenic mice demonstrating shortened stride length) (Gurney 1994). Recent work has applied more functionally relevant

22 behavioral assays, including rotarod, hanging wire and treadmill performances

(Olivan, Calvo et al. 2015). Interestingly, the onset of significant motor performance decline varied depending on which assay was used. Relative to wildtype, age-matched littermates, SOD1(G93A) mice demonstrated significant decline in the hanging wire assay at P82, and a significant decline in rotarod performance at P89. Treadmill performance, on the other hand, did not significantly decline until P116 (Olivan, Calvo et al. 2015). These comparative behavioral analyses not only further refine the disease progression in

SOD1(G93A) mice, they also demonstrate the varying degrees of sensitivity in different behavioral assays, which must be considered when designing experiments and testing therapeutics.

In addition to the aforementioned behavioral characterizations, groups have further studied the physiological decline in SOD1(G93A) mice. An early characterization of MU reduction in SOD1(G93A) mice utilized an in situ incremental muscle contraction paradigm (Hegedus, Putman et al. 2007). In this cross-sectional paradigm, muscles from either wildtype or SOD1(G93A) mice at various disease time points were individually exposed and tendons tied to a force sensing gauge. Total contractile force of the muscle and incremental contractile steps were recorded following stimulation of the sciatic nerve, and the average incremental contractile step divided into the total contractile force to estimate the number of motor units. The gastrocnemius muscle (comprised primarily of fast- twitch motor units) was the earliest muscle to present significantly reduced MU counts, at postnatal day 40 (P40). Other predominantly fast twitch muscles

(tibialis anterior and extensor digitorum longus) demonstrated MU reduction 10 days later at P50. Interestingly, the soleus muscle (comprised primarily of slow- twitch motor units) did not exhibit reduced MU counts until P90 (Hegedus,

Putman et al. 2007). The data from Hegedus and colleagues suggest that muscle phenotype may factor into the onset of MU decline. Alternatively, it suggests that larger motor units are more susceptible, possibly owing to their larger energy demands. Abnormal mitochondria is a common pathological phenotype in

SOD1(G93A) mice, including swelling and diminished Ca2+ buffering, and may inhibit adequate energy production resulting in additional stress for larger motor units (Liu, Lillo et al. 2004, Wong and Martin 2010, Dadon-Nachum, Melamed et al. 2011).

Pathological symptoms have been found to frequently precede both behavioral and physiological decline in SOD1(G93A) mice (Gould, Buss et al.

2006, Pun, Santos et al. 2006, Martin, Liu et al. 2007, Vinsant, Mansfield et al.

2013, Vinsant, Mansfield et al. 2013). Fischer and colleagues have reported a thorough pathological assessment of SOD1(G93A) mice, encompassing the entirety of the MU. Significant denervation of the gastrocnemius, soleus and tibialis anterior muscles, about 40%, are first observed at approximately P47 and would rise to as high as about 90% by disease endstage at P120 (Fischer,

Culver et al. 2004). Ventral root quantification of motor axons (at the lumbar 4 segment) would remain normal until significantly decreasing at P80. Interestingly, lumbar motor neuron counts would not significantly decline until P100, after NMJ denervation and motor axon decline were already observed. This data suggest

24 that an axonal dieback mechanism is underlying SOD1(G93A) pathology.

Additionally, a normal behavioral phenotype despite muscle denervation suggests some other underlying process is acting to compensate for denervation, possibly compensatory collateral sprouting, wherein non-denervated motor axons sprout and innervate denervated muscle fibers (Frey, Schneider et al. 2000). The timing of muscle denervation and its relationship to MU physiological decline, however, remains unclear. An improved understanding of how these features relate during disease progression may ultimately produce viable, alternative therapies.

1.5.1.5 EFFECT OF SOD1 MUTATIONS

How the G93A and other mutations in SOD1 ultimately results in ALS is still debated despite the generation of the SOD1(G93A) mouse over 20 years ago. The disease phenotype does not seem to be the result of reduced SOD1 activity, as transgenic mice with endogenous SOD1 knocked out do not exhibit the hallmarks of ALS (Reaume, Elliott et al. 1996). However, the observation of

SOD1 aggregates that arise with the G93A mutation, and other ALS causing

SOD1 mutations, suggests toxic effects of protein aggregation (Bruijn, Becher et al. 1997, Ilieva, Polymenidou et al. 2009). It is unclear what the toxic effect of misfolded mutant SOD1 aggregation is and how it ultimately leads to ALS in both patients and in the mouse model. Proposed mechanisms of toxicity include: inhibition of proteasome machinery/clearance of damaged proteins (Hoffman,

Wilcox et al. 1996, Bruijn, Houseweart et al. 1998, Kabashi, Agar et al. 2004,

Cheroni, Marino et al. 2009), mitochondria damage (possibly through diminished

ATP production or inefficient Ca2+ buffering) (Mattiazzi, D'Aurelio et al. 2002, Liu,

Lillo et al. 2004, Brown, Sullivan et al. 2006, Damiano, Starkov et al. 2006), and disruptions in axon organization/axonal transport (Williamson and Cleveland

1999, Murakami, Nagano et al. 2001, Pun, Santos et al. 2006). It should also be noted that mutations in SOD1 can negatively impact the MU via a non-cell autonomous mechanism, i.e. SOD1 mutations and the accompanying toxic effects do not necessarily have to be confined to motor neurons for ALS pathogenesis. Mice expressing mutant SOD1 in only motor neurons (under a

Thy1.2 promoter) had a slower disease progression relative to transgenic mice ubiquitously expressing mutant SOD1 (Jaarsma, Teuling et al. 2008), suggesting disease may be exacerbated when other cell types expressed the mutation.

SOD1 mutations expressed in glial cells, like astrocytes and microglia, have also demonstrated toxic properties and increased motor neuron death in vitro (Ilieva,

Polymenidou et al. 2009, Haidet-Phillips, Hester et al. 2011, Frakes, Ferraiuolo et al. 2014, Heilman, Song et al. 2017). Selectively reducing mutant SOD1 expression in astrocytes or microglia has been demonstrated to slow disease progression in vivo (Boillee, Yamanaka et al. 2006, Yamanaka, Boillee et al.

2008, Wang, Sharma et al. 2009). The effect of mutant SOD1 expression in skeletal muscle is a point of contention, however. Specifically reducing mutant

SOD1 expression in muscle had no effect on disease onset or disease progression; however, when mutant SOD1 expression is limited to muscle, mitochondria dysfunction and muscle pathology were demonstrated (Miller, Kim et al. 2006, Dobrowolny, Aucello et al. 2008, Towne, Raoul et al. 2008). It is clear

26 that further research is necessary to refine the role mutant SOD1 plays in ALS pathogenesis and its stressing of the MU. Findings from these studies may provide broader insights into MU function as well as to what extent the MU components are involved.

1.5.2 PERIPHERAL NERVE INJURY

Peripheral nerve injury (PNI) was another neuropathological stress of the

MU utilized in this work. Unlike most injuries to central nerves, peripheral nerves are capable of mounting a regenerative response following injury. Peripheral nerve injuries have varying severity, depending on what underlying structures are damaged, but for most instances the cellular responses (both injury response and repair response) are conserved across injury types.

1.5.2.1 PERIPHERAL AXON ANATOMY

Generally, individual axonal segments are myelinated by Schwann cells, which provide trophic support to axons as well as increase conduction velocity via saltatory conduction. Individual myelinated axons are encased in the endoneurium, connective tissue (Figure 1.5). Mixed sensory and motor endoneurium-wrapped axons are in turn bundled together in the perineurium, forming sensory/motor fascicles. These fascicles are then bundled together by epineurium, forming the complete nerve architecture. Blood vessels are also contained within the fascicles and epineurium, providing support and also play a crucial role in the nerve’s response to a PNI.

Figure 1.5 Overview of peripheral axon anatomy. Peripheral nerves are comprised of epithelial bundled axons interspersed with blood vessels (Siemionow and Brzezicki 2009).

1.5.2.2 CLASSIFICATION OF PERIPHERAL NERVE INJURIES

There are two frequently used classification systems when discussing

PNI—Seddon (Seddon, Medawar et al. 1943) and Sunderland (Sunderland

1951). The Seddon system has three grades of injury: neurapraxia (compression injury), axonotmesis (crush injury) and neurotmesis (transection injury) (Seddon,

Medawar et al. 1943). The Sunderland system is comprised of five grades, I through V, with grades I and V corresponding to neurapraxia and neurotmesis injuries, respectively. The Sunderland system expanded upon the Seddon system by further characterizing axonotmetic injuries into: injuries that don’t damage the connective tissue (Grade II), injuries that damage the endoneurium

(Grade III) and injuries that damage the perineurium (Grade IV) (Sunderland

1951). PNI severity will range from mild (neurapraxia/Grade I) to severe

(neurotmesis/Grade V).

Understanding the organization of the peripheral nerve and a common classification system of nerve injury are two necessary components to studying peripheral nerve regeneration. The next section will discuss the actual underlying molecular events taking place following a PNI, focusing on the more severe axonotmetic and neurotmetic injuries.

1.5.2.3 RESPONSE TO PNI – INJURY SITE AND DISTAL AXON

A nerve injury that is severe enough to cut off contact between an α-motor neuron and its muscle fibers results in the systematic breakdown of the distal axon segment and the injury site, termed “Wallerian degeneration”. The sequence of events leading to Wallerian degeneration occurs within 12 to 48 hours following nerve injury and is organized into three distinct phases: acute degeneration, a latency period and fragmentation (Wang, Medress et al. 2012).

During the acute degeneration phase the distal axon segment undergoes a rapid cytoskeletal die-back, driven by the Ca2+-dependent protease calpain, and the end swells from accumulating retrograde and anterograde vesicle transport

(George and Griffin 1994, Wang, Medress et al. 2012). The latency period is marked by retention of basic physiological function, such that distal segments of

α-motor neurons are still capable of transmitting signals to innervated muscle fibers for several hours, observable via electrophysiological readouts (Kraft 2007,

Wang, Medress et al. 2012). The precise mechanism driving the shift from the latent phase to the fragmentation phase remains uncertain, but recent work

29 suggests either a Ca2+-dependent degeneration pathway or axonal support failure arising from diminished neurotrophic support (Miller, Press et al. 2009,

Gerdts, Sasaki et al. 2011). Once fragmentation of the distal segment has been initiated, complete degeneration occurs after 1 to 2 hours (Sievers, Platt et al.

2003).

1.5.2.4 RESPONSE TO PNI – NEUROMUSCULAR JUNCTION/MUSCLE

Muscle fiber health is dependent on connections with motor axons at the

NMJ. One important way MNs maintain overall muscle fiber health is by providing trophic support, such as agrin, which clusters acetylcholine receptors (AChRs) close together in the synaptic cleft (Christie and Zochodne 2013, Gordon 2016).

A severe PNI cuts off this trophic support, and eventually leads to AChR dispersion, weakness and muscle atrophy. While not permanent in acute instances of denervation, chronic denervation eventually leads to permanent muscle atrophy (Fu and Gordon 1995, Wood, Kemp et al. 2011, Christie and

Zochodne 2013, Gordon 2016). The disconnection at the NMJ and the accompanying loss of transmission appears to closely align with the latency period of degeneration observed in the distal axon segment—motor endplate activity is preserved upwards of 10 hours following injury (Miledi and Slater

1970). Eventually, as the growth cone reforms and grows toward the target muscle fibers, the NMJs can re-establish at target muscle fibers.

1.5.2.5 PERIPHERAL NERVE RECOVERY

1.5.2.5.1 SCHWANN CELL RESPONSE AT THE DISTAL AXON SEGMENT

Schwann cells are the primary glial cells of the PNS, myelinating axons and providing trophic support. These cells are capable of transitioning from a

“myelinating state” to a “regenerative state” during nerve regeneration (Gordon

2016). The change from a myelinating state to a regenerative promoting state is thought to occur in a two-step process of dedifferentiation and regenerative reprogramming changes (Jessen and Mirsky 2016). Dedifferentiation of myelinating Schwann cells is accomplished through the rapid down-regulation of

Krox20 (a myelin transcription factor), myelin basic protein and myelin associated glycoprotein (Jessen and Mirsky 2016). Concurrent up-regulation of pre- myelinating proteins, such as GFAP and NCAM, also occurs (Jessen and Mirsky

2016). There are three routes through which regenerative promoting Schwann cells create a permissive environment and promote axon regeneration: upregulation of neurotrophic factors, accelerating an immune response and guiding regenerating axons to their target sites. BDNF, NT3, GDNF, NGF and VEGF are just some of the neurotrophic factors that are up-regulated by Schwann cells following injury and attract the re-growing proximal axon (Jessen and Mirsky

2016). Neurotrophic expression is generally transient in nature, with peak expression ranging between 5 and 25 days post-injury (Christie and Zochodne

2013, Gordon 2016). About 5 to 7 days post injury, macrophages are also recruited from the surrounding vascular system, aided by Schwann cell expression of TNFα, Il-6, MCP-1 and MAC-2 (Reichert, Saada et al. 1994). Over

31 a period of up to three weeks, the fragmented debris of the distal axonal segment is phagocytized by a combination of Schwann cells (Reichert, Saada et al. 1994) and recruited macrophages (Beuche and Friede 1984, Gordon 2016). Lastly, non-myelinating Schwann cells will form regenerative tracks for regrowing axons,

Bands of Bünger, providing trophic support and guidance cues such as NGF

(Jessen and Mirsky 2016). In addition to their role in promoting regeneration along the axon, recent evidence suggests Schwann cells at the motor endplate, terminal Schwann cells, play a role in reinnervation at the NMJ, discussed further below (Kang, Tian et al. 2014).

1.5.2.5.2 REINNERVATION OF THE NMJ

Previous studies have demonstrated that a critical window of time exists where regenerating axons must reinnervate their target muscle fibers, otherwise synapse formation and functional recovery is drastically reduced (Fu and Gordon

1995, Fu and Gordon 1995). It is then critical for functional recovery that axonal re-growth is not inhibited. This is partly achieved by a permissive growth environment created by Schwann cells and macrophages, where the re-growing proximal nerve segment will extend towards the motor endplate to reconstitute the NMJs. Denervated NMJs remain occupied by terminal Schwann cells following injury, whereupon they extended processes and eventually guide re- growing motor axons to the innervation sites (Reynolds and Woolf 1992). These terminal Schwann cell processes acting as bridges for re-growing axons result in collateral axonal sprouting (Gordon 2016, Gordon and de Zepetnek 2016). Single

32 motor neurons are capable of innervating up to 5 times as many muscle fibers as they innervated prior to PNI (Gordon 2016).

1.6 OVERVIEW OF ASSESSING MOTOR UNIT

Measuring MU integrity during neuropathological stress is essential to assess efficacy of potential therapeutics in ALS or PNI. Numerous outcome measurements can be applied, but this work will focus on those that fall under one of the following categories: MU connectivity, muscle contractility, pathological or behavioral.

1.6.1 MU CONNECTIVITY OUTCOME MEASUREMENTS

1.6.1.1 COMPOUND MUSCLE ACTION POTENTIAL

MU connectivity measurements rely on the electrophysiological properties of the MU to assess integrity. One commonly used connectivity measurement is the compound muscle action potential (CMAP). Recording a CMAP from a muscle or group of muscles is a relatively straight forward process—a supramaximal electrical current is applied to a nerve supplying the recording muscles, stimulating the motor units to fire (Arnold, Sheth et al. 2015). Action potentials are generated by the motor units when stimulated which, as detailed above, ultimately travel to the axon terminal and NMJ to release ACh, resulting in muscle membrane depolarization (Horowicz and Schneider 1981). When all motor units are stimulated, all of the corresponding innervated muscle fibers should depolarize—which can be recorded using either needle electrodes inserted in (Chan and Hsu 1991) or looped wire electrodes over the muscle

(Arnold, Sheth et al. 2015). Needle electrodes are useful when the recording muscle is either too deep to be recorded from the surface or when less electrical noise is desired (Chan and Hsu 1991, Mallik and Weir 2005). However, since this requires inserting needles directly into the muscle there is a risk of damaging the muscle, possibly minimizing this approach in longitudinal studies of neurodegenerative disease. Surface looped wire electrodes are capable of measuring the CMAP in addition to being minimally invasive. Care must be taken when using surface looped wire electrodes to measure CMAP, however, as they can be more prone to recording various electrical noise (De Luca, Gilmore et al.

2010). Sources of noise that can potentially skew CMAP recordings can be either extrinsic (e.g., surrounding electronics or recording cable movement) or intrinsic

(e.g., poor surface contact or muscle movement) (Raez, Hussain et al. 2006, De

Luca, Gilmore et al. 2010). Noise can be mitigated by applying a low frequency filter and a high frequency filter. Low and high frequency filters can affect CMAP recording by removing electrical components that fall outside the filter ranges, while retaining the nerve firing frequency responsible for the CMAP response.

There are no standard values used for low and high frequency filters, however, a low frequency range of 0.2Hz-5kHz and a high frequency range of 30Hz-10kHz have been used reliably in previous studies (Sahenk, Galloway et al. 2010,

Arnold, Porensky et al. 2014, Arnold, Sheth et al. 2015, McGovern, Iyer et al.

2015, Sheth, Iyer et al. 2018).

Owing to its relative non-invasiveness and relative ease to perform, CMAP measurements have been applied in various neurodegenerative clinical studies,

34 including: ALS (Maathuis, Drenthen et al. 2013, Mori, Yamashita et al. 2016), PNI

(Krarup, Boeckstyns et al. 2016), and spinal muscular atrophy (Kolb, Coffey et al.

2016, Kolb, Coffey et al. 2017). In these neurodegenerative diseases muscle fibers ultimately undergo denervation as the MU is stressed. As a result of less neuronal supply to muscle fibers the total number of muscle fibers that undergo depolarization following nerve stimulation is reduced corresponding to diminished

CMAP amplitude. CMAP has also been used in previous pre-clinical studies to assess MU integrity (Sahenk, Galloway et al. 2010, Srivastava, Renusch et al.

2012, Sheth, Iyer et al. 2018). The sensitivity of CMAP as a tool to assess MU integrity was demonstrated in a previous transgenic animal model from our group, where a significantly reduced CMAP of the triceps surae muscle was demonstrated at 12 months old despite no corresponding behavioral decline

(Srivastava, Renusch et al. 2012).

1.6.1.2 MOTOR UNIT NUMBER ESTIMATE

Estimating the total number of motor units is another useful tool to assess the innervation of target muscles. Motor unit number estimate (MUNE) is another assay to assess MU integrity and was originally reported as a modification of

CMAP (McComas, Fawcett et al. 1971). MUNE can be calculated using various different approaches although the two most frequently used methods are incremental nerve stimulation or multipoint stimulation (Shefner, Cudkowicz et al.

2002, Shefner, Cudkowicz et al. 2006, Gooch, Doherty et al. 2014, Arnold, Sheth et al. 2015). The incremental nerve stimulation method utilizes MU stimulation of single motor units to estimate the average motor unit potential size. A series of

35 incremental, submaximal stimuli are applied over a nerve supplying a muscle of interest and assumes that each electrical response is the result of a single MU or the addition of another MU (Gooch, Doherty et al. 2014). The average of 10 quantal steps can then be used to determine the average size of a single motor unit potential (SMUP). Then the average single motor unit size is divided into the

CMAP amplitude to calculate the MUNE of the muscle or muscles being tested.

The incremental nerve stimulation method is the most frequently used strategy to estimating the number of motor units, in part because it is relatively simple and quick to perform in experienced hands (Gooch, Doherty et al. 2014). However, a drawback of this approach is the susceptibility of MU selection bias by way of alternation—multiple MUs have similar activation thresholds but different potential sizes resulting in either of these MUs firing individually or together on different occasions (McComas, Fawcett et al. 1971).

An alternative strategy to the incremental nerve stimulation approach is multipoint stimulation. The multipoint stimulation approach approximates SMUP by applying a very low stimulation and recording single all-or-none electrical responses that are present without fractionation. This process is repeated at different locations along the nerve, after which the responses are averaged to estimate the SMUP. Alternation is thus minimized with multipoint stimulation and like incremental stimulation, multipoint stimulation can be performed rapidly by an experienced user. However, the stimulated nerve must be long enough to get the appropriate number of sampling sites—an issue that can hinder its applicability in

36 pre-clinical mouse studies (Shefner, Cudkowicz et al. 2002, Gooch, Doherty et al.

2014).

MUNE has primarily been applied in clinical studies involving motor unit degeneration, such as ALS (Shefner, Cudkowicz et al. 2002, Shefner, Cudkowicz et al. 2006, Shefner, Watson et al. 2011, Shefner, Liu et al. 2016) and spinal muscular atrophy, (Swoboda, Scott et al. 2009, Gawel, Kostera-Pruszczyk et al.

2015). Several clinical studies have performed MUNE measures to gain insights pathologically into longitudinal disease progression, including the timing and rate of MU loss (Hansen and Ballantyne 1978, Carleton and Brown 1979, Dantes and

McComas 1991, Yuen and Olney 1997, Armon and Brandstater 1999). An example of longitudinally measuring MUNE to track the rate of MU loss in disease was demonstrated by Dantes and McComas. Patient motor units had fallen ~90% after the initial screening, followed with a rapid decline in MUNE over

1 year then proceeding to slow over the next 18 months (Dantes and McComas

1991). MUNE can also be applied to assess the preservation of motor units following treatment with a potential therapeutic (Gooch and Shefner 2004,

Gooch, Doherty et al. 2014, Gooch 2017).

CMAP and MUNE are invaluable pre-clinical and clinical tools to track motor unit integrity, in part owing to their ability to make repeated measurements.

However, they do assess muscle contractions following nerve stimulation.

Methods that directly measure muscle contractility are required to measure this activity.

1.6.2 MUSCLE CONTRACTILITY MEASUREMENTS

1.6.1.3 MUSCLE CONTRACTION

Muscle contractility outcome measurements are another approach to assess MU functional integrity. Like CMAP and MUNE, muscle contraction measurements can be performed to assess the innervation status of the MU

(Navarro 2016). However, unlike the aforementioned outcome measurements described in the previous sections, information obtained from muscle contractions describe the functional force output of the MU. Muscle contractile properties have been applied in a wide range of pre-clinical studies, including muscle injury (Rathbone, Wenke et al. 2003), aging (Sheth, Iyer et al. 2018), muscle dystrophy (Capogrosso, Mantuano et al. 2017) and ALS (Hegedus,

Putman et al. 2007, Hegedus, Putman et al. 2008). Muscle contraction measurements can be sub-divided into twitch force and tetanic force. Twitch force is generated by a contracting muscle following a single stimulation, while tetanic force is generated from step-wise summation of twitches in response to rapid, successive stimulations (Sweeney and Hammers 2018).

Pre-clinical measurements of twitch and tetanic contraction can be performed using in situ, in vitro or in vivo methodologies (Mintz, Passipieri et al.

2016, Navarro 2016). An in vitro setup involves completely removing the muscle from the animal and directly stimulating the sample to measure contractile properties. This can provide researchers details of muscle health independent of neuronal input. An in situ approach is similar to the in vitro setup, in respect to its ability to isolate individual muscles, however in situ muscle contractile properties

38 are obtained following electrical stimulation of nerves as innervation is still intact

(Mintz, Passipieri et al. 2016). While both in vitro and in situ applications are suitable for assessing the contractile properties of isolated muscles, they are also highly invasive which is a limitation to their application in longitudinal studies

(Navarro 2016).

An alternative strategy for assessing muscle contraction, and thus MU integrity, is to apply an in vivo paradigm. Twitch and tetanic contractile properties can still be recorded following nerve stimulation, and unlike the aforementioned approaches an in vivo approach is minimally invasive—only requiring electrode needles to provide neuronal stimulation (Mintz, Passipieri et al. 2016). A strength of this approach is being well-suited for characterizing progressive muscle weakness in neurodegenerative disorders.

1.6.3 PATHOLOGICAL ASSESSMENT OF THE MU

1.6.3.1 NEUROMUSCULAR JUNCTION REINNERVATION

Quantification of NMJ innervation can also refine our understanding of the events underlying MU integrity in a neuropathological state.

Immunohistochemical stained pre-synaptic markers, like synaptophysin or neurofilament-200, and post-synaptic markers, like α-bungarotoxin, can be quantified to indicate a connected motor unit (Fischer, Culver et al. 2004,

Sakuma, Gorski et al. 2016). Neuropathological stress that ultimately leads to

MU denervation can be visualized microscopically when the pre- and post- synaptic markers are no longer co-labeled. Therapeutics that either protect MU

39 from degeneration or enhances reconnectivity would be expected to stabilize or increase NMJ reinnervation.

1.6.4 BEHAVIORAL RECOVERY

1.6.4.1 GRIP STRENGTH

Depending on the disease model either forelimb or hindlimb grip strength can be assessed by positioning the animal to grasp a grid connected to a grip meter, pulling away produces a tension readout (Krishnan, Vannuvel et al. 2008,

Sharp, Akbar et al. 2008, Wang, Sorenson et al. 2008, Srivastava, Renusch et al.

2012, Foust, Salazar et al. 2013). In nerve injury models, grip strength can be further refined to compare ipsilateral side with contralateral side (Wang,

Sorenson et al. 2008). The grip strength outcome measure is non-invasive, easy to perform and can be repeatedly measured over time to track muscle reinnervation, and are thus useful aids to test therapeutic efficacy of a treatment.

1.6.4.2 ROTAROD

Another commonly behavioral function assay that may be performed to study neuromuscular health is the rotarod assay. In this assay animal subjects walk on a slowly accelerating, rotating cylinder (Hamm, Pike et al. 1994, Wood,

Kemp et al. 2011). Neuromuscular health can then be assessed via the recorded time it takes for the animal to fall off the cylinder, or “latency to fall”; a short latency to fall thus corresponds with neuromuscular weakness (Hamm, Pike et al.

1994). Rotarod performance has been frequently used in numerous models where motor performance is affected, including: ALS (Weydt, Hong et al. 2003,

Zhang, Narayanan et al. 2003, Foust, Salazar et al. 2013), Alzheimer’s disease

(Jolivalt, Lee et al. 2008, Sterniczuk, Antle et al. 2010), and Huntington’s disease

(Hockly, Cordery et al. 2002, Harper, Staber et al. 2005).While grip strength and rotarod functional assays are quick and easy to perform, they are not without limitations—primarily they do not directly measure only MU (Navarro 2016).

1.7 APPLICATION OF MU CONNECTIVITY AND CONTRACTILITY IN

STRESSED STATES

There are two direct methods to measure MU integrity— electrophysiologically (“connectivity”) and muscle physiologically (“contractility”).

Connectivity measurements (e.g., CMAP and MUNE) detail the electrical generation of a muscle following nerve stimulation while contractility measurements (e.g., twitch and tetanic) detail the force generation of a muscle following nerve stimulation. Exactly when in pathophysiological stress of the MU muscle electrical excitation declines relative to muscle force generation decline is unknown. It is possible that MU connectivity is affected after muscle force production, which we can interpret to mean muscle weakness is occurring despite healthy neuronal input. Alternatively, MU connectivity decline may precede muscle force production, suggesting muscle contractility can somehow compensate for decreased neuronal input. Finally, both connectivity and contractility could be affected concurrently, which suggests muscle strength and neuronal input are directly related. We hypothesized that the timing of when connectivity is affected relative to when contractility is affected would be dependent the severity of pathophysiological stress. To address this, we utilized

41 animal models of ALS (lethal neurodegenerative disease with motor neuron death) and a mild PNI (non-lethal, absence of motor neuron death and regeneration eventually occurs). Defining this relationship will help us understand the functional status of the neuromuscular system and may ultimately lead to the development of novel therapeutic strategies. Additionally, the MU response to a pathological stressor is a dynamic process, with muscles losing innervation and neighboring axons sprouting to reinnervate. Therefore, studies that can assess

MU integrity longitudinally, with minimal invasiveness, are arguably more powerful than cross-sectional studies (those which utilize different cohorts at single points in time). Longitudinal studies help determine optimal treatment windows during MU denervation or regeneration, provide possible biomarkers and help shape the direction of future pre-clinical studies by detailing underlying mechanisms. We further apply similar paradigm to assess potential regenerative therapeutics in ALS and PNI.

CHAPTER 2: MUSCLE DYSFUNCTION PRECEDES LOSS OF

MOTOR UNIT CONNECTIVITY IN SOD1(G93A) MICE

2.1 STATEMENT OF CONTRIBUTION

Compound muscle action potential and motor unit number estimate as well as muscle physiological measurements were performed by me. Behavioral measurements were performed by Alex Crum and Anthony Reynolds. David

Arnold performed single fiber electromyography. Pathological imaging was performed both by me and Deepti Chugh, with NMJ quantification performed by me. Modeled data of electrophysiology and muscle physiology results was performed by David Kline and Marilly Palettas. I carried out writing and figure creation, with revisions and edits from: David Kline, David Arnold and Stephen

Kolb.

2.2 INTRODUCTION

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disorder affecting approximately 5 in 100,000 individuals (Al-Chalabi and Hardiman 2013,

Talbot 2016). The disease is characterized by progressive loss of upper and lower motor neurons resulting in weakness, muscle denervation and atrophy, and

43 eventual death, typically within 3-5 years of diagnosis. About 10% of ALS patients have a family history of the disease (fALS) while the remaining 90% cases arise spontaneously (sALS) (Hayashi, Homma et al. 2016, Zou, Zhou et al.

2017).

Electrophysiological measures, including compound muscle action potential (CMAP) and motor unit number estimate (MUNE), allow in vivo assessment of the motor unit (MU, a motor neuron and the myofibers it innervates) function and connectivity. Such measurements have been applied in a number of natural history studies of ALS in humans to provide important insight into the interrelationship of muscle function and MU connectivity (Kelly,

Thibodeau et al. 1990, Dantes and McComas 1991, Felice 1995, Yuen and

Olney 1997, Arasaki, Kato et al. 2002). A powerful aspect of CMAP and MUNE measurements is that the functional status of the neuromuscular system can be tracked longitudinally to identify disease onset, severity, and rate of progression.

Furthermore, MUNE techniques can provide an assessment of the neuromuscular system in response to a therapeutic, from the standpoint of preservation or regeneration of motor unit number, as well as increased output from individual motor units (i.e., improved collateral reinnervation) (Gooch and

Shefner 2004, Gooch, Doherty et al. 2014, Gooch 2017). As such, a number of clinical studies have leveraged electrophysiological measures to gain pathophysiological insight into longitudinal disease progression, including the timing and rate of MU loss, the ability of individual motor units to perform collateral sprouting, and the relationship of MU degeneration to the onset of

44 muscle weakness (Hansen and Ballantyne 1978, Carleton and Brown 1979,

Dantes and McComas 1991, Yuen and Olney 1997, Armon and Brandstater

1999). Dantes and McComas demonstrated ~90% decline in motor units during initial clinical screening, followed with a rapid decline in MUNE over 1 year— nearly halving every 6 months—before decline slowed over the next 18 months

(Dantes and McComas 1991). This study also demonstrated a corresponding increase in the single motor unit potential (SMUP), consistent with reinnervation by collateral axonal sprouting of surviving motor units—probably explaining why patients might present with absent or mild muscle weakness despite such a dramatic reduction in MUNE. Another early clinical study measured CMAP and

MUNE in addition to functional grip strength to assess the abductor digiti minimi of 10 ALS patients over the course of 6 months (Yuen and Olney 1997). The authors reported that patients had a significant decline in MUNE at 3 and 6 months post initial screening, while CMAP and grip strength had not declined, providing further evidence of the compensatory ability of remaining motor neurons via collateral sprouting. At 3 and 6 months, single fiber EMG recordings in patients demonstrated increases in the mean muscle fiber density (a measure of grouped reinnervation) of the abductor digiti minimi, again indicating compensatory MU sprouting (Yuen and Olney 1997). Clinical studies that track the natural history of MU integrity during ALS, like the aforementioned, provide several useful insights, including potential treatment windows and underlying dynamic biological processes that could not otherwise be studied (Simon, Turner

45 et al. 2014, Rutkove 2015, Shefner, Liu et al. 2016, van Eijk, Eijkemans et al.

2018).

Mutations in the Cu/Zn superoxide dismutase (SOD1) gene account for roughly 20% of fALS cases. Overexpression of about 25 copies of human SOD1 with a G93A mutation was applied to develop the first mouse model of ALS in

1994 (Gurney 1994, Zou, Zhou et al. 2017). In the SOD1(G93A) mouse, behavioral disease onset is first observed at approximately post-natal day 90

(P90), and is defined by hindlimb tremors when suspended by the tail (Vinsant,

Mansfield et al. 2013, Vinsant, Mansfield et al. 2013). Muscle atrophy and motor neuron death then rapidly progress until mice reach end stage around P120-

P150 (Gould, Buss et al. 2006). The SOD1(G93A) mouse model has been utilized in numerous pre-clinical studies that have led to clinical trials in humans, however, to date, only two compounds have become clinically approved in the

United States, and these treatments have modest effect on disease progression

(Gurney, Fleck et al. 1998, Ito, Wate et al. 2008).

Contractility of a muscle, including twitch torque or tetanic torque, is a physiological readout of the MU following neuronal stimulation. A number of studies have investigated neuromuscular function in the SOD1(G93A) mouse model using either electrophysiological or physiological measures, (Shefner,

Brown et al. 2001, Shefner, Cudkowicz et al. 2002, Shefner, Cudkowicz et al.

2006, Hegedus, Putman et al. 2007, Zhou, Zhao et al. 2007, Hegedus, Putman et al. 2008, Hegedus, Putman et al. 2009, Mancuso, Santos-Nogueira et al.

2011, Ngo, Baumann et al. 2012, Li, Sung et al. 2013, Dibaj, Schomburg et al.

2015, Li, Pacheck et al. 2016). Nevertheless, the relationships between physiological and electrophysiological readouts—muscle contractility and MU connectivity—have not been directly investigated and are poorly understood.

Therefore, we aimed to compare the longitudinal progression of muscle dysfunction with electrophysiological measurements of MU connectivity, including

CMAP and MUNE, which we have refined for use in other mouse models of neuropathy, spinal muscular atrophy, and aging (Srivastava, Renusch et al.

2012, Arnold, Porensky et al. 2014, Arnold, Sheth et al. 2015, Arnold, McGovern et al. 2016, Sheth, Iyer et al. 2018). Collateral sprouting of motor axons can compensate for denervation, however, there is a physiological limit to the number of new muscle fibers a nerve can reinnervate, typically 3-5x (Gordon 2016).

Thus, in severe diseases with motor neuron death compensation will eventually be lost. We hypothesize that MU connectivity decline will precede muscle contractility decline in an animal model of ALS. Interestingly, loss of muscle contractility was the initial phenotypic feature of disease onset. Furthermore, loss of muscle contractility and MU connectivity decline was sexually dimorphic with male SOD1(G93A) mice demonstrating much earlier deficits as compared with females.

2.3 MATERIALS AND METHODS

2.3.1 ANIMALS

All procedures were performed in accordance with NIH Guidelines and approved by the Institutional Animal Care and Use Committee (IACUC) of the

Ohio State University. A total of 54 mice were used in this study for reliability

47 testing and for the longitudinal analysis of disease progression in the

SOD1(G93A) mouse model. Adult wildtype and (Stock#: 002726) SOD1(G93A) mice were obtained from Jackson Laboratories (Bar Harbor, ME). Adult wildtype mice obtained from Taconic Biosciences (Albany, NY) were used for intrarater reliability testing. All studies were performed with blinded raters.

2.3.2 BEHAVIORAL ASSESSMENTS

Mouse body weight was recorded prior to performing grip strength and rotarod assessments (Song, Miranda et al. 2016, Frakes, Braun et al. 2017). For grip strength, the average right hindlimb grip strength, measured in grams, was calculated from four measurements per assessment using a standard grip meter

(DEFII-002, Chatillon, Largo, FL, USA). Mice were positioned to allow only the right hindpaw to grasp the grid and were pulled towards the evaluator for the length of the grip meter (Miller, Kim et al. 2006). Attempts that were ≥ ± 10g different than the other attempts were discarded and re-performed. The average of three motor coordination tests was measured using an accelerating rotarod

(LE8205, Panlab Harvard Apparatus) starting at 5RPM. Trials were stopped if

120s had passed without a fall.

2.3.3 MOUSE ANESTHESIA AND ANIMAL PREPARATION

Mice were anesthetized (isoflurane inhalation, 1.5-3%) during electrophysiological and muscle physiological recordings. Lubricant ointment was applied to the eyes to prevent corneal drying. All measurements were performed on the right hindlimb which was shaved with electric clippers prior to studies.

Electrophysiological procedures were performed on a heated platform set at

37°C (World Precision Instruments, Sarasota, FL). During muscle contractility procedures, a warm water bath HTP-1500 Heat Therapy Pump set at 37°C was used to maintain temperature of the muscle testing apparatus stage (Androit

Medical Systems, Loudon, TN). Procedures under anesthesia lasted no longer than 30 minutes and were typically less than 20 minutes.

2.3.4 ELECTROPHYSIOLOGY

CMAP and MUNE were recorded as previously described (Arnold,

Porensky et al. 2014, Arnold, Sheth et al. 2015, Arnold, McGovern et al. 2016).

Briefly, an active ring electrode was placed superficially over the right triceps surae and a reference ring electrode was placed superficially over the metatarsals of the right hindpaw (Alpine Biomed, Skovlunde, Denmark). A ground electrode was placed on the tail (Carefusion, Middleton, WI). The sciatic nerve was stimulated (0.1ms pulse, 1-10mA intensity) using two insulated monopolar needles (28G) (Teca, Oxford Instruments Medical, NY). CMAP amplitudes were recorded following maximal stimulation. Baseline-to-peak amplitudes were used for comparison of CMAP amplitudes and peak-to-peak amplitudes were used for calculation of MUNE. The single motor unit potential

(SMUP) was calculated by taking the average of 10 incremental submaximal responses. MUNE was calculated by dividing the average SMUP amplitude

(peak-to-peak) into the maximum peak-to-peak CMAP amplitude.

2.3.5 MUSCLE CONTRACTILITY

Following electrophysiological recordings, mice underwent a triceps surae plantarflexion torque assessment using an in vivo muscle contractility apparatus

(Model 1300A, Aurora Scientific Inc, Canada) (Appendix A.1) as previously detailed (Sheth, Iyer et al. 2018). Briefly, the right hindpaw was taped to the force sensor and positioned at 90°. Then, the hindlimb was extended to position the knee in the locking position and securely locked in place with compression of the femoral condyles. Two disposable monopolar EMG electrodes were inserted near the tibial nerve, just posterior to the knee (Natus Neurology, Inc, Middleton,

WI). Maximum plantarflexion twitch was recorded following a single, supramaximal stimulation (0.2ms square wave pulse). Maximum tetanic contraction was assessed following a train of supramaximal square wave stimulations at 0.2ms duration delivered at 125Hz stimulation frequency.

2.3.6 NMJ IMAGING AND QUANTIFICATION

To investigate the morphological correlates of loss of muscle contractility and reduced MU connectivity, the soleus muscle was collected from a separate cohort of SOD1(G93A) and wildtype mice for endpoint studies at P70 and fixed in

4% paraformaldehyde (PFA) at room temperature (RT) for 30min (Chai, Vukovic et al. 2011, Cheng, Morsch et al. 2013). Muscles were teased into fibers using size 55 forceps (Fine Science Tools). Teased muscle fibers were incubated in blocking buffer (10% goat serum/4% BSA/3% triton-X 100/PBS) at RT for 2hr. An overnight (O/N) primary antibody (α-NF-200, Abcam, Ab72996, [1:5,000]) incubation at 4°C was then performed. Samples then underwent 3 10min washes with PBS before receiving a 2hr incubation with secondary antibody (Alex 594 goat α-Chicken, Life Technologies, A11042, [1:1,000]) and α-Bungarotoxin-488

(Life Technologies, B13422, [1:1,000]) at RT. Samples then underwent three

10min washes with PBS at RT. Stained muscle samples were mounted onto

Superfrost positively charged glass slide (Fisher Scientific) and were sealed using Fluoromount-G (Southern Biotech). Samples were imaged at 20x and 40x magnification using a Leica confocal microscope (Leica DM IRE2) with Leica software (version 2.1). Images were viewed in FIJI (LOCI, University of

Wisconsin-Madison) to quantify NMJ innervation, the co-labeling of NF-200 and

α-Bungarotoxin. 95-120 NMJs per muscle sample (per mouse) were scored as fully innervated, partially innervated or denervated.

2.3.7 INTRARATER TESTING

Intrarater reliability was assessed for electrophysiology and muscle contractility measures.16 wildtype mice (C57BL/6 x SJL/J; 7 males and 9 females) were assessed by a blinded observer (CW), with 1 day between each trial. Reliability was reported using the intraclass correlation coefficient (ICC), which was defined as poor (ICC < 0.5), moderate (0.5 < ICC < 0.75), good (0.75

< ICC < 0.9) or excellent (ICC > 0.9) (Koo and Li 2016).

2.3.8 ASSESSMENT OF LONGITUDINAL SOD1(G93A) DISEASE

PROGRESSION

Blinded raters (AC, AR) assessed body weight, hindlimb grip strength, electrophysiological and muscle contractility weekly from P35 to P119 in a total of

20 mice [5 male/5female wildtype (C57BL/6 x SJL/J) and 5 male/5 female

SOD1(G93A) (C57BL/6 x SJL/JTg)]. Mice that displayed loss of the righting reflex for 30 seconds were euthanized for tissue harvesting and muscle wet weight recording (Foust, Salazar et al. 2013). Additional SOD1(G93A) (n = 6) and

51 wildtype (n = 6) mice were sacrificed at P70, and right triceps surae muscles were collected, weighed and processed for NMJ quantification. A single evaluator, blinded to genotype, performed the longitudinal electrophysiological and muscle physiological assessments (CW).

2.3.9 STATISTICS

Intrarater variability analyses were used to assess the degree of agreement between the electrophysiology and muscle contractility measurements across all mice, over time. The R package irr (R version 3.3.2,

The R Foundation for Statistical Computing) was used to perform these analyses.

A mixed effects model was used to model the mean of each of the 6 outcome variables for the longitudinal experiments. Fixed effects were included for the two groups, SOD1 mutants and wild type, and gender. A random intercept was used to account for repeated measures within mouse. Backward selection was used starting from an initial model with up to 3-way interactions. A square root transformation was used for CMAP, SMUP and MUNE. Furthermore, a quadratic time trend was included for CMAP, MUNE, and normalized twitch torque. Group means were compared at each time point at the 0.05 level using

Holm’s Method to adjust for multiplicity within each outcome (Holm 1979). When group effects differed significantly by gender, comparisons were made within each gender. SAS 9.4 (Cary, NC) was used for the analysis. Unpaired t-test was performed using Graphpad Prism software (version 6) to compare muscle weight and NMJ quantification. Pearson’s correlation coefficients were calculated to

52 determine correlations between: muscle contractility and MU connectivity, behavioral measurements and muscle contractility, and behavioral measurements and MU connectivity. Correlation coefficient strengths were interpreted using prior guidelines: negligible (0.0

2012). Statistical significance was set at p<0.05.

2.4 RESULTS

2.4.1 INTRARATER RELIABILITY OF MUSCLE CONTRACTILITY AND MU

CONNECTIVITY MEASUREMENTS

The intrarater reliabilities of muscle contractility and MU connectivity measurements in clinical studies have been well reported (Shefner, Watson et al.

2011, Rutkove, Caress et al. 2012, Ives and Doherty 2014, Shefner, Liu et al.

2016). However, only a handful of pre-clinical studies utilizing muscle contractility and MU connectivity measurements have reported intrarater reliability results

(Kasselman, Shefner et al. 2009, Li, Staats et al. 2012). We first sought to determine the inherent reliability and reproducibility of our muscle contractility and MU connectivity assays (Table 2.1). CMAP and twitch measurements were moderately reliable with ICCs of 0.62 and 0.68, respectively, while tetanic measurements demonstrated good reliability with an ICC of 0.81.

Outcomes Intrarater ICC (Lower bound)

CMAP 0.62 (0.28)

Twitch 0.68 (0.37)

Tetanic 0.81 (0.60) Table 2.1 Intrarater reliability correlation coefficients.

Intraclass coefficients for intrarater reliability in compound muscle action potential (CMAP), twitch torque and tetanic torque.

2.4.2 DECLINE OF MUSCLE CONTRACTILITY AND LOSS OF MOTOR UNIT

CONNECTIVITY ARE EARLY FEATURES IN SOD1(G93A) MICE

Mixed effects models for longitudinal muscle contractility and MU connectivity decline were developed to control for variability within individual mice from age P35 to P119 (Figures 2.1 and 2.2) (Appendix B.1 and B.2). We compared longitudinal muscle contractility and MU connectivity in male and female mice independently, taking into account sexual differences in disease progression (Veldink, Bar et al. 2003, Heiman-Patterson, Deitch et al. 2005).

Loss of muscle contractility preceded MU connectivity decline in both male and female SOD1(G93A) mice (Figure 2.1). Normalized twitch torque in SOD1(G93A) males was reduced compared to wildtype males at the start of the study (P35)

(0.087mN-m/g [0.073 - 0.101mN-m/g] vs 0.117mN-m/g [0.103 - 0.131mN-m/g])

(Figure 2.2A). Normalized tetanic torque was also reduced at P35 in

SOD1(G93A) males compared to wildtype males (0.48mN-m/g [0.42 - 0.54mN- m/g] vs 0.59mN-m/g [0.53 - 0.65mN-m/g]) (Figure 2.1B). Normalized twitch torque declined at P63 in SOD1(G93A) female mice compared to wildtype female mice (0.099mN-m/g [0.088 - 0.111mN-m/g] vs 0.121mN-m/g [0.110 - 0.132mN- m/g]), whereas normalized tetanic torque declined by P56 in SOD1(G03A) females relative to wildtype females (0.45mN-m/g [0.41 - 0.50mN-m/g] vs

0.54mN-m/g [0.49 - 0.58mN-m/g]) (Figure 2.1A-B).

CMAP decline first occurred at P42 in SOD1(G93A) males compared to wildtype male mice (31.0mV, [95% confidence interval: 25.4 - 37.2mV] vs

44.6mV [37.8 - 51.9mV]) and at P91 in SOD1(G93A) females compared to wildtype females (25.7mV [21.1 - 30.8mV] vs 39.2mV [33.4 - 45.5mV]) (Figure

2.2A). MUNE decline in SOD1(G93A) males relative to wildtype males occurred at P49 (252 [205 - 304] vs 368 [310 - 430]) and at P77 in SOD1(G93A) females compared to wildtype females (205 [164 - 251] vs 323 [271 - 381]) (Figure 2.2B).

Interestingly, there was no observed sexual dimorphism in SMUP, with increases occurring at P70 for both SOD1(G93A) males and females relative to wildtype males and females (252.0µV [236.6 - 267.9µV] vs 218.7µV [204.3 - 233.5µV])

(Figure 2.2C). Single fiber electromyography (SFEMG) was performed to exclude the possibility that failure of NMJ transmission could be the explanation for early contractility loss in SOD1 mice. Separate cohorts of SOD1(G93A) males and wildtype males were studied with SFEMG in the gastrocnemius at P35 to assess

NMJ integrity. There was no difference in jitter between SOD1(G93A) males

(8.02±2.29µs; range: 4.01-12.27µs and wildtype males (8.38±2.45µs; range 5.07-

12.57µs) (p=0.669) (Figure 2.3). Table 2.2 summarizes and compares onset of

MU degeneration and muscle contractility decline.

Figure 2.1 SOD1(G93A) males demonstrate earlier muscle contractility decline than SOD1(G93A) females

(A) Predicted twitch torque (normalized to body mass) (mN-m/g) outcomes of wildtype (WT) male mice (blue square, n = 5), wildtype female mice (orange x, n = 5), SOD1(G93A) male mice (red triangle, n = 5) and SOD1 female mice (purple circle, n = 5). (B) Predicted tetanic torque (normalized to body mass) (mN-m/g) outcomes in wildtype and SOD1(G93A) cohorts, organized by gender. Shaded regions depict 95% confidence interval.

Figure 2.2 SOD1(G93A) males demonstrate earlier MU connectivity decline than SOD1(G93A) females

(A) Longitudinal predicted CMAP (mV) of wildtype male mice (blue square, n = Continued

5), wildtype female mice (orange x, n = 5), SOD1(G93A) male mice (red triangle, n = 5) and SOD1 female mice (purple circle, n = 5). (B) Predicted SMUP (µV) of wildtype mice (blue square, n = 10) and SOD1(G93A) mutants (red triangle, n = 10). (C) Predicted MUNE of wildtype male mice, wildtype female mice, SOD1(G93A) male mice and SOD1(G93A) female mice. Shaded regions depict 95% confidence interval. Abbreviations: compound muscle action potential (CMAP), single motor unit potential (SMUP), motor unit number estimation (MUNE), and wildtype (WT).

Figure 2.3 SOD1(G93A) male NMJ transmission is not affected at P35

Single fiber electromyography (SFEMG) was performed in a separate cohort of SOD1(G93A) males (n=3) and wildtype males (n=3) to assess NMJ functional status. Error bars denote standard deviation. p=0.669.

Post-natal Day

Mut vs 35 42 49 56 63 70 77 84 91 98 105 112 119 WT

Males 0.0079 Twitch Females - - - - 0.0427

Males 0.0119 Tetanic Females - - - 0.0326

Males - 0.04 CMAP Females ------0.018

Males - - - - 0.0244 SMUP Females - - - - 0.0244

Males - - 0.0382 MUNE Females ------0.0157

Table 2.2 Onset of muscle contractility and MU connectivity decline

Comparison of SOD1(G93A) males (n=5) versus wildtype males (n=5) (shaded) and comparison of SOD1(G93A) females (n=5) versus wildtype females (n=5) (unshaded). Dashes represent no significant difference, arrows indicate that difference was significant for duration of study. Abbreviations: wildtype (WT), compound muscle action potential (CMAP), single motor unit potential (SMUP), motor unit number estimation (MUNE). Significance set at p<0.05.

2.4.3 MUSCLE ATROPHY OCCURS AFTER LOSS OF MUSCLE

CONTRACTILITY

We assessed wet triceps surae muscle mass (normalized to body mass) to examine whether the early findings of reduced muscle contractility and loss of

MU connectivity were associated with coexistent loss of muscle mass. Separate groups of wildtype and SOD1(G93A) were sacrificed at P70, in addition to the longitudinal cohorts at P119, and their right triceps surae muscles were harvested and weighed (Figure 2.4). Despite losses of both muscle contractility and MU connectivity there were no overt differences in normalized triceps surae wet weights at P70 between wildtype and SOD1(G93A) mice (0.0082g ±

0.0008g, n = 6 vs 0.0075g ± 0.00130g, n = 6, p = 0.123) (Figure 2.4A). At P119, normalized triceps surae wet weight was reduced in SOD1(G93A) mice compared to wildtype mice (0.0037g ± 0.0009g, n = 9 vs 0.0065g ± 0.0003g, n =

10, p<0.001) (Figure 2.4A). When we organized normalized triceps surae wet weights by sex at P70, there was no difference in normalized triceps surae weight between SOD1(G93A) males relative to wildtype males (0.0069g ±

0.0016g, n = 3 vs 0.0081g ± 0.0003g, n = 3, p = 0.19) nor was there difference between SOD1(G93A) female and wildtype female (0.0083g ± 0.0006g, n = 3 vs

0.0084g ± 0.0011g, n = 3, p = 0.52) (Figure 2.4B).

At P119, normalized triceps surae weight was reduced in SOD1(G93A) males compared to wildtype males (0.0035g ± 0.0014g, n = 4 vs 0.0066g ±

0.0003g, n = 5, p<0.001) as well as in SOD1(G93A) females compared to

61 wildtype females (0.0039g ± 0.0004g, n = 5 vs 0.0064g ± 0.0002g, n = 5, p<0.001) (Figure 2.4B).

Figure 2.4 Normalized triceps surae wet weights at P70 and P119

(A) Normalized triceps surae weight at P70 (wildtype, n = 6; SOD1(G93A), n = 6) and P119 (wildtype, n = 10; SOD1(G93A), n = 9). (B) Normalized triceps surae wet weights by sex at P70 (wildtype males, n = 3; wildtype females, n = 3; SOD1(G93A) males, n = 3; SOD1(G93A) females, n = 3) and at P119 (wildtype males, n = 5; wildtype females, n = 5; SOD1(G93A) males, n = 4; SOD1(G93A) females, n = 5). Error bars denote standard deviation. n.s. = no significance, *** = p<0.001.

2.4.4 NMJ INNERVATION IS REDUCED AT P70

NMJ innervation of the soleus muscle was quantified at P70 after contractility decline and an increase in SMUP was noted in both male and female

SOD1(G93A) mice in the longitudinal cohort (Figure 2.5). There was no change

62 in the relative percentage of denervated NMJs between SOD1(G93A) and wildtype mice at P70 (25.2% ± 18.9%, n = 6 vs 10.7% ± 4.4%, n = 6, p>0.05)

(Figure 2.5B). When we compared within gender, NMJ denervation was increased in SOD1(G93A) male mice compared to wildtype male mice (49.1% ±

4.2%, n = 3 vs 81.7% ± 5.2%, n = 3, p<0.01) (Figure 2.5C). SOD1(G93A) females did not demonstrate an increase in denervation percentage compared to wildtype females (9.99% ± 9.05%, n = 3 vs 10.0% ± 3.48%, n = 3, p>0.05)

(Figure 2.5D).

Figure 2.5 Reduced NMJ innervation in SOD1(G93A) male mice at P70

(A) Representative NMJ images from wildtype (a-c) and SOD1(G93A) (d-f) teased soleus muscle fibers. Co-labeling of NF-200 (in red) and α-Bungarotoxin (in green) demonstrated in panel c, but absent in panel f. (B) Percent innervation, percent partial innervation and percent denervation in wildtype mice (n = 6) and SOD1(G93A) mice (n = 6). (C) Percent innervation, percent partial innervation and percent denervation in wildtype males (n = 3) and SOD1(G93A) males (n = 3). (D) Percent innervation, percent partial innervation and percent denervation in Continued

64 wildtype females (n = 3) and SOD1(G93A) females (n = 3). Error bars denote standard deviation. n.s. = no significance, ** = p<0.01, and *** = p<0.001. Images taken at 40x, scale bar = 25µm.

2.4.5 CORRELATIONS BETWEEN MUSCLE CONTRACTILITY, MU

CONNECTIVITY, AND BEHAVIORAL ASSESSMENTS

Correlations were analyzed between electrophysiological measures, muscle contractility, and grip strength at week 15. MUNE showed a strong positive correlation with absolute twitch torque and absolute tetanic torque

(Figure 2.6A and 2.6C), but MUNE showed no significant correlation with normalized twitch and tetanic torque (Figure 4B and 4D). CMAP demonstrated a strong positive correlation with both absolute twitch torque (r=0.84, p<0.01) and absolute tetanic torque (r=0.76, p<0.05) and moderately positive correlated with both normalized twitch torque (r=0.68, p<0.05) and normalized tetanic torque

(r=0.65, p<0.05) (Appendix B.3). Grip strength was strongly correlated with

CMAP (r=0.82, p<0.01) and MUNE (r=0.71, p<0.05) and moderately correlated with absolute twitch torque (r=0.66, p<0.05) (Appendix B.3). There was no significant correlation for grip compared with absolute tetanic torque (r=0.579, p=0.079), normalized twitch torque (r=0.513, p=0.130) and normalized tetanic torque (r=0.516, p=0.127) (Appendix B.3).

Figure 2.6 Correlations of muscle contractility with MUNE.

(A-B) Correlation of Motor Unit Number (MUNE) with (A) absolute twitch torque and (B) normalized twitch torque. (C-D) Correlation of MUNE with (C) absolute twitch torque and (D) normalized tetanic torque. Abbreviations: motor unit number estimation (MUNE).

2.5 DISCUSSION

In this study we investigated the longitudinal progression of muscle dysfunction and MU degeneration in the SOD1(G93A) mouse model using a combination of in vivo muscle contractility, electrophysiological, behavioral and pathological outcomes. The in vivo muscle contractility assay discussed here represents a powerful tool that can rapidly and repeatedly measure the force

66 output of the MU. When combined with longitudinal electrophysiological MU connectivity measurements, muscle contractility measurements can allow a more complete analysis of motor unit innervation and functional status in degenerative models—in that both electrophysiological and force outputs can be analyzed concurrently. We show that muscle contractility can be applied in vivo and longitudinally to detect and track neuromuscular deficits. In fact, we were able to resolve deficits of muscle contractility that occur prior to motor unit deficits.

Furthermore, we showed that male SOD1(G93A) mice demonstrated earlier loss of muscle contractility compared with female SOD1(G93A) mice, consistent with prior behavioral studies showing sexual dimorphism (Veldink, Bar et al. 2003,

Heiman-Patterson, Deitch et al. 2005). We have also shown that there are moderate positive associations between muscle contractility and electrophysiological outcome measurements.

2.5.1 LOSS OF MUSCLE CONTRACTILITY IN VIVO ALIGNS WITH PRIOR IN

SITU STUDIES

In the SOD1(G93A) mouse model, overt behavioral disease onset is typically observed at approximately P90 (Gurney 1994, Chiu, Zhai et al. 1995).

However, behavioral measurements used in SOD1(G93A) mice are limited in their ability to directly and accurately measure MU integrity. A number of physiological and pathological cross-sectional studies that directly assess the MU demonstrated earlier phenotypic features relative to typical behavioral disease onset (Pun, Santos et al. 2006, Hegedus, Putman et al. 2007, Hegedus, Putman et al. 2008, Mancuso, Osta et al. 2014, Dibaj, Schomburg et al. 2015). Fast

67 fatigable muscles have increased vulnerability to denervation and are less efficient at maintaining NMJ collateral reinnervation (Pun, Santos et al. 2006).

Accordingly, in situ contractility analyses of isolated fast fatigable muscles by

Hegedus and colleagues demonstrated MU loss and corresponding contractile weakness in male mice as early as P40 (Hegedus, Putman et al. 2007).

Similarly, our results identified early loss of muscle contractility as early as P35 using non-invasive muscle contractility measurements. In contrast to prior studies, muscle contractility was assessed in the intact triceps surae muscle, which includes the gastrocnemius, a muscle predominantly comprised of fast- fatigable muscle fibers, and the soleus, predominantly composed of slow-fatigue resistant muscle fibers. Using intact muscle groups with mixed muscle fiber types closely mirrors clinical muscle testing in patients in which multiple muscles are tested in conjunction (Johnson, Sideri et al. 1973).

2.5.2 LOSS OF MUSCLE CONTRACTILITY PRECEDED LOSS OF MU

CONNECTIVITY

We had hypothesized that MU connectivity decline would precede contractility decline. However, in both SOD1(G93A) males and females, loss of contractility was noted before decline of MU connectivity. Importantly, in vivo muscle contractility measurements require intact and functioning motor axons, neuromuscular junctions and muscle excitability contraction coupling. Therefore, it was important to consider whether early NMJ degeneration might explain early twitch and tetanic muscle contraction torque losses. We found several results in our study arguing against this possibility. First, we assessed NMJ transmission

68 using SFEMG which is the most sensitive measure of NMJ transmission in vivo

(Gooch and Mosier 2001). SFEMG at P35 (when the twitch and tetanic muscle contractions were reduced) was unchanged in mutant versus wildtype male mice.

Furthermore, twitch and CMAP responses are both measured following a single supramaximal nerve stimulation, and our results showed that twitch decline occurred prior to CMAP decline in both SOD1(G93A) males and females by 7 days and 28 days respectively. Together, the discrepant twitch and CMAP findings, along with our SFEMG results, support excitation-contraction decoupling, and not NMJ transmission failure, as the cause of early contractility decline. Lastly, if muscle contractility precedes MU connectivity decline, we would expect to also see no muscle atrophy related to denervation. In our studies, wet muscle weight did not reveal triceps surae (gastrocnemius and soleus) atrophy at P70 when muscle contractility was already reduced, suggesting that loss of muscle contractility was not simply related to muscle size.

Our findings of early reduction of muscle contractility (prior to SFEMG abnormalities and CMAP and MUNE decline) suggest early subsarcolemmal abnormalities resulting in excitation-contraction decoupling. Prior studies have suggested muscle specific defects in ALS patients and SOD1(G93A) mice

(Dobrowolny, Aucello et al. 2008, Yi, Ma et al. 2011, Luo, Yi et al. 2013,

Watanabe, Atsuta et al. 2016). Increased oxidative stress in the sarcolemma and the sarcoplasmic reticulum have also been observed in SOD1(G93A) mice

(Yamada, Mishima et al. 2006, Dobrowolny, Aucello et al. 2008). The increased oxidative stress may result in muscle excitation-contraction decoupling, possibly

69 through abnormalities of calcium regulation, myofilament function or ATP production (Yamada, Mishima et al. 2006, Powers, Ji et al. 2011). One protein critical for excitation-contraction coupling which may be negatively impacted to produce early contractile decline is sarcoplasmic reticulum Ca2+ ATPase

(SERCA). SERCA pump activity is diminished following increased oxidative stress, with the intracellular Ca2+ imbalances resulting in muscle contractile weakness (Arai, Yoguchi et al. 2000, Qaisar, Bhaskaran et al. 2018).

2.5.3 LOSS OF MUSCLE CONTRACTILITY AND MU CONNECTIVITY ARE

SEXUALLY DIMORPHIC IN SOD1(G93A) MICE

There is no consensus as to why male SOD1(G93A) mice present with symptoms earlier than females. Our results are consistent with previous reports that SOD1(G93A) males exhibit earlier disease onset compared to their female counterparts, and work in human studies has suggested that the disease may be more prevalent in males (Veldink, Bar et al. 2003, Heiman-Patterson, Deitch et al. 2005, McCombe and Henderson 2010, Blasco, Guennoc et al. 2012).

Ovariectomized SOD1(G93A) females have a decreased lifespan compared to sham SOD1(G93A) females, which can be rescued with estrogen treatment, suggesting a role of hormones (Choi, Lee et al. 2008). Sex differences also appear to be background strain specific, as SJL/J and C57BL/6 x SJL/J mixed lines of SOD1(G93A) mice present earlier disease onset in males compared to females, while SOD1(G93A) expressing in the C57BL/6 line exhibit no sexual differences (Heiman-Patterson, Deitch et al. 2005). Hegedus and colleagues found no sex-specific differences in motor unit number loss or in muscle

70 contractile force deficits (Hegedus, Putman et al. 2009). Potential differences in background strains of transgenic mice used in the studies can also account for these differences. Alternatively, Hegedus et al. used an in situ paradigm to estimate MU numbers by measuring incremental muscle contractions on isolated gastrocnemius and soleus muscles. We estimated in vivo motor unit numbers electrophysiologically, and along with in vivo muscle contractility, these were performed in non-isolated gastrocnemius and soleus muscles. Recording incremental electrical responses to estimate motor units may be more sensitive to detect changes than recording incremental muscle contraction responses as we have shown in a mouse model of spinal muscular atrophy (Arnold, Porensky et al. 2014).

Our data suggest that female SOD1(G93A) mice may delay disease onset and progressive electrophysiological decline via compensatory sprouting. We showed that male SOD1(G93A) mice demonstrated morphological evidence of denervation at the NMJ at P70 whereas female SOD1(G93A) mice did not, consistent with differences we observed in the onset of muscle contractility and

MU connectivity decline. Preservation of NMJ innervation in the female

SOD1(G93A) mice (despite reduction in muscle contractility at P56) provides additional evidence that early loss of muscle contractility is a consequence downstream from NMJ integrity. Furthermore, the timing of AChR redistribution has not been characterized in SOD1(G93A) mice, however, AChRs begin to redistribute from the NMJ zone to the extrajunctional zone within 3 days of nerve transection denervation (Hartzell and Fambrough 1972). Future studies may

71 assess the timing of AChR redistribution to determine whether redistribution is delayed in female mutants.

2.5.4 IN VIVO MUSCLE CONTRACTILITY IS A POTENTIAL PRE-CLINICAL

PHYSIOLOGICAL BIOMARKER

The development of meaningful outcome measures for ALS is critically important. The methodology for measuring muscle contractility presented herein as a physiological outcome measure shares qualities of a good physiological biomarker, in that it is reproducible, minimally-invasive, easy to obtain and has the capacity to make longitudinal measures (Bowser, Turner et al. 2011). Rater reliability results are in line with intra-rater reliabilities of muscle function and MU connectivity measurements in clinical studies (ICCs ranging from 0.55 to 0.99)

(Sleivert and Wenger 1994, Clark, Cook et al. 2007, Neuwirth, Nandedkar et al.

2011, Ives and Doherty 2012, Ives and Doherty 2014, Kaya, Hoffman et al. 2014) and pre-clinical studies utilizing MU connectivity measurements in mice (ICCs ranging from 0.56 to 0.76) (Kasselman, Shefner et al. 2009).

2.6 CONCLUSION

In this chapter, we have demonstrated that when combined with longitudinal MU connectivity measurements, muscle contractility measurements allow a more complete analysis of MU innervation and functional status in degenerative models. Muscle contractility weakness preceded MU connectivity decline in both male and female SOD1(G93A) mice. Our data suggest that the contractile weakness may be the result of an early excitation-contraction decoupling mechanism within the muscle. Future pre-clinical ALS therapeutics

72 that target early muscle excitation-contraction decoupling mechanisms may be a promising treatment strategy. Furthermore, muscle contractility measurements following nerve stimulation could be performed alongside current strength tests in people with ALS, such as hand-held dynamometers, with the added benefit of not being impacted by the strength of the evaluator (Beck, Giess et al. 1999).

CHAPTER 3: AAV9 MEDIATED OVEREXPRESSION OF HSPB1

DOES NOT INFLUENCE DISEASE PROGRESSION IN THE

SOD1(G93A) MOUSE MODEL OF AMYOTROPHIC LATERAL

SCLEROSIS

3.1 STATEMENT OF CONTRIBUTION

AAV9 or PBS injections were performed by Katherine Meyer. Behavioral measurements were performed by me, while electrophysiological measurements were performed by me and David Arnold. Pathological imaging, motor neuron transduction/survival and western blot protein expression was also performed by me. Writing and figure creation was carried out by me, with David Arnold and

Stephen Kolb providing edits and revisions.

3.2 INTRODUCTION

Between 5-10% of ALS patients have a family history of the disease

(familial ALS, fALS) while in the remaining patients no pathogenic mutation can be identified (sporadic ALS, sALS). Mutations of the copper/zinc superoxide dismutase gene (SOD1) account for approximately 20% of all fALS cases (Deng,

Hentati et al. 1993, Rosen, Siddique et al. 1993, Gurney, Pu et al. 1994,

Katsuno, Tanaka et al. 2012, Hayashi, Homma et al. 2016). To date, over 180 mutations distributed throughout the SOD1 gene have been characterized in fALS patients since the first identified SOD1 mutation in 1993 (Hayashi, Homma et al. 2016). Despite no consensus regarding the mechanism by which mutations in SOD1 give rise to ALS, several toxic gain of function models have been proposed (Lin, Bristol et al. 1998, Beal 2005, Ling, Albuquerque et al. 2010, Ince,

Highley et al. 2011). One such proposed mechanism is the toxic accrual of misfolded mutant SOD1 protein (Bruijn, Becher et al. 1997, Taylor, Brown et al.

2016).

This study focuses on the ubiquitously expressed heat shock protein B1

(HSPB1), a member of the small heat shock protein family with numerous cellular functions including: protein aggregation, regulation of cytoskeleton stability, apoptosis and mRNA stability (Landry, Chretien et al. 1989, Concannon, Gorman et al. 2003, d'Ydewalle, Krishnan et al. 2011, Knapinska, Gratacos et al. 2011).

Our focus on HSPB1 as a potential therapeutic stems from a unique characteristic of the protein—overexpression of wildtype HSPB1 has demonstrated neuroprotective properties, while mutations in HSPB1 can result in the non-lethal axonal form of Charcot-Marie-Tooth disease (CMT), CMT2

(Srivastava, Renusch et al. 2012).

Transgenic animal models overexpressing the human HSPB1 appear phenotypically normal, but exhibit decreased motor neuron death in ischemia stroke models and accelerated axonal regrowth in addition to accelerated behavioral recovery in peripheral nerve injury models (Benn, Perrelet et al. 2002,

Sharp, Krishnan et al. 2006, Stetler, Cao et al. 2008, van der Weerd, Tariq Akbar et al. 2010, Ma, Omura et al. 2011). Additionally, intravenous injection with recombinant HSPB1 has demonstrated decreased cell death in stroke models

(Shimada, Tanaka et al. 2014) and injection of adenovirus overexpressing

HSPB1 has demonstrated enhanced recovery in peripheral nerve injury (Benn,

Perrelet et al. 2002).

Interestingly, over 30 mutations in HSPB1 have been found in patients with CMT. Patients develop distal muscle weakness, atrophy, and foot deformations (Rossor, Davidson et al. 2012). Furthermore, mice overexpressing mutant HSPB1 demonstrate reduced compound muscle action potentials, progressive behavioral deficits, and axonal loss (d'Ydewalle, Krishnan et al.

2011, Srivastava, Renusch et al. 2012). Elucidating the mechanism of HSPB1 neuroprotection may help to refine our understanding of how mutations in this protein selectively affect motor neurons.

The role of HSPB1 in ALS has previously been investigated using the

SOD1(G93A) transgenic mouse model. Endogenous HSPB1 levels increase in glial cells during disease progression whereas levels decrease in motor neurons prior to motor neuron death (Vleminckx, Van Damme et al. 2002, Maatkamp,

Vlug et al. 2004). It has been demonstrated that when ND7 cells overexpressing

SOD1(G93A) were virally transfected with HSPB1, there was an approximately

10% decrease in the percentage of cell death following serum removal relative to non-transfected SOD1(G93A) samples. An even greater reduction in cell death

(approximately 25%) was measured when HSPB1 and another heat shock

76 protein, HSP70, were co-transfected into these cells (Patel, Payne Smith et al.

2005). Furthermore, we have recently demonstrated a neuroprotective effect of

HSPB1 overexpression in an in vitro model. There is a toxic effect in a co-culture system of wildtype motor neurons with SOD1(G93A) mouse astrocytes, resulting in motor neuron death. Lentiviral transfection of SOD1(G93A) astrocytes to overexpress human HSPB1, however, reduced motor neuron death in co- cultures (Heilman, Song et al. 2017).

Investigations into the effect of overexpressing HSPB1 in SOD1(G93A) mice have produced a pair of conflicting papers (Krishnan, Vannuvel et al. 2008,

Sharp, Akbar et al. 2008). Krishnan et al found no neuroprotective effect from overexpression of HSPB1 (Krishnan, Vannuvel et al. 2008) while Sharp et al, reported a delay in disease-onset, as well as, increased motor neuron survival but with no corresponding survival extension in a double transgenic line compared to control SOD1(G93A) mice (Sharp, Akbar et al. 2008). Both of these groups overexpressed human HSPB1 using a chicken β-actin promoter and cytomegalovirus enhancer. One potential explanation for this discrepancy is the background strains of mice used between the groups, with Krishnan using

C57Bl6/SJL and Sharp using C57/Bl10 x CBA hybrids. The background strain of

SOD1(G93A) mice has previously been shown to influence disease progression

(Heiman-Patterson, Deitch et al. 2005), it may be that this strain difference also affected the efficacy of HSPB1 overexpression.

Our group has used self-complementary adeno associated virus 9 (AAV9) viruses to successfully transduce motor neurons and spinal cord astrocytes

(Foust, Nurre et al. 2009, Foust, Wang et al. 2010, Bevan, Duque et al. 2011,

Duque, Arnold et al. 2015, Meyer, Ferraiuolo et al. 2015). In this study, we applied this technology to test our hypothesis that HSPB1 overexpression following AAV9 injection would alters disease progression in ALS.

3.3 MATERIALS AND METHODS

3.3.1 VECTORS

A cDNA encoding the human HSPB1 gene was amplified from a pcDNA4/TO (Invitrogen) construct with primers to generate 5` AgeI and 3` HindIII restriction sites (sequences available upon request). The resultant PCR product was cloned into a dsAAV9-CB-MCS vector backbone (Foust, Nurre et al. 2009,

Foust, Wang et al. 2010) downstream of the chicken-β-actin promoter.

Expression of the transgene was confirmed by transient transfection of HEK293 cells (ATCC, Manassas VA). All scAAV viruses studied were produced by the

Viral Vector Core at Nationwide Children’s Hospital, Columbus, OH.

3.3.2 MICE

All procedures were performed in accordance with NIH Guidelines and approved by the Institutional Animal Care and Use Committee of the Research

Institute at Nationwide Children’s Hospital. B6SJLTg (SOD1(G93A)) mice and wildtype B6SJL mice were obtained from Jackson Laboratories (Bar Harbor, ME).

3.3.3 VIRAL INJECTIONS

2.0e13 vg/kg AAV9-HSPB1 virus particles were delivered via a 5μL intracerebroventricular (i.c.v.) injection (n=13) to SOD1(G93A) pups at postnatal

78 day 1 (PND1). This viral dose has been shown to effectively transduce MNs and result in disease modification in a mouse model of SMA (Meyer, Ferraiuolo et al.

2015). Control (sham) mice received an i.c.v. injection with 5μL phosphate buffered saline (PBS). The i.c.v. injection was performed using laser-pulled borosilicate glass needles (Sutter Instruments, O.D.: 1.2mm, I.D.: 0.69mm 10cm length) (Porensky, Mitrpant et al. 2012).

3.3.4 BEHAVIORAL AND SURVIVAL ASSESSMENTS

Behavioral and survival assessments were performed as previously described (Foust et al., 2013) by a blinded observer. Forelimb and hindlimb grip strengths were measured using a grip strength meter (Columbus Instruments).

Motor coordination was measured using an accelerating rotarod (Columbus

Instruments) from starting at 5RPM. Disease onset was determined by peak body weight immediately prior to decline. Disease endstage was determined by the inability to right from a prone position. Disease progression was determined by the difference between endstage and disease onset. Motor assays and body weight were measured twice a week in triplicate starting at PND55, also assessed by a blinded observer (CW).

3.3.5 ELECTROPHYSIOLOGICAL RECORDINGS

Electrophysiological recordings were performed by a blinded evaluator at

PND 60 (pre-symptomatic), 90 (disease onset), and 120 (endstage) (Gurney, Pu et al. 1994, Chiu, Zhai et al. 1995). Sciatic compound muscle action potential

(CMAP) and incremental motor unit number estimation (MUNE) were recorded from the right triceps surae muscles as previously described (Arnold, Sheth et al.

2015). Additionally, needle electromyographic recordings were performed in the right gastrocnemius muscle to assess for the presence of fibrillation potentials

(Arnold, Porensky et al. 2014). Electrophysiological recordings were assessed by a blinded observer (W.D.A.).

3.3.6 PERFUSION AND TISSUE PROCESSING

Mice were anesthetized via an intraperitoneal (ip) injection of a xylazine

(10mg/kg) and ketamine (100mg/kg) cocktail and perfused transcardially with 1X

PBS. Tissue for histology was harvested following perfusion with 4% paraformaldehyde (PFA) in 1X PBS (pH 7.4). Tissues underwent immersion fixation for 24 hours in 4% PFA at 4°C. Following fixation, tissues were washed three times in 1X PBS and stored at 4°C in 1X PBS + 0.1% sodium azide.

Tissues used for western blot analysis were harvested immediately following 1X

PBS perfusion and flash frozen in liquid nitrogen.

3.3.7 HISTOLOGY

Fixed spinal cord tissues were cryopreserved in 30% sucrose overnight at

4°C, after which tissues were embedded in optimal cutting temperature (OCT) medium and gradually frozen using liquid N2-cooled, isopentane. 20μm thick sections were collected at 150μm intervals between L3-L4 spinal segments, to assess motor neuron pools that innervate hindlimb muscles studied electrophysiologically (Nicolopoulos-Stournaras and Iles 1983).

3.3.7.1 IMMUNOHISTOCHEMISTRY

Spinal cord sections were incubated for one hour at RT in blocking buffer

(10% donkey serum/4% BSA/0.1% Triton-X 100/PBS). Primary antibody incubation was performed at RT for 2 hours using α-ChAT (Millipore, AB144P, goat [1:50]), or α-HSPB1 (Abcam, Ab1426, rabbit [1:200]). Samples were then washed 3X’s, 10min, with PBS at RT. A one hour secondary α-goat (Millipore,

A11055, donkey [1:1,000]) and α-rabbit (Millipore, A21207, donkey [1:1,000]) antibody incubation was followed by three, 10min PBS washes. Samples were mounted in Fluoromount-G (Southern Biotech). Segments were visualized at 20x and 40x magnification using a Yokogawa spinning disk confocal microscope

(Yokogawa CSU-W1) with Metamorph software (version 7.8.1.0). Total ChAT- labelled motor neurons transduced with HSPB1 were quantified by hand using

FIJI imaging software.

3.3.7.2 CRESYL VIOLET STAINING

Spinal cord segments were submerged in a progression of ethanol solutions (70%, 95%, 100%, 95%, 70% and 50%) for 5min each. Samples were stained with 0.5% cresyl violet for 5min and underwent 1min washes in 95% and

100% ethanol before being cleared with Histo-Clear (National Diagnostics).

Samples were mounted in Permount mounting media (Fisher Scientific).

3.3.8 WESTERN BLOT

Brain, spinal cord, and liver tissues were homogenized in buffer (62.5mM tris-HCl pH6.8, 10% SDS, 5mM EDTA) at room temperature (RT). Protein concentrations were determined using a Nanodrop (ND-1000). 10μg of each

81 sample was separated by SDS-PAGE and transferred to a PVDF membrane.

Blots were incubated in blocking buffer (3% BSA/1X TBS) for one hour at RT.

Blots were then incubated with α-tubulin (Abcam, ab7291, mouse [1:10,000]) or

α-HSPB1 (Abcam, ab2790, mouse [1:5,000]) primary antibodies for one hour at

RT with rocking. Blots were then washed 3X’s in 1X PBS + 0.2% Tween-20, 5min per wash. A one hour α-mouse secondary antibody incubation at RT followed

(Licor, 926-32210, goat [1:10,000]). Blots were then imaged by an infrared imaging system (Licor) and processed using Odyssey software (version 3).

3.3.9 MOTOR NEURON QUANTIFICATION

Motor neuron survival was directly visualized at 20x magnification using a

Zeiss Axiophot El-Einsatz microscope (Zeiss 45-18-89) with Metavue software

(version 6.3). 20 sections (20μm thick, 150μm separating each section) per sample were quantified. A blinded observer (CW) counted motor neurons manually using the following established criteria: cells were stained with cresyl violet acetate, were localized in the ventral horn, had a diameter ≥12.5μm and displayed a euchromatic nucleus (Ma, Omura et al. 2011).

3.3.10 STATISTICS

All statistics were performed using Graphpad Prism software (version 6).

Log rank t test was used to determine statistical significance for Kaplan-Meier survival curves (disease onset, progression and survival). Two-way, repeated measures ANOVA was performed to determine significance in grip strengths, body weight, CMAP and MUNE measurements. Unpaired t test was performed for motor neuron counts. Significance was set at p<0.05. Log rank t test with

82 multiple comparisons corrected using Bonferroni was used to determine statistical significance for Kaplan-Meier survival curves (disease onset, progression and survival) separated by gender. Significance was set at p<0.025.

3.4 RESULTS

3.4.1 HUMAN HSPB1 OVEREXPRESSED FOLLOWING AAV9-HSPB1

INJECTION IN SOD1(G93A) MICE

We first sought to determine that AAV9-HSPB1 was able to target relevant cell types. Spinal cord homogenates from presymptomatic AAV9-HSPB1 injected and sham mice were prepared, and HSPB1 protein levels were determined by western blot. Human HSPB1 was observed in mice receiving AAV9-HSPB1 and not in untreated mice, as expected (Figure 3.1A). We next characterized HSPB1 expression using immunofluorescence in the lumbar spinal cord segment from pre-symptomatic mice. We observed human HSPB1 expression in the ventral horn of injected animals (Figure 3.1B), including 40.6 ± 4.2% motor neurons co- labelled with ChAT and HSPB1 (Appendix B.3), in line with previous work

(Meyer, Ferraiuolo et al. 2015).

Figure 3.1 HSPB1 is overexpressed in SOD1(G93A) neuronal tissue

(A) Western blot of spinal cord homogenates from pre-symptomatic (PND50) SOD1(G93A) mice confirmed human HSPB1 overexpression following PND1 AAV9-HSPB1 i.c.v. injection compared to a sham control. (B) Representative immunofluorescence image of pre-symptomatic lumbar spinal cord confirming hHSPB1 expression in the ventral horn. Yellow arrowhead denotes motor neurons co-labeled ChAT/HSPB1. Scale bar 50μm.

3.4.2 HSPB1 OVEREXPRESSION DOES NOT RESCUE MOTOR

PERFORMANCE OR IMPROVE DISEASE PROGRESSION

Longitudinal motor function assays, to assess AAV9-HSPB1 treatment efficacy, were performed starting at PND55. Under our experimental conditions, we observed no significant differences between control and AAV9-HSPB1 treated mice (Table 3.1). Peak grip strength and peak latency to fall occurred at approximately the same time for each group. Disease onset, progression, and survival did not significantly differ between the cohorts either. Interestingly, gender specific differences in treatment response have been previously reported in SOD1(G93A) mice (Veldink, Bar et al. 2003, Lepore, Haenggeli et al. 2007,

Naumenko, Pollari et al. 2011, Bame, Pentiak et al. 2012). In order to determine

84 whether gender impacted treatment response in our study, we grouped the treated and control cohorts by sex. The average body mass in sham and injected male mice, as well as in female mice, remains equivalent during the course of disease progression (p>0.025 sham male vs injected male; p>0.025 sham female vs injected female) (Figure 3.2A). Similarly, grip strength and latency to fall were not significantly different between the sexes (Figures 3.2B-D). AAV9-

HSPB1 injected male mice trended towards an increased median survival compared to sham males (141 days vs 126 days), although the difference was found to be statistically insignificant (p>0.025) (Figures 3.3A). Injected male mice also trended toward delayed disease onset and an extended progression period, however these findings were also found to be statistically insignificant (p>0.025)

(Figure 3.3B-C). Conversely, AAV9-HSPB1 treated, female mice trended towards earlier disease onset, a shorter progression period and shortened survival times as compared to sham females, although were found to be statistically insignificant (p>0.025) (Figures 3.3A-C). Our results strongly indicate that overexpression of HSPB1 using AAV9 via i.c.v. injection in SOD1(G93A) mice does not delay disease onset, nor does it improve motor performance or disease outcome.

Sham AAV9-HSPB1 p Average PND86 PND91 0.41 Hindlimb Average PND86 PND86 0.31 Forelimb Average Mass PND104 PND104 0.77 Average PND94 PND94 0.12 Latency to Fall Median Disease PND105 PND105 0.85 Onset Median Disease 33 days 33 days 0.92 Progression Median Disease PND138 PND134 0.82 Survival Table 3.1 AAV9-HSPB1 has no effect on ALS disease symptoms in SOD1(G93A) mice

Ages for: peak average mass, peak average forelimb and hindlimb grip strength, peak average latency to fall and median disease outcome measures of SOD1(G93A) sham mice (n = 11) and SOD1(G93A) mice receiving AAV9- HSPB1 (n = 13). Two-way repeat measure ANOVA used for forelimb and hindlimb grip strength, latency to fall and mass. Log rank t-test performed in median disease onset, median disease progression and median disease survival.

Figure 3.2 Similar behavioral outcomes in AAV9-HSPB1 injected mice compared to sham mice

Cohorts were sub-divided by sex (AAV9-HSPB1 male, red square, n = 8; AAV9- HSPB1 female, purple upside triangle, n = 5; sham male, blue circle, n = 7; and sham female, orange triangle, n = 4). (A) No significant differences were observed in average mass between treatment cohorts. (B-D) Average forelimb and hindlimb grip strengths, as well as latency to fall off rotarod, were consistent amongst all groups. Error bars represent standard deviation. Two-way repeat measure ANOVA with multiple comparisons corrected using Bonferroni method to reduce chances of Type I error (significance at p<0.025).

Figure 3.3 Similar behavioral outcomes in AAV9-HSPB1 injected males and females compared to sham males and females

(A) Kaplan-Meir curve of disease survival in male shams (blue, dashed), AAV9- HSPB1 treated males (red, solid), female shams (orange, dashed) and AAV9- treated females (purple, solid). (B) Median survival was not significantly increased in AAV9-HSPB1 treated males compared to sham males (p=0.0252). AAV9-HSPB1 treated females and sham females had no difference in survival (p=0.21). (C-D) There were no significant differences in disease onset (C) or disease progression (D) between all sub-divided cohorts (p>0.05). Log rank t- test, with multiple comparisons corrected using Bonferroni method to reduce chances of Type I error (significance at p<0.025) performed in A-D.

3.4.3 VIRAL OVEREXPRESSION OF HSPB1 HAS NO EFFECT ON

ELECTROPHYSIOLOGICAL MEASURES OF MOTOR UNIT FUNCTION IN

SOD1(G93A) MICE

CMAP and MUNE recordings were not significantly different between cohorts at the presymptomatic, symptomatic and endstage time points (Figure

3.4A-B). These findings suggest that overexpression of HSPB1 does not improve motor unit number or connectivity in hindlimb muscles of SOD1(G93A) mice. No significant CMAP or MUNE differences were observed when cohorts were sub- divided based on sex (Figure 3.4C-D).

Figure 3.4 HSPB1 overexpression does not delay motor unit decline

(A) No significant difference in compound muscle action potential amplitude (CMAP) was measured between cohorts (p>0.05). (B) Motor unit number estimation (MUNE) recordings from the triceps surae muscles also showed no significant difference between sham and injected SOD1(G93A) mice (p>0.05). (C-D) No significant differences were observed when cohorts were further sub- divided by gender in either (C) CMAP or (D) MUNE. Error bars represent standard deviation. Two-way ANOVA performed with significance set at p<0.05 for A-B. Bonferroni post-hoc correction was performed in C-D with significance set at p<0.025 to account for splitting cohorts by gender and minimize the chances of Type I error.

3.4.4 ENDSTAGE MOTOR NEURON COUNTS ARE NOT SIGNIFICANTLY

AFFECTED FOLLOWING HSPB1 OVEREXPRESSION

End-stage motor neurons were next quantified using cresyl violet staining

(Figure 3.5A). Motor neuron counts were not significantly increased in HSPB1 overexpressing mice (n = 10) compared to sham mice (n = 9) (13 ± 3 motor neurons/section vs 12 ± 3, p>0.05) (Figure 3.5B) further supporting that overexpression of HSPB1 through AAV9 transfection does not protect motor neuron survival during disease progression.

Figure 3.5 HSPB1 overexpression did not protect against motor neuron death in SOD1(G93A) mice at disease endstage

(A) Cresyl violet stained spinal cord at lumbar segment from endstage HSPB1 overexpressing mice (left) and non-overexpressing mice (right). (B) Average motor neuron count per spinal cord section was 13 ± 3 in the AAV9-HSPB1 cohort (n=10) and 12 ± 3 in sham SOD1 mice (n=9) (p=0.075). 20x objective, scale bar = 25μm. Error bars represent standard deviation. Unpaired t test with significance set at p<0.05 used in B.

3.5 DISCUSSION

We had hypothesized that presymptomatic delivery of AAV9-HSPB1 to overexpress HSPB1 would be neuroprotective in SOD1(G93A) mice. We demonstrated using a combination of behavioral, electrophysiological and pathological measures that PND1 AAV9-HSPB1 treatment was not efficacious in

SOD1(G93A) mice. Our data from this study seems to recapitulate the findings of

Krishnan and colleagues. This group demonstrated no significant neuroprotective effect in SOD1(G93A) mice that also overexpressed HSPB1 in motor neurons

(Krishnan, Vannuvel et al. 2008). We demonstrated that SOD1(G93A) sham mice and SOD1(G93A) mice treated with AAV9-HSPB1 had symptomatic onset around PND105, as was the case in Krishnan 2008.

Before we can rule out the potential efficacy of HSPB1 overexpression in

ALS, there are caveats that should be addressed in future studies.

3.5.1 LOW DOSE AAV9-HSPB1 DOES NOT EXHIBIT NEUROPROTECTION

One possible explanation as to why we did not observe any effects is that the minimum effective dose required is larger than what we treated mutant mice with. SOD1(G93A) mice were injected with 2E+13 vector genomes/kg bodyweight

(vg/kg) of AAV9-HSPB1, based on a prior publication which demonstrated a minimum effective dose of 1.8E+13vg/kg of AAV9-SMN in a SMA mouse model

(Meyer, Ferraiuolo et al. 2015). In Meyer et al, SMAΔ7 mice demonstrated improved survival and motor performance following an AAV9-SMN i.c.v. injection

92 dose of 1.8E+13vg/kg. Higher doses of AAV9-SMN resulted in a further improved survival, however, there was no difference in motor performance between the minimum effective dose and higher doses (Meyer, Ferraiuolo et al. 2015). The minimum amount of targeted motor neurons required to see an effect may be higher for an ALS model than it is in a SMA model. We targeted about 40% of lumbar motor neurons with our AAV9-HSPB1 dose, in line with the percentage targeted by the minimum effective dose in the Meyer et al paper. The Meyer paper demonstrated an increase in the percentage of motor neurons targeted as

AAV9 dose increased, reaching a maximum of about 70% using a maximum dose of 3.3E+13vg/kg. Future studies should focus on targeting more neuronal cells by increasing the dose of AAV9-HSPB1.

3.5.2 POSSIBLE EFFICACY OF HSPB1 OVEREXPRESSION IN NON-

NEURONAL CELL TYPES

A second possible explanation as to why we did not observe a neuroprotective effect of HSPB1 overexpression in motor neurons is that overexpression in other non-neuronal cells types may be important. Other groups and ourselves have demonstrated SOD1(G93A) astrocyte toxicity towards motor neurons, which suggests that a non-cell autonomous therapeutic site may be a viable strategy (Ilieva, Polymenidou et al. 2009, Meyer, Ferraiuolo et al. 2015,

Heilman, Song et al. 2017). Indeed, we have previously demonstrated in vitro that human HSPB1 overexpression in mouse derived SOD1(G93A) astrocytes decreases motor neuron cell death in a co-culture system (Heilman, Song et al.

2017). This suggests that specifically targeting astrocytes with AAV9-HSPB1

93 transfection is an alternative therapeutic strategy. One limitation to this current study was a focus on motor neuron HSPB1 overexpression and not on astrocytes overexpressing the protein. Our group has previously demonstrated that more motor neurons than astrocytes are transfected following PND1 treatment with AAV9 (Foust, Nurre et al. 2009). In addition to targeting more motor neurons for transfection by increasing viral titer, an alternative approach is to increase the amount of astrocytes transfected to overexpress HSPB1. We hypothesize that increasing expression of HSPB1 in astrocytes would reduce toxicity and increase survival of motor neurons. Development of an AAV9 vector with an astrocyte specific promoter, such as GFAP, could be used to specifically target astrocytes (von Jonquieres, Mersmann et al. 2013).

3.6 CONCLUSIONS

We have demonstrated using a combination of behavioral, electrophysiological and pathological assessments the impracticality of AAV9-

HSPB1 as an ALS therapeutic at its current dose. Proposed future experiments to further assess the potential neuroprotective efficacy of HSPB1 in ALS include improving motor neuron transduction efficiency or alternatively specifically targeting non-neuronal cell types, like astrocytes.

CHAPTER 4: TIME COURSE OF MOTOR UNIT RECOVERY IN

THE TRICEPS SURAE FOLLOWING SCIATIC NERVE INJURY

4.1 STATEMENT OF CONTRIBUTION

Sciatic nerve crush as well as behavioral, electrophysiological and muscle physiological recovery were performed by me. Tissue harvest for pathology was performed by Anthony Reynolds and me, while NMJ imaging and quantification was also performed by me. Writing and figure creation was carried out by me, with David Arnold and Stephen Kolb providing edits and revisions.

4.2 INTRODUCTION

Peripheral nerve injury (PNI) is a relatively common occurrence, accounting for 1.6% to 3% of patients presenting to trauma centers, with annual associated healthcare costs in the billions of dollars (Evans 2001, Taylor, Braza et al. 2008, Grinsell and Keating 2014). The causes leading to PNI are diverse, although previous reports have listed severe falls, car accidents, bone fractures, gunshot wounds or knife stabbings as frequent events causing this neuropathology (Kouyoumdjian 2006). Despite the varied causes that may result in a PNI, the neuronal response is conserved—underlying pathophysiology,

95 through Wallerian degeneration, disrupts the neuronal connectivity of the motor unit (MU), causing debilitating muscle weakness in patients. Unlike most injuries to the central nervous system, however, the peripheral nervous system has the capacity to mount a regenerative response following a PNI. Up-regulation of growth enhancing proteins, such as GAP-43, CAP-23, laminin and integrin, occurs in the motor neuron and surrounding Schwann cells, promoting reformation of growth cones in the proximal nerve stump and extension towards muscle fiber innervation target site (Christie and Zochodne 2013). Despite this intrinsic regenerative capacity of peripheral motor neurons, complete functional recovery of the MU is rarely achieved in patients with more severe PNI (Ruijs,

Jaquet et al. 2005). Prolonged denervation, resulting from slow axonal regrowth

(about 1mm/day in humans vs 3mm/day in rodents), and the long distances re- growing axons must travel to reach the innervation target site, are current limitations to complete functional motor recovery (Sunderland 1947, Fu and

Gordon 1995, Ruijs, Jaquet et al. 2005, Christie and Zochodne 2013).

Assessing the efficacy of pre-clinical therapeutics and surgical outcomes in patients rely on a combination of motor functional outputs and electrophysiological output measurements (Ma, 2011; Lee and Wolfe 2000;

Ruijs, 2005). Gauging successful reconnection and functional recovery of the MU following a PNI in animal models can be accomplished through several approaches, including functional behavior recovery, MU connectivity recovery and muscle contractility recovery (Wood, Kemp et al. 2011, Navarro 2016).

Behavioral functions, such as hindlimb grip strength and rotarod performance,

96 can show when normal motor performance returns following a PNI—the ideal outcome for patients (Kemp, Cederna et al. 2017). These measurements are also easy to perform and are noninvasive, making them well-suited for longitudinal recovery studies. However, they are not without issues, chiefly their limited ability to directly assess the MU, as behavioral measurements are dependent on sensory information as well as upper motor neuron activity from the CNS (Webb, Jeffery et al. 2004, Wood, Kemp et al. 2011, Kemp, Cederna et al. 2017). Solely relying on behavioral assays may also not be reliable to assess potential therapeutics since animal behavior performances can be influenced by their motivation (Deacon 2013). Thus, assays that directly measure lower motor neurons, the motor neurons innervating skeletal muscles which are directly injured following a PNI, can provide a more detailed description of MU recovery that is not affected by whether an animal is motivated or not. Such assays can be broken down into connectivity and contractility measurements. Connectivity measurements include electrophysiological assays: compound muscle action potential (CMAP), single motor unit potential (SMUP) and motor unit number estimates (MUNE). These connectivity measurements have been frequently applied in animal models of motor neuron disease, including amyotrophic lateral sclerosis (Mancuso, Santos-Nogueira et al. 2011, Li, Sung et al. 2013), spinal muscular atrophy (Arnold, Porensky et al. 2014, Duque, Arnold et al. 2015) and distal hereditary motor neuropathy type 2 (d'Ydewalle, Krishnan et al. 2011,

Srivastava, Renusch et al. 2012). The MU loses neuronal connectivity in these neurodegenerative models, typically characterized with significantly reduced

CMAP or MUNE. Importantly, they can also demonstrate neurodegeneration in transgenic animals, even when behaviorally they appear healthy (Srivastava,

Renusch et al. 2012). While connectivity measurements directly measure the

MU, they only provide information on the electrical properties of the MU. The functional output of the MU, its ability to generate force when stimulated, can be assessed by including muscle contractility measurements. Pre-clinical studies of

PNI, which have included muscle contractility measurements, utilize highly invasive procedures, wherein muscles or groups of muscles are exposed, tendons cut and tied to force recording devices to record muscle tension following nerve stimulation (Fu and Gordon 1995, Fu and Gordon 1995, Gordon and de Zepetnek 2016, Gregor, Maas et al. 2018). This approach was carried out in two separate projects performed by Fu and Gordon in 1995, which were instrumental in demonstrating the effects on muscle contractility recovery, or lack thereof, following chronic axotomy (motor neurons are prevented from innervating a target muscle) or chronic denervation (muscles are prevented from receiving motor neuron input) peripheral nerve injuries. Muscle force recovery was capable of returning to pre-injury levels following chronic axotomy injuries but not following chronic denervation injuries, suggesting that given a long enough period of denervation muscle fiber atrophy may be an insurmountable hurdle to functional recovery (Fu and Gordon 1995, Fu and Gordon 1995).

The area of peripheral nerve regeneration is presently lacking a robust, longitudinal characterization of the relationship between MU connectivity and muscle contractility of the MU during regeneration. Highly invasive contractility

98 assays have also made it challenging to characterize MU functional output recovery during nerve regeneration. Longitudinal characterization of MU connectivity and muscle contractility recovery following PNI can help elucidate future therapeutic interventions. First, we will be able to refine treatment times during recovery. And second, by comparing the recovery timelines of MU connectivity with contractility, we can optimize target sites to address. For example, if MU connectivity recovers before muscle contractility recovers, we may prioritize development of muscle specific therapeutics to improve contractile recovery.

In this chapter, we have detailed the early timeline of MU connectivity and muscle contractility recovery in a murine PNI model. We utilized a relatively mild

PNI paradigm where motor neuron death does not occur, double sciatic nerve crush, in FVB wildtype mice (Ma, Omura et al. 2011). We compared regeneration between MU connectivity and muscle contractility to test our hypothesis that connectivity and contractility would recover concurrently. Interestingly, two phases of MU recovery over a span of 28 days post-injury (28dpi) were demonstrated, an early phase of contractility and behavioral recovery and a latent phase of connectivity recovery. This in vivo longitudinal paradigm of assessing MU connectivity and muscle contractility demonstrates the underlying discrepancy between electrophysiological recovery and functional force recovery—two important outcomes following a PNI. It further suggests that some other underlying biological process, other than innervation status, may be involved in the functional status of the MU during peripheral nerve recovery.

4.3 METHODS

4.3.1 ANIMALS

All surgical, behavioral, electrophysiological, muscle physiological and pathological procedures were performed in accordance with NIH Guidelines and approved by the Institutional Animal Care and Use Committee of the Ohio State

University. Wildtype FVB/NJ mice were obtained from Jackson Laboratories (Bar

Harbor, ME).

4.3.2 SURGERY

A total of 16 adult (2-3 months) female mice underwent a double sciatic nerve crush procedure on their right hindlimb. Anesthesia induction was achieved under 5% isoflurane. Once mice demonstrated a lack of response following a hind paw pinch, they were transferred to a heating pad (set at 37°C) under an anesthesia nosecone. The surgical procedure was performed under 2-3% isoflurane. The sciatic nerve received two 15s standard crushes using locked forceps with a 15s release in between; the crush site was approximately 1cm distal from the spinal cord and marked with powdered carbon (Figure 4.1). After surgery, the incision site was closed using three to four surgical staples and mice recovered on heating pads before being returned to their home cage. The entire procedure, from anesthesia induction to waking up from anesthesia, lasted no longer than 20 minutes.

100

Figure 4.1 Position of sciatic nerve crush site

The right sciatic nerve was exposed following an initial incision and crushed mid- thigh (red star). The crush site was located before the sciatic nerve branches to the peroneal and tibial nerves (white dashed lines).

4.3.3 RECOVERY TIMELINE

The timeline following sciatic nerve crush is shown in Figure 4.2. Baseline behavioral, electrophysiological and muscle physiological measurements were made prior to PNI. Behavioral, electrophysiological and muscle physiological measurements were then performed at 7dpi, 14dpi, 18dpi, 21dpi and 28dpi.

Muscle samples were harvested from a separate cohort of mice at 0dpi (n = 3) and 14dpi (n = 3) to perform NMJ imaging and quantification.

101

Figure 4.2 Timeline for assessing motor unit reconnectivity

Longtiduinal motor unit reconnection was assessed using a combination of outcome measurements.

4.3.4 BEHAVIOR - HINDLIMB GRIP STRENGTH

Behavioral assessments were performed prior to electrophysiological and muscle physiological measurements to prevent any potential negative effects of residual isoflurane from influencing behaviors.

The average of four ipsilateral hindlimb grip strengths were measured using a Chatillon DEF2-002 grip strength meter (Columbus Instruments). Mice were positioned to allow only the ipsilateral hindpaw to grasp a grid attached to a force tension sensor and were pulled toward the evaluator for the length of the grip meter (Miller, Kim et al. 2006).

4.3.5 ELECTROPHYSIOLOGY

CMAP, SMUP and MUNE measurements were recorded from the ipsilateral hindlimb using established parameters (Arnold, Sheth et al. 2015). An active ring electrode was placed superficially over the ipsilateral triceps surae

102 with the reference ring electrode placed superficially over the metatarsals of the ipsilateral hindpaw. The sciatic nerve was stimulated (0.1ms pulse, 1-10mA intensity) using two insulated monopolar needles (28G). Peak CMAP amplitude was recorded using the baseline-to-peak amplitude following supramaximal stimulation. The distance between the positive wave and negative wave peaks was used to determine the “peak-to-peak” distance (Figure 4.3A). The sciatic nerve was next sub-maximally stimulated in increments of 0.3mA until 10 incremental, submaximal responses were recorded, which was averaged to estimate the SMUP (Figure 4.3B). MUNE was calculated via the following equation:

퐶푀퐴푃 ("푝푒푎푘 푡표 푝푒푎푘") 푀푈푁퐸 = 퐴푣푒푟푎푔푒 푆푀푈푃

103

Figure 4.3 Representative motor unit electrophysiological waveforms

(A) Representative compound muscle action potential (CMAP). The distance between baseline (1) and peak negative wave (2) is the peak CMAP amplitude (mV) and the distance between the peak negative wave and peak positive wave (3) is the peak-to-peak distance (mV). The stereotypical biphasic waveform of the CMAP results from volume conduction, ie the traveling of electric potential relative to the recording electrode (Lateva, McGill et al. 1996). The initial negative peak results when the muscle action potential occurs directly under the recording electrode, with the latent positive peak the result of the muscle action potential traveling away from the recording electrode. The waveform is the result of voltage-gated sodium channels opening and an inflow of sodium into the muscle fiber. (B) Representative tracing of 10 submaximal stimulations to estimate an average single motor unit potential (SMUP) (µV). The peak-to-peak distance and the average SMUP can be used to calculate the motor unit number estimate (MUNE). Note the sensitivity differences between (A) and (B), reflecting the submaximal responses in SMUP. Image modified from Arnold, Sheth et al. 2015.

104

4.3.6 MUSCLE PHYSIOLOGY

Following connectivity recordings, mice underwent a triceps surae plantarflexion torque assessment using an in vivo physiology apparatus (Model

1300A, Aurora Scientific Inc, Canada). The right hindpaw was taped to the force sensor and positioned at 90°. The leg was then extended to position the knee in the locking position then securely locked in place. Two monopolar electrodes were inserted over the tibial nerve, just posterior to the knee. Peak plantarflexion twitch was measured following instantaneous supramaximal stimulation (0.2ms square wave pulse). Maximum tetanic contraction was measured following a train of supramaximal square wave stimulations at 0.2ms duration delivered at 125Hz stimulation frequency.

4.3.7 PERFUSION AND TISSUE PROCESSING

Tissue samples were collected and weighed at 0dpi and at 14dpi. Mice were anesthetized under 5% isoflurane and perfused transcardially with 1X PBS followed immediately with 4% PFA. Ipsilateral and contralateral soleus muscles were harvested and teased into individual muscle fibers under a dissecting microscope. Following teasing, soleus muscle fibers immediately underwent an immunohistochemistry procedure to image and quantify NMJs.

4.3.8 IMMUNOFLUORESCENCE

Teased fixed soleus muscle fibers were blocked for two hours at room temperature (RT) in blocking buffer (10% goat serum, 4% BSA, 3.0% TritonX-

100) after which they were incubated in NF-200 (Abcam, ab72996, chicken

[1:5000]) primary antibody solution for 24hr at 4°C. Samples were washed three

105 times in PBS, 10min each, then incubated in α-Chicken-594 secondary antibody

(ThermoFisher, A11042, goat [1:1000]) and α-Bungarotoxin-488 (ThermoFisher,

B13422, [1:1000]) for 2hr at RT. Samples were washed three times in PBS,

10min each, then mounted under a glass slide with Fluoromount-G (Southern

Biotech). Samples were visualized at 20x magnification using a Leica confocal microscope (Leica DM IRE2) with Leica software (version 2.1). Images were viewed in FIJI (LOCI, University of Wisconsin-Madison) to quantify NMJ innervation, the co-labeling of NF-200 and α-Bungarotoxin. 95-120 NMJs per muscle sample (per mouse) were scored as fully innervated or denervated.

4.3.9 STATISTICS

All statistics were performed using Graphpad Prism software (version 6).

Repeat-measure two-way ANOVA was performed to test for significance in longitudinal recovery and longitudinal percent recovery of CMAP, MUNE, SMUP, twitch, tetanic, and grip strength. Standard t-test was performed to determine significance of NMJ denervation. Pearson correlations were performed to test for associations between outcome measurements, with correlation coefficient strengths interpreted using prior guidelines (Mukaka 2012). Significance was set at p<0.05.

106

4.4 RESULTS

4.4.1 MUSCLE CONTRACTILITY RECOVERY PRECEDES MU

CONNECTIVITY REGENERATION

MU connectivity and muscle contractility recovery were measured out to

28 days post-injury (28dpi) (Table 4.1). By 7dpi MU connectivity and muscle contractility outcome assays demonstrated characteristics of PNI relative to their pre-injury baseline measures. Reductions were recorded in CMAP (4.87 mV ±

2.12mV vs 41.4mV ± 5.36mV, p<0.001), MUNE (14 ± 8 vs 345 ± 96, p<0.001), twitch torque (0.33mN-m ± 0.2mN-m vs 2.28mN-m ± 0.47mN-m, p<0.001) and tetanic torque (1.44mN-m ± 0.71mN-m vs 11.29mN-m ± 1.24mN-m, p<0.001) while SMUP had increased (529.4µV ± 254.48µV vs 221.5µV ± 55.11µV, p<0.001). By 18dpi twitch torque had recovered to pre-injury baseline levels

(1.97mN-m ± 0.41mN-m vs 2.28mN-m ± 0.47mN-m, p>0.05), shortly followed by

SMUP recovery at 21dpi (316.3µV ± 150.91µV vs 221.5µV ± 55.11µV, p>0.05).

There was no complete recovery to baseline in CMAP (20.8mV ± 6.21mV vs

41.4mN ± 5.36mV, p<0.001), MUNE (126 ± 58 vs 345 ± 96, p<0.001) or tetanic torque (9.15mN-m ± 0.6mN vs 11.29mN ± 1.24mN, p<0.001) measurements by the end of the study at 28dpi.

107

Grip Body CMAP SMUP Twitch Tetanic MUNE Strength Mass dpi (mV) (µV) (mN-m) (mN-m) (±SD) (kg) (g) (±SD) (±SD) (±SD) (±SD) (±SD) (±SD)

41.4 221.5 345 2.28 11.29 54.3 22.89 0 (5.36) (55.11) (96) (0.47) (1.24) (11.89) (1.88)

4.87 529.4 14 0.33 1.44 0 22.55 7 (2.12) (254.48) (8) (0.20) (0.71) (0) (1.77) *** *** *** *** *** ***

5.13 423.9 20 1.03 4.25 29.7 23.11 14 (2.03) (161.96) (6) (0.56) (2.19) (7.13) (1.54) *** *** *** *** *** ***

6.42 331.5 36 8.05 40 1.97 22.97 18 (2.21) (141.5) (23) (1.80) (13.3) (0.41) (1.51) *** * *** *** *

8.18 46 8.34 47.2 316.3 2.09 23.4 21 (1.44) (22) (0.76) (9.28) (150.91) (0.35) (1.42) *** *** ***

20.8 126 9.15 54.1 303.2 2.32 24.01 28 (6.21) (58) (0.60) (6.14) (72.04) (0.24) (1.95) *** *** ***

Table 4.1 Longitudinal recovery of MU connectivity, muscle contractility behavior following PNI

Extent of recovery relative to pre-injury baseline for MU connectivity (CMAP, MUNE, and SMUP), muscle contractility (twitch and tetanic torque) and raw hindlimb grip strength. Body mass did not significantly change over the course of recovery.* = p < 0.05 and *** = p < 0.001.

We calculated the percent recovery (relative to baseline pre-injury measurements) at each time point to better compare the extent of recovery

108 between measurements (Figure 4.4). By 14dpi the percent recovery of twitch torque (39% ± 21%) was significantly improved compared to percent recovery of

CMAP (12% ± 5%) and percent recovery of MUNE (6% ± 2%) (Figure 4.4A).

Significant percent recovery of tetanic torque (39% ± 21%) relative to percent recovery of CMAP (12% ± 5%) and MUNE (6% ± 2%) was measured at 14dpi

(Figure 4.4B).

Figure 4.4 Muscle contractility recovery precedes MU connectivity recovery following sciatic nerve crush

(A) By 14dpi the percent recovery of twitch torque (blue circle) had improved compared to percent recovery of CMAP (black square) and MUNE (red triangle). (B) Percent recovery of tetanic torque (blue circle) had also improved by 14dpi compared to both percent recovery of CMAP (black square) and MUNE (red triangle). Error bars denote standard deviation. * p<0.05, ** p<0.01, *** p<0.001.

109

4.4.2 MUSCLE CONTRACTILITY RECOVERY ALIGNS WITH GRIP

STRENGTH RECOVERY

Hindlimb grip tests were also performed during recovery to compare with muscle contractility recovery (Table 4.1). A significant drop in hindlimb grip relative to pre-injury baseline was recorded at 7dpi (0kg ± 0kg vs 54.3kg

±11.89kg, p<0.001). Mice exhibited recovered hindlimb grip strength relative to pre-injury baseline levels by 21dpi (47.2kg ± 9.28kg vs 54.3kg ± 11.89kg, p>0.05). Percent recovery of hindlimb grip strength was further calculated to compare with muscle contractility recovery (Figure 4.5). There were no significant differences between the percent recovery of hindlimb grip strength and twitch torque throughout recovery (Figure 4.5A). The percent recovery of hindlimb grip strength and tetanic torque were similar for most of the recovery timeline, however by 28dpi hindlimb grip strength had significantly diverged (102% ± 16% vs 82% ± 9%, p<0.05) (Figure 4.5B).

110

Figure 4.5 Muscle contractility recovery closely aligns with behavioral recovery following sciatic nerve crush

(A) Percent recovery of hindlimb grip strength (red square) and twitch torque (blue circle) were similar over 28dpi (B) The percent recovery of hindlimb grip strength (red square) and tetanic torque (blue circle) were similar out to 21dpi, but by 28dpi grip recovery was significantly more improved than tetanic torque. Error bars denote standard deviation. * = p<0.05.

4.4.3 SOLEUS MUSCLE PARTIALLY INNERVATED EARLY IN

REGENERATION

We further characterized recovery by imaging and quantifying soleus muscle innervation (Figure 4.6). There was no significant difference in NMJ innervation between ipsilateral and contralateral soleus muscles at baseline

(83.2% ± 6.0% vs 86.7% ± 5.5%, respectively) (Figure 4.6A-B). NMJ innervation in the soleus muscle was then quantified at 14dpi, when MU connectivity recovery and muscle contractility recovery were first beginning to diverge (Figure

4.6A). Ipsilateral soleus NMJ innervation was reduced compared to contralateral soleus NMJ innervation (32.0% ± 11% vs 87.0% ± 3%, p<0.001) (Figure 4.6B).

111

Figure 4.6 Partial NMJ innervation of soleus muscle at 14dpi

(A) Representative images of soleus muscles at 0dpi (n = 3) (a-f) and 14dpi (n = Continued

112

3) (g-l). Co-labelling of NF-200 (red) and α-Bungarotoxin (green) was imaged in both contralateral (a-c, g-i) and ipsilateral (d-f, j-l) soleus muscles. (B) Percent innervation of ipsilateral soleus muscles (blue) was significantly reduced compared to the contralateral soleus muscles (red). Error bars denote standard deviation. *** = p<0.001. Images taken at 20x, scale bar = 50µm.

4.4.4 CORRELATIONS BETWEEN MUSCLE CONTRACTILITY, MU

CONNECTIVITY AND BEHAVIOR DURING REGENERATION

We next investigated possible correlations between MU connectivity, muscle contractility and behavioral recovery (Table 4.2). Twitch force was moderately correlated with CMAP (r = 0.53, p<0.001) and MUNE recovery (r =

0.47, p<0.001) while strongly correlated with grip strength (r = 0.88, p<0.001).

Tetanic force recovery was also moderately correlated with CMAP (r = 0.68, p<0.001) and MUNE recovery (r = 0.63, p<0.001) while strongly correlated with grip strength (r = 0.89, p<0.001). Twitch and tetanic recovery were moderately negatively correlated with SMUP (r = -0.55 and r = -0.51, respectively, p<0.001) while CMAP, MUNE and grip strength were weakly negatively correlated with

SMUP (r = -0.37, r = -0.47 and r = -0.49, respectively, p<0.001).

113

Twitch Tetanic Grip CMAP MUNE SMUP Force Force Strength

Twitch - 0.93 0.53 0.47 -0.55 0.88 Force

Tetanic 0.93 - 0.68 0.63 -0.51 0.89 Force

CMAP 0.53 0.68 - 0.94 -0.37 0.55

MUNE 0.47 0.63 0.94 - -0.47 0.47

SMUP -0.55 -0.51 -0.37 -0.47 - -0.49

Grip 0.88 0.89 0.55 0.47 -0.49 -

Table 4.2 Pearson correlations during nerve recovery between muscle contractility, MU connectivity and behavioral measurements

Correlation analysis between MU connectivity, muscle contractility and behavioral measurements during recovery as well. Significance set at p<0.05. All coefficients were significant, with p<0.001.

4.5 DISCUSSION

In this study we characterized the longitudinal progression of MU integrity in wildtype FVB mice utilizing a strategy of combined electrophysiological, muscle physiological, behavioral and pathological outcome measurements. We show that in vivo muscle contractility recovery precedes MU connectivity recovery following a double sciatic nerve crush injury. Underlying NMJ pathology demonstrated partial innervation of the soleus muscle, coinciding with the observed partial muscle contractile recovery at the same time point. Furthermore,

114 in vivo muscle contractility recovery closely aligned with hindlimb grip strength recovery and strongly correlated with one another.

4.5.1 DISCREPANCY BETWEEN MU CONNECTIVITY RECOVERY AND

MUSCLE CONTRACTILITY SUGGESTS EARLY AND LATENT RECOVERY

PHASES

Numerous pre-clinical studies assessing MU recovery following nerve injury have utilized either electrophysiological or muscle physiological tools (Fu and Gordon 1995, Fu and Gordon 1995, Sahenk, Galloway et al. 2010, Morrison,

Tsingalia et al. 2015, Glat, Benninger et al. 2016). However, the field is currently lacking a detailed, longitudinal time course of recovery, relating MU connectivity with muscle contractility. We demonstrated earlier muscle contractility recovery compared to connectivity recovery. One possible explanation for this result is terminal axon collateral sprouting. Axons that reach their target muscle site can sprout to innervate up to five times more muscle fibers relative to their pre-injury innervation, increasing their MU size, before eventually being replaced by other regenerating axons (Brown and Ironton 1978). An early paper from Fu and

Gordon demonstrated using an in situ muscle contractile paradigm that muscle contractility can recover to pre-injury levels despite significantly reduce MU numbers (Fu and Gordon 1995). The reduced motor units demonstrated a larger innervation ratio, indicating increased collateral sprouting in the target muscle.

Despite incomplete MU recovery, the increased innervation ratio compensated reduced motor units, resulting in “recovered” contractility (Fu and Gordon 1995).

115

We hypothesized that connectivity and contractility would recover concurrently following a mild PNI, possibly through collateral sprouting of axons to muscle fibers. Interestingly, we observed increased SMUP at 7dpi and 14dpi, which typically indicates collateral sprouting and increased MU size. However, in this same timeframe we only observed compensation of contractility but not of

CMAP. Both CMAP and muscle contractile assays assess the integrity of the

MU, either electrophysiologically (CMAP) or via force production (twitch and tetanic) and are dependent on muscle fiber innervation. The absence of CMAP recovery despite both increased SMUP and muscle contractile recovery may suggest some masking effect of CMAP recovery.

CMAP recovery could possibly be masked by other regenerative events occurring during MU reconnection. One such event may involve Schwann cells and myelination of peripheral nerves, which breaks down and recycled following a PNI (Jessen and Mirsky 2016). During axonal regrowth towards the terminal muscle site, juvenile Schwann cells reform and myelinate around axons at different rates (Jessen and Mirsky 2016). Immature Schwann cells and different rates of myelination can influence the action potential propagation rate of motor units when stimulated, resulting in temporal dispersion (Figure 4.7).

Asynchronous MU responses can cause positive and negative electrophysiological waves to cancel out, resulting in a reduced CMAP readout.

Unlike CMAP, muscle contraction measurements are not dependent on the myelinated state of motor axons. This is supported by the NMJ quantification of the soleus muscle, which demonstrated increased innervation by 14dpi—

116 corresponding with the earlier recovery of muscle contraction compared to MU connectivity.

Figure 4.7 Temporal dispersion can mask MU connectivity recovery

(A) Normal MU electrical output is recorded when firing rates are synchronized. (B) Asynchronous MU responses, which can arise from demyelinated axons, results in positive and negative peak wave cancellation. This results in a reduced MU electrical output, despite no evidence of denervation. Image modified from (Mallik and Weir 2005).

117

4.5.2 BEHAVIORAL FUNCTION AND UNDERLYING MUSCLE

CONTRACTILITY

Unlike MU connectivity recovery, muscle contractility and hindlimb grip strength recovery were closely aligned following a sciatic nerve injury. Contractile recovery and behavioral recovery also demonstrated a stronger positive correlation compared to contractile and connectivity recovery. One possible explanation is unlike MU connectivity assays, hindlimb grip strength and muscle contractility assays measure the ability of the MU to produce a force output (i.e. tension). It seemed as though twitch torque, more so than tetanic torque, was even closer aligned to hindlimb grip strength. All three outcome measurements demonstrated similar recovery trends early in regeneration however, by 28dpi twitch and hindlimb grip strength, but not tetanic torque, recovered to baseline pre-injury levels. As in the previous section, this suggests a latent recovery phase for tetanic torque generation in the muscle. Tetanic responses arise from the summation of rapid twitches, and the inability of tetanic torque to recover by

28dpi could also reflect asynchronous MU firing causing inadequate summation of twitches. Alternatively, subsarcolemmal factors could also impact twitch summation such as: sarcolemma membrane potential, Ca2+ concentration, and

ATP availability in the muscle (Fitts 1994). Future work can investigate subsarcolemmal activity during peripheral nerve regeneration to better understand the discrepancy between twitch recovery and tetanic recovery.

Overall, these data suggest two phases of MU recovery. First, an early phase, marked by contractile functional recovery, then a latent phase of MU

118 recovery, marked by eventual neuronal connectivity recovery. We also demonstrated similar recovery timelines for contractility and behavioral assessments. Future studies can further investigate underlying pathology at additional time points to further refine how underlying molecular mechanisms affect muscle contractile and MU connectivity recovery, such as the extent of remyelination at 14dpi—when MU connectivity and muscle contractility recovery first began to diverge.

119

CHAPTER 5: ASSESSING THE REGENERATIVE EFFICACY OF

AAV9-SMN AND AAV9-HSPB1 FOLLOWING A PERIPHERAL

NERVE INJURY

5.1 STATEMENT OF CONTRIBUTION

Sciatic nerve crush and cisterna magna injections were performed by me.

Behavioral and electrophysiological recovery were also performed by me. Along with me, Anthony Reynolds assisted with tissue harvest for pathology, motor neuron transduction and protein expression; motor neuron and NMJ imaging and quantification was carried by me. Writing and figure creation was carried out by me, with David Arnold and Stephen Kolb providing edits and revisions.

5.2 INTRODUCTION

Frequently in cases of severe nerve crush or transection, the only viable treatment is surgery, including: direct re-suture, nerve graft and nerve conduit.

Direct nerve re-suture entails suturing the proximal and distal segments together and is typically only performed for short (<2-3cm) gaps (Barton, Morley et al.

2014). Care must be taken when applying sutures, as they can cause undue stress on the nerve, including compression and an immune response (Menovsky

120 and Beek 2003). A nerve graft can either be performed using an expendable nerve branch within the patient (“autograft”) or using the same nerve but from a cadaver donor (“allograft”). A nerve graft would be favorable over a nerve re- suture in cases of larger gaps (>3cm), where increased tension on the nerve from suturing would negatively impact recovery (Flores, Lavernia et al. 2000).

Nerve grafts additionally provide a biological scaffold containing axonal regrowth promoting Schwann cells, and in the case for autografts immune rejection would be negated (Barton, Morley et al. 2014). Nerve grafts are not without limitations, as sacrificing expendable nerve branches still results in minor sensory/motor deficits—termed donor site morbidity—and if using an allograft the risk of a donor rejection is present (Barton, Morley et al. 2014). The third surgical procedure, a nerve conduit, is similar to nerve grafts in that it can be used to cover larger

(>3cm) gaps in a nerve. Nerve conduits may be synthetic or biological and optimized to promote axon regeneration. A previous study demonstrated improved electrophysiological recovery in transected rat sciatic nerves when

Schwann cells were aligned within the nerve conduit (Gonzalez-Perez,

Hernandez et al. 2018). Unlike nerve grafts, there is no risk of donor site morbidity since no minor branches are sacrificed. However, previous materials used for nerve conduits, such as silicone, have been demonstrated to be too rigid producing additional nerve compression in patients (Barton, Morley et al. 2014).

Despite the wide range of potential nerve repair surgeries available to patients, functional recovery is considerably lacking following surgery, with only about half of patients experiencing complete motor recovery (Ruijs, Jaquet et al.

121

2005). It is clear that additional work is necessary to improve functional recovery in patients following nerve injury. One possible surgical alternative is gene therapy—providing a potentially therapeutic gene to treat disease (de Winter,

Hoyng et al. 2013). Researchers have been refining various viral vectors to better target the cells constituting the PNS over several decades (de Winter, Hoyng et al. 2013). One viral vector that can potentially be used to enhance recovery following PNI is the adeno-associated virus (AAV), owing to its gene transfer efficiency and relatively low immunogenicity (Sun, Anand-Jawa et al. 2003). Of these, AAV9 is a promising tool for delivering neuroprotective genes and accelerating MU recovery, in part due to the aforementioned strengths of AAV as well as AAV9’s previous use in neurodegenerative disorders—spinal muscular atrophy (Foust, Wang et al. 2010, Duque, Arnold et al. 2015), Alzheimer’s disease (He, Pan et al. 2017), Parkinson’s disease (Xue, Ma et al. 2010), and amyotrophic lateral sclerosis (Foust, Salazar et al. 2013)—as well as in a clinical trial for spinal muscular atrophy (Mendell, Al-Zaidy et al. 2017). Two promising candidates for improving MU recovery following a PNI are heat shock protein B1 and survival motor neuron protein.

As a stress chaperone protein, heat shock protein B1 (HSPB1) possesses numerous cellular activities including: protein aggregation, regulation of cytoskeleton stability, regulation of apoptosis and mRNA stability (Landry,

Chretien et al. 1989, Concannon, Gorman et al. 2003, d'Ydewalle, Krishnan et al.

2011, Knapinska, Gratacos et al. 2011). Increased expression following cellular stress in vitro is transient, with HspB1 protein expression returning to basal levels

122 once the stress subsides (Knapinska, Gratacos et al. 2011). HSPB1 has also demonstrated neuroprotective characteristics in several disease models, including stroke and amyotrophic lateral sclerosis (ALS) (Vleminckx, Van Damme et al. 2002, Sharp, Akbar et al. 2008, Stetler, Cao et al. 2008, van der Weerd,

Tariq Akbar et al. 2010, Shimada, Tanaka et al. 2014). Transgenic mice with neuronal HSPB1 overexpression have smaller areas of cell death in the brain following middle cerebral artery occlusion (MCAO) compared to wildtype mice

(van der Weerd, Tariq Akbar et al. 2010). In ALS mouse models, HSPB1 overexpression preserves some muscle function and delays disease onset, though mice still ultimately develop ALS symptoms (Sharp, Akbar et al. 2008).

Additionally, using transgenic mice to neuronally overexpress HSPB1, groups have demonstrated a neuroprotective role following a PNI (Benn, Perrelet et al.

2002, Ma, Omura et al. 2011). It has previously been demonstrated that HSPB1 protein expression increases in mouse dorsal and ventral horns of lumbar spinal cord following a sciatic nerve injury (Benn, Perrelet et al. 2002). Furthermore, immunofluorescence demonstrated HSPB1 co-labeling with ATF3, a marker for neuronal injury, suggesting that injured neurons in particular were overexpressing HSPB1 (Benn, Perrelet et al. 2002). Transgenic mice with neuronal HSPB1 overexpression have also demonstrated earlier behavioral recovery as well as accelerated axonal regrowth following sciatic nerve crush and sciatic nerve transection injuries (Ma, Omura et al. 2011).

Survival motor neuron protein (SMN) is a second potential therapeutic that will be assessed in this thesis. The best characterized function of SMN is its role

123 in assembling small nuclear ribonucleoprotein (snRNP) complexes, which actively play a role in splicing pre-mRNA introns in the nucleus (Beattie and Kolb

2018). The role SMN plays in motor neuron health is perhaps best characterized in spinal muscular atrophy (SMA), a motor neuron disease common in infants

(1:10,000 births) (Sugarman, Nagan et al. 2012). SMA, like ALS and denervating injuries of the PNS, results in denervation at the NMJ and eventual muscle atrophy. While potentially lethal like ALS, clinicians can classify patients as SMA type 0 to type 4, ranging from most severe to least severe (Arnold and Burghes

2013, Kolb and Kissel 2015). Common clinical features of SMA typically include proximal muscle weakness and atrophy, with the more severe cases requiring respiratory support. While SMN is predominantly associated with infants and early childhood as it relates to SMA, work from Kariya and colleagues has demonstrated the importance of SMN in mature, adult motor neurons. A Cre- inducible transgenic mouse line was developed to determine what effect knockdown of SMN protein in adult mice had on the MU and motor function

(Kariya, Obis et al. 2014). Neither significant behavioral deficits nor diminished lifespans were observed in adult mice with SMN knocked down—contrary to when SMN is knocked down in pups. However, pathological assessment in these mice demonstrated abnormal NMJ morphology as they aged—AChR fragmentation and poor pre- and post- synaptic overlap. Furthermore, these SMN knockdown mice demonstrated delayed behavioral and electrophysiological recovery compared to control mice following a sciatic nerve crush procedure. The findings suggest that SMN protein may be necessary for MU integrity.

124

Here, we utilized AAV9 vectors to test our hypothesis that overexpression of HSPB1 or SMN following a PNI will accelerate MU recovery. We performed a longitudinal assessment of MU integrity using a combination of behavioral and electrophysiological measurements. No differences in recovery times were observed in mice overexpressing HSPB1 or SMN compared to control mice.

However, mice overexpressing SMN did demonstrate a significant increase in hindlimb grip strength later in the study time course, which may indicate a possible beneficial effect following nerve injury.

5.3 MATERIALS AND METHODS

5.3.1 ANIMALS

All surgical, behavioral, electrophysiological, and pathological procedures were performed in accordance with NIH Guidelines and approved by the

Institutional Animal Care and Use Committee of the Ohio State University.

FVB/NJ wildtype breeder mice were obtained from Jackson Laboratories (Bar

Harbor, ME). A total of 51 mice were used throughout the study. 48 mice were randomly split into four cohorts—AAV9-SMN treated, AAV9-HSPB1 treated,

PBS-1 sham and PBS-2 sham. PBS-1 sham mice served as negative controls for

AAV9-HSPB1 treated mice and PBS-2 sham mice served as negative controls for AAV9-SMN treated mice. Three mice were treated with AAV9-GFP to quantify motor neuron transduction.

125

5.3.2 VIRAL VECTORS

5.3.2.1 AAV9-GFP

AAV9-GFP plasmid vector was generously provided by Dr. Foust. Briefly, enhanced GFP was cut and amplified from the pCMV-GFP plasmid to generate

5’ AgeI and 3’ HindIII restriction sites (McCarty, Monahan et al. 2001, Foust,

Nurre et al. 2009). The PCR product was cloned into a scAAV9-CB-MCS vector backbone (Foust, Nurre et al. 2009, Foust, Wang et al. 2010) downstream of a cytomegalovirus (CMV) enhancer/chicken-β-actin (CBA) promoter (Appendix

A.2). Expression of the transgene was confirmed by transient transfection of

HEK293 cells (ATCC, Manassas VA). scAAV virus was produced by Vector

Biolabs (Malver, PA).

5.3.2.2 AAV9-SMN

AAV9-SMN plasmid vector was provided by Dr. Foust. cDNA encoding human SMNΔ7 was amplified from a pcDNA3 construct (Le, Pham et al. 2005) with primers to generate 5’ EcoRV and 3’ PmeI restriction sites. The amplified product was cloned into a scAAV-CB-MCS vector backbone, downstream of a

CMV enhancer and CBA promoter (Appendix A.3). Expression of the transgene was confirmed by transient transfection of HEK293 cells (ATCC, Manassas VA). scAAV virus was produced by Vector Biolabs (Malver, PA).

5.3.2.3 AAV9-HSPB1

A cDNA encoding the human HSPB1 gene was amplified from a pcDNA4/TO (Invitrogen) construct with primers to generate 5` AgeI and 3` HindIII

126 restriction sites. The resultant PCR product was cloned into a scAAV9-CB-MCS vector backbone downstream of a CMV enhancer/CBA promoter (Appendix A.4).

Expression of the transgene was confirmed by transient transfection of HEK293 cells (ATCC, Manassas VA). scAAV virus was produced by the Viral Vector Core at Nationwide Children’s Hospital, Columbus, OH.

5.3.3 SURGICAL PROCEDURES

5.3.3.1 SCIATIC NERVE CRUSH

A total of 48 mice (24 male, 24 female) at post-natal day (PND) 30 underwent a sciatic nerve crush procedure on their right hindlimb. Anesthesia induction was achieved under 5% isoflurane. Once mice demonstrated a lack of response following a hind paw pinch, they were transferred to a heating pad

(37°C) and positioned under an anesthesia nosecone. The surgical procedure was performed under 2-3% isoflurane. The sciatic nerve received two 15s standard crushes using locked forceps with a 15s release in between; the crush site was approximately 1cm distal from the spinal cord (mid-thigh) and marked with powdered carbon. After surgery, the incision site was closed using three to four surgical staples and mice were prepared for cisterna magna injection. The entire procedure lasted no longer than 20 minutes.

5.3.3.2 CISTERNA MAGNA INJECTION

Immediately following the sciatic nerve crush procedure, while still anesthetized, a cisterna magna (CM) injection was performed. Mice were positioned at approximately 140° for the duration of the surgery to aid in exposing

127 the CM. A small incision was made between the nape of the neck to just anterior of the occipital bone following removal of hair. Forceps and scalpel were used to expose dura mater around the CM (Figure 5.1). Any blood was cleared using sterile cotton tip swabs. A pulled glass capillary tube (Sutter Instruments, O.D.:

1.2mm, I.D.: 0.69mm 10cm length) slowly injected 7µL of AAV9-HSPB1

(3.3e+12vg/kg), AAV9-SMN (3.3e+12vg/kg) or PBS solution into CM. Following

CM injection, three to four sterile surgical staples were used to close the incision.

Mice recovered on heating pads before being returned to their home cage. The entire procedure lasted no longer than 10 minutes.

Figure 5.1 Exposed cisterna magna

Anesthetized mouse was prepared for cisterna magna (CM) injection by creating a small incision from the nape of the neck to the base of the occipital bone. Blunt dissection was performed to expose dura mater overlaying the CM (yellow arrow).

128

5.3.4 RECOVERY TIMELINE

The MU recovery timeline following sciatic nerve crush is shown in Figure

5.2. Baseline behavioral and electrophysiological measurements were made prior to PNI. Behavioral and electrophysiological measurements were then performed at 7dpi, 14dpi, 18dpi, 21dpi, 25dpi, 28dpi, 35dpi, 42dpi, and 60dpi.

Muscle samples were harvested at 0dpi and 60dpi to perform NMJ imaging and quantification.

Figure 5.2 Longitudinal recovery timeline following PNI and CM injection

Baseline electrophysiological and behavioral measurements were made prior to sciatic nerve crush and cisterna magna (CM) injection. CM injection was performed immediately following nerve injury, while mouse was still anesthetized. Longitudinal electrophysiological and behavioral measurements were performed out to 60dpi. Mice were sacrificed at the conclusion of the study for pathology.

129

5.3.5 BEHAVIOR

Behavioral assessments were performed prior to performing electrophysiological measurements to prevent any negative effects of residual isoflurane. The average of four ipsilateral and contralateral hindlimb grip strengths were measured using a Chatillon DEF2-002 grip strength meter

(Columbus Instruments). Mice were positioned to allow only the right or left hindpaw to grasp a grid attached to a force tension sensor and were pulled toward the evaluator away from the grid until the mouse had lost grip.

5.3.6 ELECTROPHYSIOLOGY

Compound muscle action potential (CMAP) and motor unit number estimates (MUNE) were recorded from the ipsilateral hindlimb using previous parameters (Arnold, Sheth et al. 2015). An active ring electrode was placed over the ipsilateral triceps surae with a reference ring electrode placed over the metatarsals of the ipsilateral hindpaw. The sciatic nerve was stimulated (0.1ms pulse, 1-10mA intensity) using two 28G insulated monopolar needles. Peak

CMAP amplitude was recorded using the baseline-to-peak amplitude following maximal stimulation. The distance between the positive wave and negative wave peaks was used to determine the “peak to peak” distance. The sciatic nerve was then stimulated submaximally in increments of 0.3mA until 10 distinct stepwise incremental submaximal responses were recorded. The average of these 10 incremental responses provided an estimate of the single motor unit potential

(SMUP) (min: 25µV; max: peak-to-peak value of recorded CMAP). MUNE was calculated by dividing the average SMUP into the peak-to-peak CMAP amplitude.

130

5.3.7 PERFUSION AND TISSUE PROCESSING

Mice were deeply anesthetized at 5% isoflurane. A transcardial PBS (1X) perfusion was then performed when a toe-pinch response was absent. Spinal cord and soleus muscle samples were then collected for further immunofluorescence and western blot processing.

5.3.8 IMMUNOFLUORESCENCE

5.3.8.1 AAV9 TRANSDUCTION

4% fixed spinal cord tissues were cryopreserved in 30% sucrose overnight at 4°C, after which they were embedded in optimal cutting temperature (OCT) medium and gradually frozen using liquid N2-cooled isopentane. 20μm thick sections were collected at 150μm intervals between L3-L4 spinal segments, to assess motor neuron pools that innervate hindlimb muscles (Nicolopoulos-

Stournaras and Iles 1983). Spinal cord sections were incubated for one hour at room temperature (RT) in blocking buffer (10% donkey serum/4% BSA/0.1%

Triton-X 100/PBS). Primary antibody incubation was performed at RT for 2 hours using α-ChAT (Millipore, AB144P, goat [1:50]), and α-GFP (Abcam, Ab16901, chicken [1:500]). Samples were then washed three times, 10min each, with PBS at RT. A one hour secondary α-goat-594 (Millipore, A11058, donkey [1:1,000]) and α-chicken-488 (Jackson ImmunoResearch, 703-545-155, donkey [1:1,000]) antibody incubation was followed by three, 10min PBS washes. Samples were mounted in Fluoromount-G (Southern Biotech). Segments were visualized at 20x magnification using a Yokogawa spinning disk confocal microscope (Yokogawa

CSU-W1) with Metamorph software (version 7.8.1.0). Total ChAT-labelled motor

131 neurons transduced with GFP were quantified by hand using FIJI imaging software.

5.3.8.2 NMJ QUANTIFICATION

Teased soleus muscle fibers were blocked for 2 hour at RT in blocking buffer (10% goat serum, 4% BSA, 3.0% TritonX-100), then incubated in NF-200

(Abcam, ab72996, chicken [1:5000]) primary antibody solution for overnight at

4°C. Samples were washed 3X in PBS, 10min each, then incubated in α-

Chicken-594 secondary antibody (ThermoFisher, A11042, goat [1:1000]) and α-

Bungarotoxin-488 (ThermoFisher, B13422, [1:1000]) for 2hr at RT. Samples were washed 3X in PBS, 10min each, then mounted under a glass slide with

Fluoromount-G (Southern Biotech). Samples were visualized at 40x magnification using a Leica confocal microscope (Leica DM IRE2) with Leica software (version 2.1). Images were viewed in FIJI (LOCI, University of

Wisconsin-Madison) to quantify NMJ innervation, the co-labeling of NF-200 and

α-Bungarotoxin. 110-150 NMJs per muscle sample (per mouse) were scored as fully innervated, partially innervated or denervated.

5.3.9 WESTERN BLOT

Spinal cord samples were homogenized in buffer (62.5mM tris-HCl pH6.8,

10% SDS, 5mM EDTA) at RT. Protein concentrations were determined using a

Nanodrop (ND-1000). 10μg of each sample was separated by SDS-PAGE and transferred to a PVDF membrane. Blots were incubated in blocking buffer (3%

BSA/1X TBS) for one hour at RT. A 1 hour rocking incubation at RT was performed with α-tubulin (Abcam, ab7291, mouse [1:10,000]), α-SMN (BD

132

Biosciences, 610647, mouse [1:1,000]) or α-HSPB1 (Abcam, ab2790, mouse

[1:5,000]) primary antibodies. Blots were then washed three times in 1X PBS +

0.2% Tween-20, 5min per wash. A one hour α-mouse secondary antibody incubation at RT followed (Licor, 926-32210, goat [1:10,000]). Blots were imaged by an infrared imaging system (Licor) and processed using Odyssey software

(version 3).

5.3.10 STATISTICS

All statistics were performed using Graphpad Prism software (version 6).

Repeat-measure two-way ANOVA was performed to test for significance in longitudinal CMAP, MUNE, SMUP, and grip strength recovery between cohorts.

Standard t-test was performed to determine significance of NMJ denervation.

Significance was set at p<0.05.

5.4 RESULTS

5.4.1 NERVOUS TISSUE TRANSFECTED FOLLOWING CM INJECTION

We first checked for successful transfection of the spinal cord following cisterna magna injection (Figure 5.3). Western blot demonstrated increased human HSPB1 and SMN expression in spinal cord homogenate 7 days post- injection (Figure 5.3A). Mice receiving a CM injection with AAV9-GFP were used to determine whether relevant cells, motor neurons, were transfected (Figure

5.3B). 27.1% ± 5.2% of quantified ChAT positive motor neurons co-stained with

GFP.

133

Figure 5.3 Spinal cord transduction following AAV9 cisterna magna injection

(A) Western blot of spinal cord homogenates from positive control mice overexpressing HSPB1 (+ control), mice receiving a cisterna magna injection with AAV9-HSPB1, AAV-SMN, or PBS (sham). (B) Representative immunofluorescence image of lumbar spinal cord segment confirming GFP expression in the ventral horn (yellow). Scale bar 50μm.

5.4.2 AAV9 OVEREXPRESSION OF SMN OR HSPB1 DOES NOT

ACCELERATE MU CONNECTIVITY FOLLOWING PNI

We next investigated the therapeutic efficacy of AAV9-HSPB1 or AAV9-

SMN immediately delivered after a sciatic nerve crush. Delivering either AAV9-

HSPB1 or AAV9-SMN following nerve crush did not accelerate MU reconnectivity

(Figure 5.4). HSPB1 overexpressing mice demonstrated similar CMAP recovery as PBS-1 sham mice, both cohorts recovering to baseline by 32dpi (31.7mV ±

4.57mV vs 28.8mV ± 7.50mV) (Figure 5.4A). There was no difference in MUNE recovery between HSPB1 overexpressing mice and PBS-1 sham mice, with recovery to baseline occurring by 35dpi (231 ± 110 vs 251 ± 114) (Figure 5.4C).

134

SMN overexpressing mice and PBS-2 sham mice also demonstrated similar

CMAP recovery, reaching pre-injury baseline levels by 32dpi (30.6mV ± 5.52mV vs 27.7mV 6.38mV) (Figure 5.4B). MUNE recovered to baseline by 35dpi in both

SMN overexpressing and PBS-2 sham mice (239 ± 98 vs 225 ± 78) (Figure

5.4D).

Figure 5.4 No accelerated MU connectivity recovery following delivery of AAV9 vectors

(A) Longitudinal CMAP recovery (mV) of AAV9-HSPB1 cohort (blue line, n = 10) compared to PBS-1 cohort (red line, n = 10). (B) Longitudinal CMAP recovery (mV) of the AAV9-SMN cohort (blue line, n = 10) relative to recovery in the PBS- 2 cohort (red line, n = 10). (C) Longitudinal average MUNE recovery in AAV9- HSPB1 and PBS-1 cohorts. (D) Longitudinal average MUNE recovery in AAV9- SMN and PBS-2 cohorts. Error bars denote standard deviation.

135

5.4.3 AAV9-SMN TREATED MICE DEMONSTRATED MILD HINDLIMB GRIP

STRENGTH INCREASE LATE IN RECOVERY TIMELINE

We further assessed behavioral functional recovery in mice using hindlimb grip strength (Figure 5.5). To account for mice continuing to grow, ipsilateral raw grip strength and ipsilateral grip strength normalized to body mass are reported.

Mice overexpressing HSPB1 did not exhibit accelerated raw grip strength recovery when compared to PBS-1 sham mice, with both cohorts returning to baseline by 14dpi (0.025kg ± 0.008kg vs 0.020kg ± 0.006kg) (Figure 5.5A). SMN overexpressing mice also did not demonstrate accelerated raw grip strength recovery relative to PBS-2 sham mice, both cohorts returning to pre-injury grip strength at 14dpi (0.024kg ± 0.008kg vs 0.021kg ± 0.008kg) (Figure 5.5B). At

35dpi and 42dpi AAV9-SMN treated mice appeared to demonstrate stronger raw hindlimb grip strength than PBS-2 sham mice (0.047kg ± 0.006kg vs 0.038 ±

0.006kg and 0.055kg ± 0.01kg vs 0.046kg ± 0.007kg, respectively), which t-tests revealed to be significant. This difference was still observed when ipsilateral grip strength was normalized to body weight (Table 5.1). No such effect was observed in mice treated with AAV9-HSPB1. To rule out the possibility that this difference was on account of PBS sham cohort, AAV9-SMN treated mice grip strengths were compared to PBS-1 sham mice (cohort of mice randomly selected to compare with AAV9-HSPB1 treated mice). SMN overexpressing mice still exhibited increased hindlimb grip strength when compared to the PBS-1 sham cohort.

136

Figure 5.5 Post-crush delivery of AAV9-HSPB1 or AAV9-SMN did not accelerate grip strength recovery to pre-injury level

(A) Longitudinal ipsilateral hindlimb grip strength recovery (kg) in AAV9-HSPB1 (blue line, n = 10) and PBS-1 (red line, n = 10) cohorts. (B) Ipsilateral hindlimb grip strength recovery in AAV9-SMN (blue line, n = 10) and PBS-2 (red line, n = 10) cohorts. Error bars denoted standard deviation. T-tests were performed at 35dpi and 42dpi to test for significance. * = p<0.05.

137

Normalized Grip Strength (±StDev) dpi AAV9-HSPB1 PBS-1 AAV9-SMN PBS-2 1.53E-3 1.58E-3 1.76E-3 1.55E-3 0 (3.6E-4) (2.99E-4) (4.19E-4) (2.81E-4) 9.0E-5 0.0 5.0E-5 4.0E-5 7 (1.7E-4) (0.0) (1.12E-4) (7.47E-5) 1.05E-3 8.6E-4 1.05E-3 8.9E-4 14 (3.46E-4) (2.72E-4) (3.71E-4) (3.55E-4) 1.44E-3 1.28E-3 1.45E-3 1.21E-3 18 (3.92E-4) (4.51E-4) (5.37E-4) (4.93E-4) 1.43E-3 1.36E-3 1.65E-3 1.44E-3 21 (3.57E-4) (3.92E-4) (4.86E-4) (4.49E-4) 1.58E-3 1.41E-3 1.65E-3 1.56E-3 25 (3.0E-4) (3.19E-4) (4.67E-4) (3.6E-4) 1.57E-3 1.57E-3 1.65E-3 1.53E-3 28 (3.72E-4) (1.9E-4) (4.08E-4) (3.0E-4) 1.63E-3 1.58E-3 1.76E-3 1.5E-3 32 (3.88E-4) (4.08E-4) (3.84E-4) (3.3E-4) 1.59E-3 1.58E-3 1.83E-3 * 1.43E-3 35 (2.56E-4) (2.03E-4) (2.71E-4) (2.67E-4) 1.89E-3 1.72E-3 2.08E-3 * 1.71E-3 42 (3.09E-4) (3.19E-4) (3.78E-4) (2.9E-4) 2.06E-3 1.80E-3 2.02E-3 1.86E-3 60 (3.69E-4) (5.09E-4) (2.87E-4) (5.25E-4)

Table 5.1 Ipsilateral hindlimb grip strength recovery normalized to body weight

Mice receiving AAV9-SMN demonstrated significantly increased normalized grip strength at 35dpi and 42dpi compared to both PBS-1 and PBS-2 cohorts. n = 10 for each cohort. T-tests were performed at 35dpi and 42dpi to test for significance. * = p<0.05.

5.4.4 NMJ REINNERVATION RECOVERED BY 60DPI

Quantification of NMJ reinnervation in soleus muscles was performed at the end of the study (Figure 5.6). There was no discernable difference in NMJ reinnervation between HSPB1 overexpressing mice and PBS-1 sham mice

(86.9% ± 4.63% vs 90.9% ± 2.31%, p>0.05) (Figure 5.6A&C). Additionally, NMJ

138 reinnervation was similar between SMN overexpressing mice and PBS-2 sham mice (88.3% ± 6.7% vs 87.2% ± 3.3%, p>0.05) (Figure 5.6B&D).

139

Figure 5.6 Reinnervation is complete by 60dpi

(A) Representative images of reinnervated NMJs in teased soleus muscles at Continued

140

60dpi from PBS-1 (a-c) and AAV9-HSPB1 (d-f) mice. (B) Representative images of reinnervated NMJs in teased soleus muscles at 60dpi from PBS-2 (g-i) and AAV9-SMN (j-l) mice. (C-D) Percent NMJ reinnervation, co-labelled NF-200 (red) and α-Bungarotoxin (green), between PBS-1 and AAV9-HSPB1 mice (C) or PBS- 2 and AAV9-SMN mice (D). Error bars denote standard deviation. n = 4 – 6 per cohort. Images taken at 20x, scale bar = 50µm.

5.5 DISCUSSION

Patient motor and strength recovery following a PNI is dependent on the severity (compression, crush or complete transection) and where along the nerve

(distal or proximal) the injury occurs. Patients with more severe injuries typically experience incomplete motor and strength recovery, even with surgical intervention (Ruijs, Jaquet et al. 2005). One proposed explanation for a lack of recovery is because axonal regrowth in patients is too slow and missing a critical window of NMJ reinnervation (Ma, Omura et al. 2011). A therapeutic that can accelerate MU recovery pre-clinically could perhaps improve patient recovery outcome. Here, two potential therapeutics—AAV9-HSPB1 and AAV9-SMN— were assessed to test our hypothesis that HSPB1 or SMN overexpression after a

PNI would improve MU recovery. Post-injury delivery of either of viral vectors did not accelerate MU recovery to baseline levels relative to PBS sham control mice.

It is possible that overexpressing HSPB1 or SMN via AAV9 may not be a feasible therapeutic, however there are several issues that should be addressed before making that conclusion, including effect of motor neuron transduction efficiency and potential protective effects being masked by the mild PNI.

141

5.5.1 AAV9-SMN AND AAV9-HSPB1 DO NOT ACCELERATE MU RECOVERY

FOLLOWING PNI

We reported relatively few lumbar motor neurons, about 25%, targeted following CM-injection of AAV9-GFP (at a dose of 3.3E+12 vg/kg). One alternative possibility accounting for an absence of accelerated MU recovery is that not enough motor neurons were transfected to have a meaningful effect. It has previously been reported that motor neuron transduction efficiency can be increased by increasing the viral dosage delivered (Meyer, Ferraiuolo et al.

2015). Meyer and colleagues reported that compared to PBS sham mice, the minimum effective dose of AAV9-SMN, 1.8E+13 vg/kg, extended median survival from less than 20 days to about 160 days in SMA pups, and transfected about

40% of lumbar motor neurons. Furthermore, the largest effective viral dose,

3.3E+13 vg/kg, was able to further extended median survival to about 280 days and transfected about 70% of lumbar motor neurons (Meyer, Ferraiuolo et al.

2015). A future study could investigate potential accelerated MU recovery using a

“high-dose” cohort—where by increasing the viral dose delivered post-PNI, more motor neurons could be targeted to express the potentially neuroprotective proteins.

5.5.2 NEUROPROTECTIVE EFFECTS MAY BE MASKED BY NATURALLY

ROBUST RECOVERY IN MICE

Alternatively, our findings may relate to the PNI severity used in our study.

An interesting challenge to using mice for assessing the therapeutic efficacy of

AAV9-HSPB1 and AAV9-SMN is the mouse’s already robust nerve recovery

142 following injury. Previous studies have demonstrated complete behavioral recovery in mice following a sciatic nerve crush (Ma, Omura et al. 2011), which is not typically observed in patients (Ruijs, Jaquet et al. 2005, Navarro 2016, Kemp,

Cederna et al. 2017). This is thought to be due to the differences in axonal regrowth (humans: approximately 1mm/day; mice: 2-3mm/day) (Christie and

Zochodne 2013, Gordon 2016) in addition to the longer distances re-growing axons must traverse in humans. It is possible that a mild, earlier protective effect of either HSPB1 or SMN overexpression was masked by the naturally robust recovery in mice following a less severe nerve injury. Future studies investigating earlier times of underlying pathology, such as NMJ quantification, are also necessary to rule out possible masked effects. It is interesting that we did demonstrate a mild increase in hindlimb grip strength later in the time course in mice overexpressing SMN. However, given the potential unreliability of behavioral grip strength measurements in mice (as it may sometimes be affected by motivation), we must exercise caution before we claim some positive effect of

SMN overexpression. At the same time, this unreliability with grip strength may falsely exclude any potential positive effects. Assessing muscle strength independent of potential confounding variables such as animal motivation, such as with contractility assays, may be a more reliable approach to analyze muscle strength following a PNI.

Finally, utilizing a more severe injury model—such as a complete sciatic nerve transection or repeated sciatic nerve crushes (Ma, Omura et al. 2011,

Sakuma, Gorski et al. 2016)—where motor recovery is often incomplete could

143 provide a more significant insult to better assess any potential therapeutic efficacy of AAV9-HSPB1 or AAV9-SMN treatments. Longitudinally assessing the efficacy of AAV9-HSPB1 or AAV9-SMN in a more severe injury paradigm would also be more clinically relevant, reflecting the typical poor motor function recovery demonstrated in PNI patients.

144

CHAPTER 6: CONCLUSION

6.1 CONCLUSION

The final common pathway of the neuromuscular system, the motor unit

(MU), is comprised of an α-motor neuron and all the muscle fibers it innervates.

The integrity of the MU can be stressed through various neuropathological insults; this dissertation investigated two—amyotrophic lateral sclerosis (ALS) and peripheral nerve injury (PNI). While ALS is a lethal motor neuron disease and some functional MU recovery can occur following a PNI, both ALS and PNI result in denervation, collateral sprouting, reinnervation and axonal degeneration, occurring over a period of time. These dynamic responses at the MU level can be measured using electrophysiological (MU connectivity) or muscle physiological

(muscle contractility) assessments. The relationship of declined muscle electrical excitation in relation to declined force production during neuropathological stress is unknown. To address this, we developed and refined methodologies to longitudinally measure MU integrity during neuropathological stress. This provided key insights towards the connectivity and contractility relationships in

ALS and PNI models—contractility decline preceded connectivity decline in ALS and contractility recovery occurred before connectivity recovery after a PNI.

145

Ultimately, we utilized these paradigms to assess potential therapeutics towards preserving MU integrity.

Chapter two of this work investigated the relationship between MU connectivity and muscle contractility in SOD1(G93A) mice. One obstacle to understanding connectivity and contractility decline in SOD1(G93A) mice had been the difficulty in performing these assessments in the same mouse over time to define the natural history of the model. Previous groups have investigated MU function in neuropathological disease models, however, such reports have utilized highly invasive procedures. Hegedus and colleagues demonstrated fast- twitch MU loss around post-natal day 40 (P40), and Mancuso and colleagues measured CMAP decline by P56—both of which preceded behavioral decline around P90 (Hegedus, Putman et al. 2007, Mancuso, Osta et al. 2014). While these cross-sectional studies can examine aggregated differences over time, they are not able to report individualized disease trajectories. We have previously reported the application of longitudinal connectivity and contractility in vivo assessments in mouse models of neuropathy (Srivastava, Renusch et al. 2012), spinal muscular atrophy (Arnold, Porensky et al. 2014, Arnold, McGovern et al.

2016) and aging (Sheth, Iyer et al. 2018), which we then applied in a natural history of SOD1(G93A) mice.

The pathophysiological origins in ALS are currently debated, with one proposal arguing that pathology arises proximally at the motor neuron cell body and ultimately progresses anterogradely (“dying-forward”) until muscle denervation and motor neuron death occur (Kiernan, Vucic et al. 2011).

146

Glutamate toxicity at the motor neuron cell body, triggering toxic levels of intracellular Ca2+ and cellular damage, is one such proposed mechanism for pathology in ALS (Van Den Bosch, Van Damme et al. 2006). Alternatively, another proposal argues pathology arises distally at the NMJ and progresses retrogradely towards the cell body (“dying-backward”) (Dadon-Nachum, Melamed et al. 2011). In this hypothesis, pathology at the NMJ or muscle arises before motor neuron death is observed. While it is possible that the dying-forward and dying-backward hypotheses are not mutually exclusive, our results in chapter two possibly suggests a dying-backward mechanism. We demonstrated muscle weakness—via diminished muscle twitch and tetanic torque—despite healthy motor neuronal input, as measured by MUNE and CMAP. We would propose future studies in animal models where SOD1(G93A) is restricted to skeletal muscle, discussed in more detail below, to further advance our understanding of how pathology arising in skeletal muscle may eventually result in motor neuron degeneration.

Since our muscle twitch and tetanic torque contractility assays relied on tibial nerve stimulation and muscle excitation, an important consideration is that contractile decline can result from NMJ dysfunctions or a problem arising after excitation of the muscle, via excitation-contraction decoupling—where despite normal electrical excitation the muscle is unable to contract. Our results suggest that early weakness was arising as a result of some muscle specific pathology and not NMJ instability or loss of muscle fiber excitation. Both twitch and CMAP responses are measured following a single, brief, supramaximal excitation of all

147

NMJs. However, loss of muscle twitch torque was observed while the CMAP remained unchanged.

To further consider and address this limitation, we applied single fiber electromyography (SFEMG) to directly analyze NMJ transmission integrity at P35 in SOD1(G93A) and wildtype males (when mutant male mice demonstrated decline contractile strength). SFEMG has previously been demonstrated to detect early increased NMJ transmission failure in ALS patients (Cui, Liu et al.

2004). We hypothesized that if muscle weakness at P35 was independent of the functional integrity of the NMJ, then NMJ transmission would not differ at P35 between mutant and wildtype males. Indeed, we did not observe a difference in jitter (the response latency at the NMJ) between mutant and wildtype males, arguing in support of some excitation-contraction decoupling pathology in the muscle. Further molecular studies—such as altered gene expression, protein activity or mitochondrial activity—can then be performed to elucidate the underlying pathological events leading to excitation-contraction decoupling and muscle weakness and help refine therapeutic intervention strategies.

One proposed future direction to assess how mutation in SOD1 potentially affects the excitation-contraction coupling pathway to reduce muscle contractility and MU connectivity would be to use a transgenic mouse model where

SOD1(G93A) overexpression is limited to skeletal muscle. Dobrowolny and colleagues have developed a transgenic mouse line with human SOD1(G93A) under a myosin light chain (MLC) promoter (MLC-SOD1(G93A)) limits mutant

SOD1 expression to skeletal muscle at levels comparable to the ubiquitous

148

SOD1(G93A) mouse model (Dobrowolny, Aucello et al. 2008). MLC-

SOD1(G93A) demonstrate muscle weakness and NMJ fragmentation around 4 months of age. However, critical questions with respect to proteins involved in excitation-contraction coupling remain—are expression levels altered, what is their functional status and how might these contribute to early contractile decline? Addressing these questions may help define potential muscle therapeutic targets in ALS patients.

Chapter three of this dissertation demonstrates the importance of a well characterized disease time course presented in chapter two. Chapter three assesses the neuroprotective efficacy of HSPB1 overexpression after postnatal day 1 (PND1) AAV9-HSPB1 in SOD1(G93A) mice. We were interested in the potential therapeutic efficacy of HSPB1 overexpression in part because of its protection in other neurodegenerative models, such as peripheral nerve injury

(Ma, Omura et al. 2011), Alzheimer’s disease (Bjorkdahl, Sjogren et al. 2008) and Parkinson’s disease (Renkawek, Stege et al. 1999) as well as in previous work from our lab using in vitro models of ALS (Heilman, Song et al. 2017).

Additionally, our group is interested in that, unlike other neuroprotective proteins when overexpressed, mutations in HSPB1 can give rise to the neuropathic disease Charcot-Marie-Tooth disease/distal hereditary motor neuropathy—a non- lethal disease characterized by distal muscle weakness, abnormal gait and diminished CMAP (Srivastava, Renusch et al. 2012).

We delivered AAV9-HSPB1 in P1 SOD1(G93A) mice (C57Bl6/SJL) and hypothesized that mice overexpressing HSPB1 would demonstrate delayed ALS

149 pathology, as measured by MU connectivity and behavioral measurements. This study was performed prior to our thorough MU characterization during disease progression. As such, we measured MU connectivity only at P60, P90 and

P120—corresponding to “pre-symptomatic”, “early symptomatic”, and “disease endstage” (Gurney 1994, Chiu, Zhai et al. 1995). However, based off our longitudinal characterization of SOD1(G93A) mice, we know these data points may not accurately capture critical disease time points. If AAV9-HSPB1 had a mild protective effect for MU connectivity, it is possible we may have missed it on account of our limited window of assessments. Alternatively, as these studies were performed in C57Bl6/SJL background of SOD1(G93A) mice, the same as

Krishnan and colleagues who similarly report absent neuroprotective effect, it is possible some unknown strain-specific effect is muting a potential effect.

Background strain differences and their potential to mask therapeutic effect is one of several practical limitations in SOD1(G93A) mice, discussed further below.

SOD1(G93A) UTILITY AS A MODEL FOR ALS

The SOD1(G93A) mouse model was utilized in chapters two and three to investigate neuronal connectivity and muscle contractility during disease progression and assess the therapeutic efficacy of HSPB1 overexpression.

Despite this animal model’s role in the field, as well as in this dissertation, there are several drawbacks that must be considered. These drawbacks include the underlying genetics of the model in addition to the field’s lack of viable clinical treatments.

150

The first drawback of this model is the discrepancy between mice and humans in mutant SOD1 protein expression levels. The SOD1(G93A) model utilized here is the “high-copy number” transgenic line (about 25 copies), where unlike in human patients, the mutant protein is markedly overexpressed

(Acevedo-Arozena, Kalmar et al. 2011). This raises the first issue with the model—to what extent is the disease phenotype a result of mutant protein expression versus simple toxicity from SOD1 expression levels. The possibility of toxic protein levels, and not the mutation itself, as the cause of disease is apparent in transgenic mice overexpressing wildtype human SOD1 (wtSOD1 mice) (Graffmo, Forsberg et al. 2013). These wtSOD1 mice also contain 25 copies of human SOD1 and also develop ALS symptoms, suggesting toxic aggregation of SOD1 and not the mutation itself may give rise to the disease. We cannot rule out that the early muscle specific disease phenotype discussed in chapter two was due to high SOD1 expression and not a relevant phenotype.

Future studies utilizing models with lower/biologically relevant SOD1 expression levels can elucidate whether our finding is a relevant ALS phenotype.

A second consideration that one must consider when designing experiments with the SOD1(G93A) mouse is potential limited applicability to human patients. About 90% of ALS cases arise spontaneously (sporadic, or sALS) with an inherited genetic cause in the remaining 10% (familial, or fALS)

(Al-Chalabi and Hardiman 2013). Mutations in SOD1 comprise between 10-20% of fALS patients, or 1-2% of all ALS patients (Al-Chalabi and Hardiman 2013).

This begs the question—will future therapies developed using the SOD1(G93A)

151 mouse model be applicable to human patients with ALS that is not the result of a

SOD1 mutation? Therapies that address the potential toxic effect of mutant

SOD1 in mice may not be a viable paradigm for human clinical studies. Indeed, this problem of limited applicability as well as over expression of the mutant protein leads to the third potential pitfall with the SOD1(G93A) model—minimal development of effective therapeutics.

The ALS field has developed numerous therapies that improve lifespan and delay onset of symptoms in SOD1 mutant mice. However, with the exception of riluzole and edaravone—both having minimal efficacy—no therapies have been strikingly effective in human patients (Petrov, Mansfield et al. 2017). Thus, this discrepancy—efficacious therapies in SOD1(G93A) mice not having effects in human patients—suggests a limitation in the applicability of the model in developing therapeutics for broader causes of ALS in human patients.

Regardless of the aforementioned limitations of the SOD1(G93A) mouse model, we still elected to utilize this model in chapters two and three of this dissertation. The high-copy number line of SOD1(G93A) mice we utilized still demonstrates a hallmark characteristic of ALS—muscle denervation and the accompanying muscle weakness. It is arguable that our understanding of these hallmark characteristics may result in refining the potential target sites for therapies. For example, in chapter two we demonstrated that in both male and female mutant SOD1 mice, muscle strength (measured by twitch and tetanic torque contractility) weakness preceded motor neuron connectivity decline (as measured by compound muscle action potential and motor unit number

152 estimates) by 7 and 21 days, respectively—suggesting early pathology is occurring downstream of the NMJ. Our characterization of MU integrity suggests that the muscle may be an important therapeutic target site. Through our ability to directly assess both muscle contractile and neuronal connectivity, we could also address the issue of whether a therapy that improves muscle weakness also improves neuronal connectivity.

Similarly, this idea of refining potential target sites and the role of the muscle help us to improve future experiments, especially as it relates to our therapeutic paradigm in chapter three—where AAV9-HSPB1 was applied to specifically target cells of the central nervous system. A future study could investigate the effect of HSPB1 overexpression in the muscle. It is possible that the efficacy of HSPB1 overexpression is apparent in non-motor neuronal cells; we have previously demonstrated in vitro that HSPB1 overexpression in mutant

SOD1(G93A) astrocytes reduces motor neuron cell death in a co-culture system

(Heilman, Song et al. 2017). We would hypothesize that if HSPB1 overexpression is non-cell autonomously protective, then direct muscle injection with AAV9-HSPB1 would alleviate disease symptoms. Alternatively, combining increased astrocyte expression with muscle expression could also be assessed, as we have demonstrated in chapter two that muscle may be an important therapeutic site in the disease.

Lastly, while there are few therapies, the two FDA approved drug treatments in human patients—riluzole and edaravone—have also been applied in the SOD1(G93A) high copy mouse with modest increases in survival (Gurney,

153

Fleck et al. 1998, Ito, Wate et al. 2008). While these drugs extend the lifespan in

SOD1(G93A) mice, the clinically translatable MU connectivity and contractility measurements have not been used to assess how they might impact MU integrity. We hypothesize that in addition to extended lifespan, these therapies would extend the time until connectivity and contractility decline. Riluzole is thought to extend survival in ALS patients by blocking the release of glutamate from neurons, minimizing potential toxicity; edaravone is believed to delay ALS progression through its function as a scavenger of free oxygen radicals (Hugon

1996, Shichinohe, Kuroda et al. 2004). These drugs have been given to ALS patients with genetic mutations other than SOD1(G93A). While the treatments only moderately extend the lives of patients by less than a year, it suggests that despite numerous therapies that work in the animal model but not in humans there remains the possibility that therapeutics applied in SOD1(G93A) can in fact be more broadly applied to ALS patients. As the field continues to develop more transgenic mouse models for ALS (Philips and Rothstein 2015), perhaps an improved practice will be to assess potential therapies in multiple models. Such an approach could address the potential issue of a therapy being limited to a niche mutation and not the broader disease.

PERIPHERAL NERVE INJURY AND POTENTIAL THERAPEUTICS

Chapter four of this dissertation demonstrates a natural history of MU connectivity and contractility recovery following a peripheral nerve injury (PNI).

Studies of MU recovery have previously been performed (Gordon and de

Zepetnek 2016, Gregor, Maas et al. 2018), and an influential study from Fu and

154

Gordon demonstrated that the functional strength recovery of muscles in mice is absent following prolonged muscle denervation (Fu and Gordon 1995). In chapter four we applied our electrophysiological and muscle physiological measurements to address the question of what is the temporal relationship between neuronal connectivity and strength recovery of the MU following a PNI.

Applying this minimally-invasive, novel paradigm in a regenerative model demonstrated earlier muscle strength recovery relative to MU nerve connectivity.

An unaddressed question, however, from chapter four is what accounts for this discrepancy. We propose that one possible explanation is temporal dispersion, which can occur when MUs propagate action potentials asynchronously, resulting in electric phase cancelation and ultimately diminished electrophysiological measurements (Mallik and Weir 2005). It is possible that despite neuronal innervation to the muscle, resulting in measurable contractile force production following nerve stimulation, the myelination status of motor axons is insufficient to produce synchronous action potential propagation, resulting in diminished CMAP amplitude. A future study could address whether temporal dispersion is occurring by performing two-point nerve stimulation. In this approach, CMAP amplitude and latency are recorded following both proximal and distal nerve stimulation (Mallik and Weir 2005). Distal nerve stimulation results in normal amplitude and latency of CMAP, while proximal nerve stimulation results in a decreased amplitude and increased latency of CMAP (relative to distal stimulation). We would expect a decrease in CMAP amplitude as well as increased latency of CMAP response following proximal nerve stimulation at 14

155 days post injury, when contractility recovery first started to diverge from connectivity recovery.

Chapter five assessed the therapeutic efficacy of either AAV9-HSPB1 or

AAV9-SMN following a PNI. PNI accounts for up to 3% of trauma center patients and can have a healthcare cost in the billions of dollars (Evans 2001, Grinsell and Keating 2014). Unlike injuries to the central nervous system, injuries to peripheral nervous system, with the focus on motor neurons in this dissertation, are capable of mounting a regenerative response. While motor neurons are capable of regenerating following a PNI, recovery is dependent on injury severity and location, with patients experiencing incomplete recovery in more severe, proximal injuries (Ruijs, Jaquet et al. 2005). The field has investigated numerous genetic therapy targets to improve recovery such as overexpressing BDNF, NGF,

NT-3 or VEGF (Gao 2016; Hoyng 2014). In chapter five we assessed two other potential therapeutics in nerve recovery—HSPB1 and SMN.

Previously, transgenic mice overexpressing human HSPB1 from birth have demonstrated faster axonal regrowth, improved NMJ reinnervation and ultimately accelerated behavioral recovery following a PNI (Ma, Omura et al.

2011). While this previous study demonstrates the promise of HSPB1 as a neuroprotective protein, we wanted to assess a more clinically relevant scenario wherein a PNI in a wildtype mouse was followed by an intervention to overexpress HSPB1. We hypothesized that utilizing an AAV9 vector to overexpress HSPB1 following a PNI would accelerate motor functional recovery.

156

In addition to HSPB1, we also assessed SMN overexpression as a potential therapeutic. Transgenic adult mice with SMN knocked down appeared behaviorally normal, however, when these mice were stressed with a PNI NMJ reinnervation and behavioral recovery were significantly delayed relative to wildtype mice (Kariya, Park et al. 2008). This suggests that SMN serves some as of yet undefined, necessary function towards MU integrity following a stress.

Extrapolating from the apparent necessity of SMN for MU integrity, we investigated whether SMN was also sufficient for MU integrity--hypothesizing that overexpressing SMN via AAV9 would accelerate MU recovery following a PNI.

Neither of these aforementioned interventions demonstrated accelerated

MU reconnection relative to control animals. Mice treated with AAV9-SMN did demonstrate an increase in hindlimb grip strength later in the recovery time course (after already recovering to baseline pre-injury levels); however given how animal motivation may affect grip strength measurements (Deacon 2013), muscle contractile measurements following nerve stimulation are required to better determine the therapeutic effect. There are other potential issues that should be addressed in future studies. First, as mentioned previously only about

50% of human patients experience a complete return of pre-injury muscle strength following a PNI, however regeneration in untreated wildtype mice following a sciatic nerve crush results in complete recovery (Ma, Omura et al.

2011, Christie and Zochodne 2013, Gordon 2016). We also demonstrated this in chapter four where wildtype mice exhibited approximately 50% recovery in both muscle contractile and behavioral measurements as early as 14dpi. This may be

157 partially due to the shorter distances motor axons must regrow in mice compared to humans or partially due to the faster axonal regrowth in mice relative humans

(Christie and Zochodne 2013). The sciatic nerve crush injury to model regeneration following a PNI may have been a potential problem because potential therapies could be mildly protective, which could be masked by an already robust regenerative response. A more severe PNI, such as complete sciatic nerve transection, not only better models real life poor patient recovery— as mice do not go on to fully recover behaviorally (Ma, Omura et al. 2011,

Sakuma, Gorski et al. 2016)—we would also hypothesize that previously masked mild effects would be resolved and not lost in a robust regenerative response.

Furthermore, the first time-point we assessed MU connectivity following sciatic nerve crush was 7 days post injury. Prior work has shown that MU connectivity is lost 1 to 3 days post-injury when NMJ transmission fails, muscle contractility is reduced, and muscle atrophy ensues (Ma, Shen, Garrett 2007;

Wu, Chawla 2014). One unaddressed question from chapter five is whether treatment with AAV9-HSPB1 or AAV9-SMN has any effect during this early phase after PNI, possibly in delaying these pathologies. A future study could include earlier MU connectivity and muscle contractility measurements to test our hypothesis that treatment with AAV9-HSPB1 or AAV9-SMN delays the loss of connectivity and muscle contractility. Additionally, we could also apply a complete nerve transection to test whether denervation-associated muscle atrophy is reduced following AAV9-HSPB1 or AAV9-SMN treatment.

158

UTILITY OF AAV9 TO DELIVER POTENTIAL THERAPEUTICS

We utilized AAV9 as a tool to overexpress potential therapeutics in models of ALS and PNI. There were several reasons why we decided to utilize this tool for our research strategy. First, our group has previously reported successful application of AAV9 to deliver effective therapeutics in animal models of ALS

(Foust, Salazar et al. 2013) and SMN (Meyer, Ferraiuolo et al. 2015), thus demonstrating this tool’s utility in neurodegenerative models. Similarly, AAV9 has further been successfully utilized to deliver SMN to patients with spinal muscular atrophy (SMA) in a clinical study from our group (Mendell, Al-Zaidy et al. 2017).

Despite this rationale, we must still be cognizant of the potential pitfalls to using

AAV9 as a tool to deliver potential therapeutics, particularly as it relates to our experimental paradigms in chapters three and five.

The first potential pitfall of our strategy relates to the timing of AAV9 delivery in an ALS model and the capacity to translate to a viable clinical therapy.

In chapter three we treated presymptomatic SOD1(G93A) transgenic mice at post-natal day one with an intracerebroventricular injection of AAV9-HSPB1 to overexpress HSPB1. At this age, motor neurons and astrocytes are primarily targeted (Foust, Nurre et al. 2009, Foust, Salazar et al. 2013, Meyer, Ferraiuolo et al. 2015); we similarly demonstrated lumbar motor neuron transduction via

ChAT/HSPB1 antibody co-staining. This approach, however, may have a real- world limitation—any potential therapeutic that utilizes AAV9 as a delivery instrument would be given to ALS patients after symptomatic onset had already occurred. Until the development of a presymptomatic ALS biomarker, capable of

159 diagnosing a patient with ALS prior to symptomatic onset thus permitting earlier therapeutic intervention, a challenge to the field is the development of therapies that can be applied in later stages of disease progression (Petrov, Mansfield et al. 2017). Interestingly, our group has demonstrated that delivering AAV9 later in development primarily targets astrocytes. Furthermore, our lab has recently published that HSPB1 overexpression in SOD1(G93A) mouse derived astrocytes diminishes motor neuron death in an in vitro co-culture system, demonstrating

HSPB1’s potential neuroprotective role in a via a non-cell autonomous route

(Heilman, Song et al. 2017). A future study delaying AAV9-HSPB1 until a later time point in SOD1(G93A) mice, potentially after symptomatic onset, would address whether a similar non-cell autonomous neuroprotective effect is observed in vivo. We’d hypothesize that that targeting SOD1(G93A) astrocytes to overexpress HSPB1 would delay ALS symptom onset by extending motor neuron survival. This alternative approach would also address the issue of clinical translation, where intervention would occur after symptom onset. Such an approach could further be compared to a muscle specific overexpression of

HSPB1—our contractility data from chapter two argues that muscle health, in addition to motor neuron health, should also be considered in therapeutic strategies.

In addition to the potential problem in our approach to using AAV9 in an

ALS model, there is one important consideration we must make to our strategy in chapter five, where we utilized AAV9 to overexpress HSPB1 or SMN following a

PNI. Applying AAV9 results in the long-term overexpression of the genetic cargo

160

(Foust, Salazar et al. 2013). While this would be necessary in a disease that would last the life of a patient, as in ALS, we are unsure what effect constitutive overexpression of HSPB1 or SMN would have following an acute stress, such as

PNI. We should note that our group has demonstrated no adverse effects in transgenic mice overexpressing HSPB1 or SMN (Le, Pham et al. 2005,

Srivastava, Renusch et al. 2012). None-the-less, care still must be given as previous studies have demonstrated that the continual overexpression of a neuroprotective protein may have unintended consequences following PNI. Glial cell-line derived neurotrophic factor (GDNF) can promote motor axon regrowth following a PNI. However, the continued overexpression of GDNF also results in the mis-direction of axonal regrowth, reducing muscle reinnervation (Eggers,

Hendriks et al. 2008, Tannemaat, Eggers et al. 2008, de Winter, Hoyng et al.

2013). A more precise vector could be applied in future studies where the overexpression of a therapeutic following the acute stress (a PNI) is then stopped once the regenerative process is complete and the therapeutic is no longer needed. Previous groups have developed tet-inducible viral vectors, where the gene of interest is expressed only in the presence of tetracycline or doxycycline

(Chtarto, Bender et al. 2003, Vanrell, Di Scala et al. 2011, Shakhbazau, Mohanty et al. 2013), which could be applied in a PNI model.

HUMAN CANDIDATE GENES IN ANIMAL MODELS

While on the topic of how we delivered our candidate genes in neurodegenerative models, a broader question is why we elected to assess human candidate genes in animal models. It is possible that the reason we did

161 not observe accelerated recovery to baseline following overexpression of either human HSPB1 or human SMN is because they are nonfunctional in mice. While the functional status of human HSPB1 or SMN following AAV9 delivery were not investigated in chapters three and five, previous findings suggest the absence of efficacy was not the result of incompatibility of human protein in mice. Both human HSPB1 and SMN have previously been applied in other animal models with success. Human HSPB1 and mouse HSPB1 (mHSPB1) are approximately

85% homologous, furthermore previous studies have demonstrated that human

HSPB1 expressed in mice is still capable of performing its functions, including: chaperoning/protein refolding (Ojha, Masilamoni et al. 2011), regulation of apoptosis (Concannon, Gorman et al. 2003, Sharp, Krishnan et al. 2006), and cytoskeleton stabilization (Ma, Omura et al. 2011, Almeida-Souza, Asselbergh et al. 2013).

Similarly, human SMN and mouse SMN (mSMN) are approximately 80% homologous. Functional similarities between human SMN and mSMN is best observed in mouse models of SMA. Transgenic mice with mSMN knocked out or reduced expression produces animal models of SMA, with phenotypes ranging from embryonic lethal to an average lifespan of about 15 days (Schrank, Gotz et al. 1997, Le, Pham et al. 2005). However, our group has previously demonstrated rescued behavioral phenotype and survival in these transgenic mice when human SMN overexpression is present (Le, Pham et al. 2005, Foust,

Wang et al. 2010, Meyer, Ferraiuolo et al. 2015), suggesting that the human form is able to compensate functionally for the endogenous mSMN.

162

The principal reason we elected to utilize human HSPB1 and SMN in mouse models of MU dysfunction for this dissertation was due to our desire to develop potential therapies for human patients. Using the endogenous proteins in

ALS and PNI animal models would have addressed the question of what neuroprotective effect these proteins have towards MU stability. However, if either AAV9-HSPB1 or AAV9-SMN were to eventually go to clinic for ALS or PNI patients, we will have to do so with the human, and not mouse, gene. Thus, by delivering human genetic cargo to mouse models of MU dysfunction, we are concurrently assessing the cargo’s therapeutic efficacy as well as taking an early small step towards the development of a viable patient therapeutic.

CONCLUDING REMARKS

This work represents the application of minimally-invasive outcome measurements to longitudinally characterize MU neuronal input and functional force production in a comprehensive manner to advance our understanding of two pathophysiological stressors—ALS and PNI. Characterization of the natural history of injury, degeneration, and regeneration of the neuromuscular system helps to resolve the issue of treatment timing and timeframe when pathophysiological processes are occurring. In our studies, muscle contractility decline was demonstrated before MU connectivity decline in an animal model of

ALS. These results suggest that excitation-contraction decoupling events may be an important therapeutic consideration for interventions in ALS. Furthermore, in our PNI studies muscle contractility recovery preceded MU connectivity recovery.

Utilizing our paradigm, we have proposed novel therapeutic interventions to

163 target the potential underlying pathologies of MU dysfunction. We performed a thorough, in-depth assessment of potential therapeutics in two neurodegenerative models. Future pre-clinical studies will be strengthened by incorporating the clinically-relevant outcomes of MU connectivity and muscle contractility that we have developed and refined. Overall, our results have established and highlight the strengths of longitudinally assessment of MU integrity, using connectivity and contractility in tandem.

164

REFERENCES

Acevedo-Arozena, A., B. Kalmar, S. Essa, T. Ricketts, P. Joyce, R. Kent, C. Rowe, A. Parker, A. Gray, M. Hafezparast, J. R. Thorpe, L. Greensmith and E. M. Fisher (2011). "A comprehensive assessment of the SOD1G93A low-copy transgenic mouse, which models human amyotrophic lateral sclerosis." Dis Model Mech 4(5): 686-700. Al-Chalabi, A. and O. Hardiman (2013). "The epidemiology of ALS: a conspiracy of genes, environment and time." Nat Rev Neurol 9(11): 617-628. Alami, N. H., R. B. Smith, M. A. Carrasco, L. A. Williams, C. S. Winborn, S. S. W. Han, E. Kiskinis, B. Winborn, B. D. Freibaum, A. Kanagaraj, A. J. Clare, N. M. Badders, B. Bilican, E. Chaum, S. Chandran, C. E. Shaw, K. C. Eggan, T. Maniatis and J. P. Taylor (2014). "Axonal transport of TDP-43 mRNA granules is impaired by ALS-causing mutations." Neuron 81(3): 536-543. Albuquerque, E. X., E. F. Pereira, M. Alkondon and S. W. Rogers (2009). "Mammalian nicotinic acetylcholine receptors: from structure to function." Physiol Rev 89(1): 73-120. Almeida-Souza, L., B. Asselbergh, V. De Winter, S. Goethals, V. Timmerman and S. Janssens (2013). "HSPB1 facilitates the formation of non- centrosomal microtubules." PLoS One 8(6): e66541. Alvarez, F. J., K. L. Bullinger, H. E. Titus, P. Nardelli and T. C. Cope (2010). "Permanent reorganization of Ia afferent synapses on motoneurons after peripheral nerve injuries." Ann N Y Acad Sci 1198: 231-241. Arai, M., A. Yoguchi, T. Takizawa, T. Yokoyama, T. Kanda, M. Kurabayashi and R. Nagai (2000). "Mechanism of doxorubicin-induced inhibition of sarcoplasmic reticulum Ca(2+)-ATPase gene transcription." Circ Res 86(1): 8-14. Arasaki, K., Y. Kato, A. Hyodo, R. Ushijima and M. Tamaki (2002). "Longitudinal study of functional spinal alpha motor neuron loss in amyotrophic lateral sclerosis." Muscle Nerve 25(4): 520-526. Armand, J. (1982). "The origin, course and terminations of corticospinal fibers in various mammals." Prog Brain Res 57: 329-360. Armon, C. and M. E. Brandstater (1999). "Motor unit number estimate-based rates of progression of ALS predict patient survival." Muscle Nerve 22(11): 1571-1575.

165

Arnold, W., V. L. McGovern, B. Sanchez, J. Li, K. M. Corlett, S. J. Kolb, S. B. Rutkove and A. H. Burghes (2016). "The neuromuscular impact of symptomatic SMN restoration in a mouse model of spinal muscular atrophy." Neurobiol Dis 87: 116-123. Arnold, W. D. and A. H. Burghes (2013). "Spinal muscular atrophy: development and implementation of potential treatments." Ann Neurol 74(3): 348-362. Arnold, W. D., P. N. Porensky, V. L. McGovern, C. C. Iyer, S. Duque, X. Li, K. Meyer, L. Schmelzer, B. K. Kaspar, S. J. Kolb, J. T. Kissel and A. H. Burghes (2014). "Electrophysiological Biomarkers in Spinal Muscular Atrophy: Preclinical Proof of Concept." Ann Clin Transl Neurol 1(1): 34-44. Arnold, W. D., K. A. Sheth, C. G. Wier, J. T. Kissel, A. H. Burghes and S. J. Kolb (2015). "Electrophysiological Motor Unit Number Estimation (MUNE) Measuring Compound Muscle Action Potential (CMAP) in Mouse Hindlimb Muscles." J Vis Exp(103). Bagust, J., D. M. Lewis and R. A. Westerman (1981). "Motor units in cross- reinnervated fast and slow twitch muscle of the cat." J Physiol 313: 223- 235. Bame, M., P. A. Pentiak, R. Needleman and W. S. Brusilow (2012). "Effect of sex on lifespan, disease progression, and the response to methionine sulfoximine in the SOD1 G93A mouse model for ALS." Gend Med 9(6): 524-535. Bang, M. L., T. Centner, F. Fornoff, A. J. Geach, M. Gotthardt, M. McNabb, C. C. Witt, D. Labeit, C. C. Gregorio, H. Granzier and S. Labeit (2001). "The complete gene sequence of titin, expression of an unusual approximately 700-kDa titin isoform, and its interaction with obscurin identify a novel Z- line to I-band linking system." Circ Res 89(11): 1065-1072. Barber, S. C., R. J. Mead and P. J. Shaw (2006). "Oxidative stress in ALS: a mechanism of neurodegeneration and a therapeutic target." Biochim Biophys Acta 1762(11-12): 1051-1067. Barton, M. J., J. W. Morley, M. A. Stoodley, A. Lauto and D. A. Mahns (2014). "Nerve repair: toward a sutureless approach." Neurosurg Rev 37(4): 585- 595. Beal, M. F. (2005). "Mitochondria take center stage in aging and neurodegeneration." Ann Neurol 58(4): 495-505. Bear, M. F., Connors, B.W., and Paradiso, M.A. (2007). Neuroscience Exploring the Brain, Lippincott Williams & Wilkins. Beattie, C. E. and S. J. Kolb (2018). "Spinal muscular atrophy: Selective motor neuron loss and global defect in the assembly of ribonucleoproteins." Brain Res. Beck, M., R. Giess, W. Wurffel, T. Magnus, G. Ochs and K. V. Toyka (1999). "Comparison of maximal voluntary isometric contraction and Drachman's hand-held dynamometry in evaluating patients with amyotrophic lateral sclerosis." Muscle Nerve 22(9): 1265-1270. Ben-Ari, Y., J. L. Gaiarsa, R. Tyzio and R. Khazipov (2007). "GABA: a pioneer transmitter that excites immature neurons and generates primitive oscillations." Physiol Rev 87(4): 1215-1284.

166

Benn, S. C., D. Perrelet, A. C. Kato, J. Scholz, I. Decosterd, R. J. Mannion, J. C. Bakowska and C. J. Woolf (2002). "Hsp27 upregulation and phosphorylation is required for injured sensory and motor neuron survival." Neuron 36(1): 45-56. Beuche, W. and R. L. Friede (1984). "The role of non-resident cells in Wallerian degeneration." J Neurocytol 13(5): 767-796. Bevan, A. K., S. Duque, K. D. Foust, P. R. Morales, L. Braun, L. Schmelzer, C. M. Chan, M. McCrate, L. G. Chicoine, B. D. Coley, P. N. Porensky, S. J. Kolb, J. R. Mendell, A. H. Burghes and B. K. Kaspar (2011). "Systemic gene delivery in large species for targeting spinal cord, brain, and peripheral tissues for pediatric disorders." Mol Ther 19(11): 1971-1980. Bewick, G. S. and R. W. Banks (2015). "Mechanotransduction in the muscle spindle." Pflugers Arch 467(1): 175-190. Bigland, B. and O. C. Lippold (1954). "The relation between force, velocity and integrated electrical activity in human muscles." J Physiol 123(1): 214-224. Bjorkdahl, C., M. J. Sjogren, X. Zhou, H. Concha, J. Avila, B. Winblad and J. J. Pei (2008). "Small heat shock proteins Hsp27 or alphaB-crystallin and the protein components of neurofibrillary tangles: tau and neurofilaments." J Neurosci Res 86(6): 1343-1352. Blasco, H., A. M. Guennoc, C. Veyrat-Durebex, P. H. Gordon, C. R. Andres, W. Camu and P. Corcia (2012). "Amyotrophic lateral sclerosis: a hormonal condition?" Amyotroph Lateral Scler 13(6): 585-588. Block, B. A., T. Imagawa, K. P. Campbell and C. Franzini-Armstrong (1988). "Structural evidence for direct interaction between the molecular components of the transverse tubule/sarcoplasmic reticulum junction in skeletal muscle." J Cell Biol 107(6 Pt 2): 2587-2600. Boillee, S., K. Yamanaka, C. S. Lobsiger, N. G. Copeland, N. A. Jenkins, G. Kassiotis, G. Kollias and D. W. Cleveland (2006). "Onset and progression in inherited ALS determined by motor neurons and microglia." Science 312(5778): 1389-1392. Bondan, E. F., P. R. Custodio, M. A. Lallo, H. D. Bentubo and D. L. Graca (2009). "Ethidium bromide-induced demyelination in the sciatic nerve of diabetic rats." Arq Neuropsiquiatr 67(4): 1066-1070. Bowser, R., M. R. Turner and J. Shefner (2011). "Biomarkers in amyotrophic lateral sclerosis: opportunities and limitations." Nat Rev Neurol 7(11): 631- 638. Brown, A. (2003). "Axonal transport of membranous and nonmembranous cargoes: a unified perspective." J Cell Biol 160(6): 817-821. Brown, M. C. and R. Ironton (1978). "Sprouting and regression of neuromuscular synapses in partially denervated mammalian muscles." J Physiol 278: 325-348. Brown, M. R., P. G. Sullivan and J. W. Geddes (2006). "Synaptic mitochondria are more susceptible to Ca2+overload than nonsynaptic mitochondria." J Biol Chem 281(17): 11658-11668. Bruijn, L. I., M. W. Becher, M. K. Lee, K. L. Anderson, N. A. Jenkins, N. G. Copeland, S. S. Sisodia, J. D. Rothstein, D. R. Borchelt, D. L. Price and D.

167

W. Cleveland (1997). "ALS-linked SOD1 mutant G85R mediates damage to astrocytes and promotes rapidly progressive disease with SOD1- containing inclusions." Neuron 18(2): 327-338. Bruijn, L. I., M. K. Houseweart, S. Kato, K. L. Anderson, S. D. Anderson, E. Ohama, A. G. Reaume, R. W. Scott and D. W. Cleveland (1998). "Aggregation and motor neuron toxicity of an ALS-linked SOD1 mutant independent from wild-type SOD1." Science 281(5384): 1851-1854. Buller, A. J., J. C. Eccles and R. M. Eccles (1960). "Interactions between motoneurones and muscles in respect of the characteristic speeds of their responses." J Physiol 150: 417-439. Burke, R. E. (1967). "Motor unit types of cat triceps surae muscle." J Physiol 193(1): 141-160. Burke, R. E., D. N. Levine, P. Tsairis and F. E. Zajac, 3rd (1973). "Physiological types and histochemical profiles in motor units of the cat gastrocnemius." J Physiol 234(3): 723-748. Calderon, J. C., P. Bolanos and C. Caputo (2014). "Tetanic Ca2+ transient differences between slow- and fast-twitch mouse skeletal muscle fibres: a comprehensive experimental approach." J Muscle Res Cell Motil 35(5-6): 279-293. Capogrosso, R. F., P. Mantuano, A. Cozzoli, F. Sanarica, A. M. Massari, E. Conte, A. Fonzino, A. Giustino, J. F. Rolland, A. Quaranta, M. De Bellis, G. M. Camerino, R. W. Grange and A. De Luca (2017). "Contractile efficiency of dystrophic mdx mouse muscle: in vivo and ex vivo assessment of adaptation to exercise of functional end points." J Appl Physiol (1985) 122(4): 828-843. Carleton, S. A. and W. F. Brown (1979). "Changes in motor unit populations in motor neurone disease." J Neurol Neurosurg Psychiatry 42(1): 42-51. Carunchio, I., C. Mollinari, M. Pieri, D. Merlo and C. Zona (2008). "GAB(A) receptors present higher affinity and modified subunit composition in spinal motor neurons from a genetic model of amyotrophic lateral sclerosis." Eur J Neurosci 28(7): 1275-1285. Chai, R. J., J. Vukovic, S. Dunlop, M. D. Grounds and T. Shavlakadze (2011). "Striking denervation of neuromuscular junctions without lumbar motoneuron loss in geriatric mouse muscle." PLoS One 6(12): e28090. Chan, R. C. and T. C. Hsu (1991). "Quantitative comparison of motor unit potential parameters between monopolar and concentric needles." Muscle Nerve 14(10): 1028-1032. Cheng, A., M. Morsch, Y. Murata, N. Ghazanfari, S. W. Reddel and W. D. Phillips (2013). "Sequence of age-associated changes to the mouse neuromuscular junction and the protective effects of voluntary exercise." PLoS One 8(7): e67970. Cheroni, C., M. Marino, M. Tortarolo, P. Veglianese, S. De Biasi, E. Fontana, L. V. Zuccarello, C. J. Maynard, N. P. Dantuma and C. Bendotti (2009). "Functional alterations of the ubiquitin-proteasome system in motor neurons of a mouse model of familial amyotrophic lateral sclerosis." Hum Mol Genet 18(1): 82-96.

168

Chiu, A. Y., P. Zhai, M. C. Dal Canto, T. M. Peters, Y. W. Kwon, S. M. Prattis and M. E. Gurney (1995). "Age-dependent penetrance of disease in a transgenic mouse model of familial amyotrophic lateral sclerosis." Mol Cell Neurosci 6(4): 349-362. Choi, C. I., Y. D. Lee, B. J. Gwag, S. I. Cho, S. S. Kim and H. Suh-Kim (2008). "Effects of estrogen on lifespan and motor functions in female hSOD1 G93A transgenic mice." J Neurol Sci 268(1-2): 40-47. Christie, K. J. and D. Zochodne (2013). "Peripheral axon regrowth: new molecular approaches." Neuroscience 240: 310-324. Chtarto, A., H. U. Bender, C. O. Hanemann, T. Kemp, E. Lehtonen, M. Levivier, J. Brotchi, T. Velu and L. Tenenbaum (2003). "Tetracycline-inducible transgene expression mediated by a single AAV vector." Gene Ther 10(1): 84-94. Clamann, H. P. (1993). "Motor unit recruitment and the gradation of muscle force." Phys Ther 73(12): 830-843. Clark, B. C., S. B. Cook and L. L. Ploutz-Snyder (2007). "Reliability of techniques to assess human neuromuscular function in vivo." J Electromyogr Kinesiol 17(1): 90-101. Concannon, C. G., A. M. Gorman and A. Samali (2003). "On the role of Hsp27 in regulating apoptosis." Apoptosis 8(1): 61-70. Conway, B. A., D. M. Halliday, S. F. Farmer, U. Shahani, P. Maas, A. I. Weir and J. R. Rosenberg (1995). "Synchronization between motor cortex and spinal motoneuronal pool during the performance of a maintained motor task in man." J Physiol 489 ( Pt 3): 917-924. Cui, L. Y., M. S. Liu and X. F. Tang (2004). "Single fiber electromyography in 78 patients with amyotrophic lateral sclerosis." Chin Med J (Engl) 117(12): 1830-1833. Cullheim, S. (1978). "Relations between cell body size, axon diameter and axon conduction velocity of cat sciatic alpha-motoneurons stained with horseradish peroxidase." Neurosci Lett 8(1): 17-20. d'Ydewalle, C., J. Krishnan, D. M. Chiheb, P. Van Damme, J. Irobi, A. P. Kozikowski, P. Vanden Berghe, V. Timmerman, W. Robberecht and L. Van Den Bosch (2011). "HDAC6 inhibitors reverse axonal loss in a mouse model of mutant HSPB1-induced Charcot-Marie-Tooth disease." Nat Med 17(8): 968-974. Dadon-Nachum, M., E. Melamed and D. Offen (2011). "The "dying-back" phenomenon of motor neurons in ALS." J Mol Neurosci 43(3): 470-477. Damiano, M., A. A. Starkov, S. Petri, K. Kipiani, M. Kiaei, M. Mattiazzi, M. Flint Beal and G. Manfredi (2006). "Neural mitochondrial Ca2+ capacity impairment precedes the onset of motor symptoms in G93A Cu/Zn- superoxide dismutase mutant mice." J Neurochem 96(5): 1349-1361. Dantes, M. and A. McComas (1991). "The extent and time course of motoneuron involvement in amyotrophic lateral sclerosis." Muscle Nerve 14(5): 416- 421.

169

De Luca, C. J., L. D. Gilmore, M. Kuznetsov and S. H. Roy (2010). "Filtering the surface EMG signal: Movement artifact and baseline noise contamination." J Biomech 43(8): 1573-1579. de Noordhout, A. M., G. Rapisarda, D. Bogacz, P. Gerard, V. De Pasqua, G. Pennisi and P. J. Delwaide (1999). "Corticomotoneuronal synaptic connections in normal man: an electrophysiological study." Brain 122 ( Pt 7): 1327-1340. de Winter, F., S. Hoyng, M. Tannemaat, R. Eggers, M. Mason, M. Malessy and J. Verhaagen (2013). "Gene therapy approaches to enhance regeneration of the injured peripheral nerve." Eur J Pharmacol 719(1-3): 145-152. Deacon, R. M. (2013). "Measuring the strength of mice." J Vis Exp(76). Deng, H. X., A. Hentati, J. A. Tainer, Z. Iqbal, A. Cayabyab, W. Y. Hung, E. D. Getzoff, P. Hu, B. Herzfeldt, R. P. Roos and et al. (1993). "Amyotrophic lateral sclerosis and structural defects in Cu,Zn superoxide dismutase." Science 261(5124): 1047-1051. Dibaj, P., E. D. Schomburg and H. Steffens (2015). "Contractile characteristics of gastrocnemius-soleus muscle in the SOD1G93A ALS mouse model." Neurol Res 37(8): 693-702. Dobrowolny, G., M. Aucello, E. Rizzuto, S. Beccafico, C. Mammucari, S. Boncompagni, S. Belia, F. Wannenes, C. Nicoletti, Z. Del Prete, N. Rosenthal, M. Molinaro, F. Protasi, G. Fano, M. Sandri and A. Musaro (2008). "Skeletal muscle is a primary target of SOD1G93A-mediated toxicity." Cell Metab 8(5): 425-436. Downes, L., P. Ashby and J. Bugaresti (1995). "Reflex effects from Golgi tendon organ (Ib) afferents are unchanged after spinal cord lesions in humans." Neurology 45(9): 1720-1724. Duque, S. I., W. D. Arnold, P. Odermatt, X. Li, P. N. Porensky, L. Schmelzer, K. Meyer, S. J. Kolb, D. Schumperli, B. K. Kaspar and A. H. Burghes (2015). "A large animal model of spinal muscular atrophy and correction of phenotype." Ann Neurol 77(3): 399-414. Eccles, J. C., R. M. Eccles and A. Lundberg (1957). "The convergence of monosynaptic excitatory afferents on to many different species of alpha motoneurones." J Physiol 137(1): 22-50. Edwards, J. N., T. R. Cully, T. R. Shannon, D. G. Stephenson and B. S. Launikonis (2012). "Longitudinal and transversal propagation of excitation along the tubular system of rat fast-twitch muscle fibres studied by high speed confocal microscopy." J Physiol 590(3): 475-492. Eggers, R., W. T. Hendriks, M. R. Tannemaat, J. J. van Heerikhuize, C. W. Pool, T. P. Carlstedt, A. Zaldumbide, R. C. Hoeben, G. J. Boer and J. Verhaagen (2008). "Neuroregenerative effects of lentiviral vector-mediated GDNF expression in reimplanted ventral roots." Mol Cell Neurosci 39(1): 105-117. England, J. and S. Loughna (2013). "Heavy and light roles: myosin in the morphogenesis of the heart." Cell Mol Life Sci 70(7): 1221-1239. Enoka, R. M. and J. Duchateau (2017). "Rate Coding and the Control of Muscle Force." Cold Spring Harb Perspect Med 7(10).

170

Evans, G. R. (2001). "Peripheral nerve injury: a review and approach to tissue engineered constructs." Anat Rec 263(4): 396-404. Fabricius, C., C. H. Berthold and M. Rydmark (1993). "Axoplasmic organelles at nodes of Ranvier. II. Occurrence and distribution in large myelinated spinal cord axons of the adult cat." J Neurocytol 22(11): 941-954. Felice, K. J. (1995). "Thenar Motor Unit Number Estimates Using the Multiple Point Stimulation Technique - Reproducibility Studies in Als Patients and Normal Subjects." Muscle & Nerve 18(12): 1412-1416. Ferrucci, M., G. Lazzeri, M. Flaibani, F. Biagioni, F. Cantini, M. Madonna, D. Bucci, F. Limanaqi, P. Soldani and F. Fornai (2018). "In search for a gold- standard procedure to count motor neurons in the spinal cord." Histol Histopathol: 11983. Fischer, L. R., D. G. Culver, P. Tennant, A. A. Davis, M. Wang, A. Castellano- Sanchez, J. Khan, M. A. Polak and J. D. Glass (2004). "Amyotrophic lateral sclerosis is a distal axonopathy: evidence in mice and man." Exp Neurol 185(2): 232-240. Fitts, R. H. (1994). "Cellular mechanisms of muscle fatigue." Physiol Rev 74(1): 49-94. Flores, A. J., C. J. Lavernia and P. W. Owens (2000). "Anatomy and physiology of peripheral nerve injury and repair." Am J Orthop (Belle Mead NJ) 29(3): 167-173. Foust, K. D., E. Nurre, C. L. Montgomery, A. Hernandez, C. M. Chan and B. K. Kaspar (2009). "Intravascular AAV9 preferentially targets neonatal neurons and adult astrocytes." Nat Biotechnol 27(1): 59-65. Foust, K. D., D. L. Salazar, S. Likhite, L. Ferraiuolo, D. Ditsworth, H. Ilieva, K. Meyer, L. Schmelzer, L. Braun, D. W. Cleveland and B. K. Kaspar (2013). "Therapeutic AAV9-mediated suppression of mutant SOD1 slows disease progression and extends survival in models of inherited ALS." Mol Ther 21(12): 2148-2159. Foust, K. D., X. Wang, V. L. McGovern, L. Braun, A. K. Bevan, A. M. Haidet, T. T. Le, P. R. Morales, M. M. Rich, A. H. Burghes and B. K. Kaspar (2010). "Rescue of the spinal muscular atrophy phenotype in a mouse model by early postnatal delivery of SMN." Nat Biotechnol 28(3): 271-274. Frakes, A. E., L. Braun, L. Ferraiuolo, D. C. Guttridge and B. K. Kaspar (2017). "Additive amelioration of ALS by co-targeting independent pathogenic mechanisms." Ann Clin Transl Neurol 4(2): 76-86. Frakes, A. E., L. Ferraiuolo, A. M. Haidet-Phillips, L. Schmelzer, L. Braun, C. J. Miranda, K. J. Ladner, A. K. Bevan, K. D. Foust, J. P. Godbout, P. G. Popovich, D. C. Guttridge and B. K. Kaspar (2014). "Microglia induce motor neuron death via the classical NF-kappaB pathway in amyotrophic lateral sclerosis." Neuron 81(5): 1009-1023. Franzini-Armstrong, C. and A. O. Jorgensen (1994). "Structure and development of E-C coupling units in skeletal muscle." Annu Rev Physiol 56: 509-534. Franzini-Armstrong, C. and K. R. Porter (1964). "Sarcolemmal Invaginations Constituting the T System in Fish Muscle Fibers." J Cell Biol 22: 675-696.

171

Frey, D., C. Schneider, L. Xu, J. Borg, W. Spooren and P. Caroni (2000). "Early and selective loss of neuromuscular synapse subtypes with low sprouting competence in motoneuron diseases." J Neurosci 20(7): 2534-2542. Fu, S. Y. and T. Gordon (1995). "Contributing factors to poor functional recovery after delayed nerve repair: prolonged axotomy." J Neurosci 15(5 Pt 2): 3876-3885. Fu, S. Y. and T. Gordon (1995). "Contributing factors to poor functional recovery after delayed nerve repair: prolonged denervation." J Neurosci 15(5 Pt 2): 3886-3895. Fu, S. Y. and T. Gordon (1997). "The cellular and molecular basis of peripheral nerve regeneration." Mol Neurobiol 14(1-2): 67-116. Fuchtbauer, E. M., A. M. Rowlerson, K. Gotz, G. Friedrich, K. Mabuchi, J. Gergely and H. Jockusch (1991). "Direct correlation of parvalbumin levels with myosin isoforms and succinate dehydrogenase activity on frozen sections of rodent muscle." J Histochem Cytochem 39(3): 355-361. Fuglevand, A. J., R. A. Lester and R. K. Johns (2015). "Distinguishing intrinsic from extrinsic factors underlying firing rate saturation in human motor units." J Neurophysiol 113(5): 1310-1322. Fuglevand, A. J., V. G. Macefield and B. Bigland-Ritchie (1999). "Force- frequency and fatigue properties of motor units in muscles that control digits of the human hand." J Neurophysiol 81(4): 1718-1729. Gawel, M., A. Kostera-Pruszczyk, A. Lusakowska, M. Jedrzejowska, B. Ryniewicz, M. Lipowska, D. Gawel and A. Kaminska (2015). "Motor unit loss estimation by the multipoint incremental MUNE method in children with spinal muscular atrophy--a preliminary study." Neuromuscul Disord 25(3): 216-221. George, R. and J. W. Griffin (1994). "The proximo-distal spread of axonal degeneration in the dorsal columns of the rat." J Neurocytol 23(11): 657- 667. Gerdts, J., Y. Sasaki, B. Vohra, J. Marasa and J. Milbrandt (2011). "Image-based screening identifies novel roles for IkappaB kinase and glycogen synthase kinase 3 in axonal degeneration." J Biol Chem 286(32): 28011-28018. Glat, M. J., F. Benninger, Y. Barhum, T. Ben-Zur, E. Kogan, I. Steiner, D. Yaffe and D. Offen (2016). "Ectopic Muscle Expression of Neurotrophic Factors Improves Recovery After Nerve Injury." J Mol Neurosci 58(1): 39-45. Goetz, C. G. (2000). "Amyotrophic lateral sclerosis: early contributions of Jean- Martin Charcot." Muscle Nerve 23(3): 336-343. Gomes, A. V., J. D. Potter and D. Szczesna-Cordary (2002). "The role of troponins in muscle contraction." IUBMB Life 54(6): 323-333. Gonzalez-Perez, F., J. Hernandez, C. Heimann, J. B. Phillips, E. Udina and X. Navarro (2018). "Schwann cells and mesenchymal stem cells in laminin- or fibronectin-aligned matrices and regeneration across a critical size defect of 15 mm in the rat sciatic nerve." J Neurosurg Spine 28(1): 109- 118. Gooch, C. L. (2017). "The canaries in the coal mine: mune and munix in amyotrophic lateral sclerosis." Muscle Nerve 56(2): 183-184.

172

Gooch, C. L., T. J. Doherty, K. M. Chan, M. B. Bromberg, R. A. Lewis, D. W. Stashuk, M. J. Berger, M. T. Andary and J. R. Daube (2014). "Motor unit number estimation: a technology and literature review." Muscle Nerve 50(6): 884-893. Gooch, C. L. and D. R. Mosier (2001). "Stimulated single fiber electromyography in the mouse: techniques and normative data." Muscle Nerve 24(7): 941- 945. Gooch, C. L. and J. M. Shefner (2004). "ALS surrogate markers. MUNE." Amyotroph Lateral Scler Other Motor Neuron Disord 5 Suppl 1: 104-107. Gordon, T. (2016). "Nerve Regeneration: Understanding Biology and Its Influence on Return of Function After Nerve Transfers." Hand Clin 32(2): 103-117. Gordon, T. and J. E. T. de Zepetnek (2016). "Motor unit and muscle fiber type grouping after peripheral nerve injury in the rat." Exp Neurol 285(Pt A): 24- 40. Gordon, T., R. B. Stein and C. K. Thomas (1986). "Organization of motor units following cross-reinnervation of antagonistic muscles in the cat hind limb." J Physiol 374: 443-456. Gould, T. W., R. R. Buss, S. Vinsant, D. Prevette, W. Sun, C. M. Knudson, C. E. Milligan and R. W. Oppenheim (2006). "Complete dissociation of motor neuron death from motor dysfunction by Bax deletion in a mouse model of ALS." J Neurosci 26(34): 8774-8786. Graffmo, K. S., K. Forsberg, J. Bergh, A. Birve, P. Zetterstrom, P. M. Andersen, S. L. Marklund and T. Brannstrom (2013). "Expression of wild-type human superoxide dismutase-1 in mice causes amyotrophic lateral sclerosis." Hum Mol Genet 22(1): 51-60. Gregor, R. J., H. Maas, M. A. Bulgakova, A. Oliver, A. W. English and B. I. Prilutsky (2018). "Time course of functional recovery during the first 3 mo after surgical transection and repair of nerves to the feline soleus and lateral gastrocnemius muscles." J Neurophysiol 119(3): 1166-1185. Grinsell, D. and C. P. Keating (2014). "Peripheral nerve reconstruction after injury: a review of clinical and experimental therapies." Biomed Res Int 2014: 698256. Gurney, M. E. (1994). "Transgenic-mouse model of amyotrophic lateral sclerosis." N Engl J Med 331(25): 1721-1722. Gurney, M. E., T. J. Fleck, C. S. Himes and E. D. Hall (1998). "Riluzole preserves motor function in a transgenic model of familial amyotrophic lateral sclerosis." Neurology 50(1): 62-66. Gurney, M. E., H. Pu, A. Y. Chiu, M. C. Dal Canto, C. Y. Polchow, D. D. Alexander, J. Caliendo, A. Hentati, Y. W. Kwon, H. X. Deng and et al. (1994). "Motor neuron degeneration in mice that express a human Cu,Zn superoxide dismutase mutation." Science 264(5166): 1772-1775. Gutmann, E. and F. K. Sanders (1943). "Recovery of fibre numbers and diameters in the regeneration of peripheral nerves." J Physiol 101(4): 489- 518.

173

Haidet-Phillips, A. M., M. E. Hester, C. J. Miranda, K. Meyer, L. Braun, A. Frakes, S. Song, S. Likhite, M. J. Murtha, K. D. Foust, M. Rao, A. Eagle, A. Kammesheidt, A. Christensen, J. R. Mendell, A. H. Burghes and B. K. Kaspar (2011). "Astrocytes from familial and sporadic ALS patients are toxic to motor neurons." Nat Biotechnol 29(9): 824-828. Hamm, R. J., B. R. Pike, D. M. O'Dell, B. G. Lyeth and L. W. Jenkins (1994). "The rotarod test: an evaluation of its effectiveness in assessing motor deficits following traumatic brain injury." J Neurotrauma 11(2): 187-196. Hammad, M., A. Silva, J. Glass, J. T. Sladky and M. Benatar (2007). "Clinical, electrophysiologic, and pathologic evidence for sensory abnormalities in ALS." Neurology 69(24): 2236-2242. Hansen, S. and J. P. Ballantyne (1978). "A quantitative electrophysiological study of motor neurone disease." J Neurol Neurosurg Psychiatry 41(9): 773-783. Harper, S. Q., P. D. Staber, X. He, S. L. Eliason, I. H. Martins, Q. Mao, L. Yang, R. M. Kotin, H. L. Paulson and B. L. Davidson (2005). "RNA interference improves motor and neuropathological abnormalities in a Huntington's disease mouse model." Proc Natl Acad Sci U S A 102(16): 5820-5825. Hartzell, H. C. and D. M. Fambrough (1972). "Acetylcholine receptors. Distribution and extrajunctional density in rat diaphragm after denervation correlated with acetylcholine sensitivity." J Gen Physiol 60(3): 248-262. Hayashi, Y., K. Homma and H. Ichijo (2016). "SOD1 in neurotoxicity and its controversial roles in SOD1 mutation-negative ALS." Adv Biol Regul 60: 95-104. He, Y., S. Pan, M. Xu, R. He, W. Huang, P. Song, J. Huang, H. T. Zhang and Y. Hu (2017). "Adeno-associated virus 9-mediated Cdk5 inhibitory peptide reverses pathologic changes and behavioral deficits in the Alzheimer's disease mouse model." FASEB J 31(8): 3383-3392. Heckman, C. J. and R. M. Enoka (2012). "Motor unit." Compr Physiol 2(4): 2629- 2682. Hegedus, J., C. T. Putman and T. Gordon (2007). "Time course of preferential motor unit loss in the SOD1 G93A mouse model of amyotrophic lateral sclerosis." Neurobiol Dis 28(2): 154-164. Hegedus, J., C. T. Putman and T. Gordon (2009). "Progressive motor unit loss in the G93A mouse model of amyotrophic lateral sclerosis is unaffected by gender." Muscle Nerve 39(3): 318-327. Hegedus, J., C. T. Putman, N. Tyreman and T. Gordon (2008). "Preferential motor unit loss in the SOD1 G93A transgenic mouse model of amyotrophic lateral sclerosis." J Physiol 586(14): 3337-3351. Heilman, P. L., S. Song, C. J. Miranda, K. Meyer, A. K. Srivastava, A. Knapp, C. G. Wier, B. K. Kaspar and S. J. Kolb (2017). "HSPB1 mutations causing hereditary neuropathy in humans disrupt non-cell autonomous protection of motor neurons." Exp Neurol 297: 101-109. Heiman-Patterson, T. D., J. S. Deitch, E. P. Blankenhorn, K. L. Erwin, M. J. Perreault, B. K. Alexander, N. Byers, I. Toman and G. M. Alexander (2005). "Background and gender effects on survival in the TgN(SOD1- G93A)1Gur mouse model of ALS." J Neurol Sci 236(1-2): 1-7.

174

Heizmann, C. W., M. W. Berchtold and A. M. Rowlerson (1982). "Correlation of parvalbumin concentration with relaxation speed in mammalian muscles." Proc Natl Acad Sci U S A 79(23): 7243-7247. Henneman, E., G. Somjen and D. O. Carpenter (1965). "Excitability and inhibitability of motoneurons of different sizes." J Neurophysiol 28(3): 599- 620. Herbison, G. J., M. M. Jaweed and J. F. Ditunno (1982). "Muscle fiber types." Arch Phys Med Rehabil 63(5): 227-230. Hilber, K., S. Galler, B. Gohlsch and D. Pette (1999). "Kinetic properties of myosin heavy chain isoforms in single fibers from human skeletal muscle." FEBS Lett 455(3): 267-270. Hockly, E., P. M. Cordery, B. Woodman, A. Mahal, A. van Dellen, C. Blakemore, C. M. Lewis, A. J. Hannan and G. P. Bates (2002). "Environmental enrichment slows disease progression in R6/2 Huntington's disease mice." Ann Neurol 51(2): 235-242. Hodgkin, A. L. and A. F. Huxley (1952). "Movement of sodium and potassium ions during nervous activity." Cold Spring Harb Symp Quant Biol 17: 43- 52. Hodgkin, A. L. and A. F. Huxley (1952). "Propagation of electrical signals along giant nerve fibers." Proc R Soc Lond B Biol Sci 140(899): 177-183. Hodgkin, A. L., A. F. Huxley and B. Katz (1952). "Measurement of current-voltage relations in the membrane of the giant axon of Loligo." J Physiol 116(4): 424-448. Hoffman, E. K., H. M. Wilcox, R. W. Scott and R. Siman (1996). "Proteasome inhibition enhances the stability of mouse Cu/Zn superoxide dismutase with mutations linked to familial amyotrophic lateral sclerosis." J Neurol Sci 139(1): 15-20. Hollingworth, S. and S. M. Baylor (2013). "Comparison of myoplasmic calcium movements during excitation-contraction coupling in frog twitch and mouse fast-twitch muscle fibers." J Gen Physiol 141(5): 567-583. Holm, S. (1979). "A Simple Sequentially Rejective Multiple Test Procedure." Scandinavian Journal of Statistics 6(2): 65-70. Horowicz, P. and M. F. Schneider (1981). "Membrane charge movement in contracting and non-contracting skeletal muscle fibres." J Physiol 314: 565-593. Howland, D. S., J. Liu, Y. She, B. Goad, N. J. Maragakis, B. Kim, J. Erickson, J. Kulik, L. DeVito, G. Psaltis, L. J. DeGennaro, D. W. Cleveland and J. D. Rothstein (2002). "Focal loss of the glutamate transporter EAAT2 in a transgenic rat model of SOD1 mutant-mediated amyotrophic lateral sclerosis (ALS)." Proc Natl Acad Sci U S A 99(3): 1604-1609. Hugon, J. (1996). "Riluzole and ALS therapy." Wien Med Wochenschr 146(9-10): 185-187. Huxley, A. F. and R. Niedergerke (1954). "Structural changes in muscle during contraction; interference microscopy of living muscle fibres." Nature 173(4412): 971-973.

175

Huxley, H. and J. Hanson (1954). "Changes in the cross-striations of muscle during contraction and stretch and their structural interpretation." Nature 173(4412): 973-976. Hynes, T. R., S. M. Block, B. T. White and J. A. Spudich (1987). "Movement of myosin fragments in vitro: domains involved in force production." Cell 48(6): 953-963. Ilieva, H., M. Polymenidou and D. W. Cleveland (2009). "Non-cell autonomous toxicity in neurodegenerative disorders: ALS and beyond." J Cell Biol 187(6): 761-772. Ince, P. G., J. R. Highley, J. Kirby, S. B. Wharton, H. Takahashi, M. J. Strong and P. J. Shaw (2011). "Molecular pathology and genetic advances in amyotrophic lateral sclerosis: an emerging molecular pathway and the significance of glial pathology." Acta Neuropathol 122(6): 657-671. Ito, H., R. Wate, J. Zhang, S. Ohnishi, S. Kaneko, H. Ito, S. Nakano and H. Kusaka (2008). "Treatment with edaravone, initiated at symptom onset, slows motor decline and decreases SOD1 deposition in ALS mice." Exp Neurol 213(2): 448-455. Ives, C. T. and T. J. Doherty (2012). "Intra- and inter-rater reliability of motor unit number estimation and quantitative motor unit analysis in the upper trapezius." Clin Neurophysiol 123(1): 200-205. Ives, C. T. and T. J. Doherty (2014). "Intra-rater reliability of motor unit number estimation and quantitative motor unit analysis in subjects with amyotrophic lateral sclerosis." Clin Neurophysiol 125(1): 170-178. Jaarsma, D., E. Teuling, E. D. Haasdijk, C. I. De Zeeuw and C. C. Hoogenraad (2008). "Neuron-specific expression of mutant superoxide dismutase is sufficient to induce amyotrophic lateral sclerosis in transgenic mice." J Neurosci 28(9): 2075-2088. Jessen, K. R. and R. Mirsky (2016). "The repair Schwann cell and its function in regenerating nerves." J Physiol 594(13): 3521-3531. Johnson, M. A., G. Sideri, D. Weightman and D. Appleton (1973). "A comparison of fibre size, fibre type constitution and spatial fibre type distribution in normal human muscle and in muscle from cases of spinal muscular atrophy and from other neuromuscular disorders." J Neurol Sci 20(4): 345- 361. Jolivalt, C. G., C. A. Lee, K. K. Beiswenger, J. L. Smith, M. Orlov, M. A. Torrance and E. Masliah (2008). "Defective insulin signaling pathway and increased glycogen synthase kinase-3 activity in the brain of diabetic mice: parallels with Alzheimer's disease and correction by insulin." J Neurosci Res 86(15): 3265-3274. Jones, S. L., F. Korobova and T. Svitkina (2014). "Axon initial segment cytoskeleton comprises a multiprotein submembranous coat containing sparse actin filaments." J Cell Biol 205(1): 67-81. Joosten, E. A., R. L. Schuitman, M. E. Vermelis and P. J. Dederen (1992). "Postnatal development of the ipsilateral corticospinal component in rat spinal cord: a light and electron microscopic anterograde HRP study." J Comp Neurol 326(1): 133-146.

176

Kabashi, E., J. N. Agar, D. M. Taylor, S. Minotti and H. D. Durham (2004). "Focal dysfunction of the proteasome: a pathogenic factor in a mouse model of amyotrophic lateral sclerosis." J Neurochem 89(6): 1325-1335. Kang, H., L. Tian, M. Mikesh, J. W. Lichtman and W. J. Thompson (2014). "Terminal Schwann cells participate in neuromuscular synapse remodeling during reinnervation following nerve injury." J Neurosci 34(18): 6323-6333. Kariya, S., T. Obis, C. Garone, T. Akay, F. Sera, S. Iwata, S. Homma and U. R. Monani (2014). "Requirement of enhanced Survival Motoneuron protein imposed during neuromuscular junction maturation." J Clin Invest 124(2): 785-800. Kariya, S., G. H. Park, Y. Maeno-Hikichi, O. Leykekhman, C. Lutz, M. S. Arkovitz, L. T. Landmesser and U. R. Monani (2008). "Reduced SMN protein impairs maturation of the neuromuscular junctions in mouse models of spinal muscular atrophy." Hum Mol Genet 17(16): 2552-2569. Kasselman, L. J., J. M. Shefner and S. B. Rutkove (2009). "Motor unit number estimation in the rat tail using a modified multipoint stimulation technique." Muscle Nerve 40(1): 115-121. Katsuno, M., F. Tanaka and G. Sobue (2012). "Perspectives on molecular targeted therapies and clinical trials for neurodegenerative diseases." J Neurol Neurosurg Psychiatry 83(3): 329-335. Kaur, S. J., S. R. McKeown and S. Rashid (2016). "Mutant SOD1 mediated pathogenesis of Amyotrophic Lateral Sclerosis." Gene 577(2): 109-118. Kaya, R. D., R. L. Hoffman and B. C. Clark (2014). "Reliability of a modified motor unit number index (MUNIX) technique." J Electromyogr Kinesiol 24(1): 18-24. Kelly, J. J., Jr., L. Thibodeau, P. L. Andres and L. J. Finison (1990). "Use of electrophysiologic tests to measure disease progression in ALS therapeutic trials." Muscle Nerve 13(6): 471-479. Kemp, S. W., P. S. Cederna and R. Midha (2017). "Comparative outcome measures in peripheral regeneration studies." Exp Neurol 287(Pt 3): 348- 357. Kevenaar, J. T. and C. C. Hoogenraad (2015). "The axonal cytoskeleton: from organization to function." Front Mol Neurosci 8: 44. Kiernan, M. C., S. Vucic, B. C. Cheah, M. R. Turner, A. Eisen, O. Hardiman, J. R. Burrell and M. C. Zoing (2011). "Amyotrophic lateral sclerosis." Lancet 377(9769): 942-955. Knapinska, A. M., F. M. Gratacos, C. D. Krause, K. Hernandez, A. G. Jensen, J. J. Bradley, X. Wu, S. Pestka and G. Brewer (2011). "Chaperone Hsp27 modulates AUF1 proteolysis and AU-rich element-mediated mRNA degradation." Mol Cell Biol 31(7): 1419-1431. Kolb, S. J., C. S. Coffey, J. W. Yankey, K. Krosschell, W. D. Arnold, S. B. Rutkove, K. J. Swoboda, S. P. Reyna, A. Sakonju, B. T. Darras, R. Shell, N. Kuntz, D. Castro, S. T. Iannaccone, J. Parsons, A. M. Connolly, C. A. Chiriboga, C. McDonald, W. B. Burnette, K. Werner, M. Thangarajh, P. B. Shieh, E. Finanger, M. E. Cudkowicz, M. M. McGovern, D. E. McNeil, R. Finkel, E. Kaye, A. Kingsley, S. R. Renusch, V. L. McGovern, X. Wang, P.

177

G. Zaworski, T. W. Prior, A. H. Burghes, A. Bartlett, J. T. Kissel, N. C. T. N. Neuro and N. N. S. M. A. B. I. on behalf of the (2016). "Baseline results of the NeuroNEXT spinal muscular atrophy infant biomarker study." Ann Clin Transl Neurol 3(2): 132-145. Kolb, S. J., C. S. Coffey, J. W. Yankey, K. Krosschell, W. D. Arnold, S. B. Rutkove, K. J. Swoboda, S. P. Reyna, A. Sakonju, B. T. Darras, R. Shell, N. Kuntz, D. Castro, J. Parsons, A. M. Connolly, C. A. Chiriboga, C. McDonald, W. B. Burnette, K. Werner, M. Thangarajh, P. B. Shieh, E. Finanger, M. E. Cudkowicz, M. M. McGovern, D. E. McNeil, R. Finkel, S. T. Iannaccone, E. Kaye, A. Kingsley, S. R. Renusch, V. L. McGovern, X. Wang, P. G. Zaworski, T. W. Prior, A. H. M. Burghes, A. Bartlett, J. T. Kissel and N. C. T. N. o. b. o. t. N. N. S. M. A. B. I. Neuro (2017). "Natural history of infantile-onset spinal muscular atrophy." Ann Neurol 82(6): 883- 891. Kolb, S. J. and J. T. Kissel (2015). "Spinal Muscular Atrophy." Neurol Clin 33(4): 831-846. Koo, T. K. and M. Y. Li (2016). "A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research." J Chiropr Med 15(2): 155-163. Kouyoumdjian, J. A. (2006). "Peripheral nerve injuries: a retrospective survey of 456 cases." Muscle Nerve 34(6): 785-788. Kraft, G. H. (2007). "The electromyographer's guide to the motor unit." Phys Med Rehabil Clin N Am 18(4): 711-732, vi. Krarup, C., M. Boeckstyns, A. Ibsen, M. Moldovan and S. Archibald (2016). "Remodeling of motor units after nerve regeneration studied by quantitative electromyography." Clin Neurophysiol 127(2): 1675-1682. Krishnan, J., K. Vannuvel, M. Andries, E. Waelkens, W. Robberecht and L. Van Den Bosch (2008). "Over-expression of Hsp27 does not influence disease in the mutant SOD1(G93A) mouse model of amyotrophic lateral sclerosis." J Neurochem 106(5): 2170-2183. Kuypers, H. G. (1964). "The Descending Pathways to the Spinal Cord, Their Anatomy and Function." Prog Brain Res 11: 178-202. Lai, F. A., H. P. Erickson, E. Rousseau, Q. Y. Liu and G. Meissner (1988). "Purification and reconstitution of the calcium release channel from skeletal muscle." Nature 331(6154): 315-319. Landry, J., P. Chretien, H. Lambert, E. Hickey and L. A. Weber (1989). "Heat shock resistance conferred by expression of the human HSP27 gene in rodent cells." J Cell Biol 109(1): 7-15. Lateva, Z. C., K. C. McGill and C. G. Burgar (1996). "Anatomical and electrophysiological determinants of the human thenar compound muscle action potential." Muscle Nerve 19(11): 1457-1468. Le, T. T., L. T. Pham, M. E. Butchbach, H. L. Zhang, U. R. Monani, D. D. Coovert, T. O. Gavrilina, L. Xing, G. J. Bassell and A. H. Burghes (2005). "SMNDelta7, the major product of the centromeric survival motor neuron (SMN2) gene, extends survival in mice with spinal muscular atrophy and associates with full-length SMN." Hum Mol Genet 14(6): 845-857.

178

Lemon, R. N. and J. Griffiths (2005). "Comparing the function of the corticospinal system in different species: organizational differences for motor specialization?" Muscle Nerve 32(3): 261-279. Lepore, A. C., C. Haenggeli, M. Gasmi, K. M. Bishop, R. T. Bartus, N. J. Maragakis and J. D. Rothstein (2007). "Intraparenchymal spinal cord delivery of adeno-associated virus IGF-1 is protective in the SOD1G93A model of ALS." Brain Res 1185: 256-265. Li, J., A. Pacheck, B. Sanchez and S. B. Rutkove (2016). "Single and modeled multifrequency electrical impedance myography parameters and their relationship to force production in the ALS SOD1G93A mouse." Amyotroph Lateral Scler Frontotemporal Degener 17(5-6): 397-403. Li, J., W. L. Staats, A. Spieker, M. Sung and S. B. Rutkove (2012). "A technique for performing electrical impedance myography in the mouse hind limb: data in normal and ALS SOD1 G93A animals." PLoS One 7(9): e45004. Li, J., M. Sung and S. B. Rutkove (2013). "Electrophysiologic biomarkers for assessing disease progression and the effect of riluzole in SOD1 G93A ALS mice." PLoS One 8(6): e65976. Lin, C. L., L. A. Bristol, L. Jin, M. Dykes-Hoberg, T. Crawford, L. Clawson and J. D. Rothstein (1998). "Aberrant RNA processing in a neurodegenerative disease: the cause for absent EAAT2, a glutamate transporter, in amyotrophic lateral sclerosis." Neuron 20(3): 589-602. Ling, S. C., C. P. Albuquerque, J. S. Han, C. Lagier-Tourenne, S. Tokunaga, H. Zhou and D. W. Cleveland (2010). "ALS-associated mutations in TDP-43 increase its stability and promote TDP-43 complexes with FUS/TLS." Proc Natl Acad Sci U S A 107(30): 13318-13323. Liu, J., C. Lillo, P. A. Jonsson, C. Vande Velde, C. M. Ward, T. M. Miller, J. R. Subramaniam, J. D. Rothstein, S. Marklund, P. M. Andersen, T. Brannstrom, O. Gredal, P. C. Wong, D. S. Williams and D. W. Cleveland (2004). "Toxicity of familial ALS-linked SOD1 mutants from selective recruitment to spinal mitochondria." Neuron 43(1): 5-17. Lorand, L. (1953). "Adenosine triphosphate-creatine transphosphorylase as relaxing factor of muscle." Nature 172(4391): 1181-1183. Luo, G., J. Yi, C. Ma, Y. Xiao, F. Yi, T. Yu and J. Zhou (2013). "Defective mitochondrial dynamics is an early event in skeletal muscle of an amyotrophic lateral sclerosis mouse model." PLoS One 8(12): e82112. Ma, C. H., T. Omura, E. J. Cobos, A. Latremoliere, N. Ghasemlou, G. J. Brenner, E. van Veen, L. Barrett, T. Sawada, F. Gao, G. Coppola, F. Gertler, M. Costigan, D. Geschwind and C. J. Woolf (2011). "Accelerating axonal growth promotes motor recovery after peripheral nerve injury in mice." J Clin Invest 121(11): 4332-4347. Maathuis, E. M., J. Drenthen, P. A. van Doorn, G. H. Visser and J. H. Blok (2013). "The CMAP scan as a tool to monitor disease progression in ALS and PMA." Amyotroph Lateral Scler Frontotemporal Degener 14(3): 217- 223.

179

Maatkamp, A., A. Vlug, E. Haasdijk, D. Troost, P. J. French and D. Jaarsma (2004). "Decrease of Hsp25 protein expression precedes degeneration of motoneurons in ALS-SOD1 mice." Eur J Neurosci 20(1): 14-28. Macefield, V. G., A. J. Fuglevand and B. Bigland-Ritchie (1996). "Contractile properties of single motor units in human toe extensors assessed by intraneural motor axon stimulation." J Neurophysiol 75(6): 2509-2519. MacLennan, D. H., C. J. Brandl, B. Korczak and N. M. Green (1985). "Amino-acid sequence of a Ca2+ + Mg2+-dependent ATPase from rabbit muscle sarcoplasmic reticulum, deduced from its complementary DNA sequence." Nature 316(6030): 696-700. Maier, M. A., J. Armand, P. A. Kirkwood, H. W. Yang, J. N. Davis and R. N. Lemon (2002). "Differences in the corticospinal projection from primary motor cortex and supplementary motor area to macaque upper limb motoneurons: an anatomical and electrophysiological study." Cereb Cortex 12(3): 281-296. Mallik, A. and A. I. Weir (2005). "Nerve conduction studies: essentials and pitfalls in practice." J Neurol Neurosurg Psychiatry 76 Suppl 2: ii23-31. Mancuso, R., R. Osta and X. Navarro (2014). "Presymptomatic electrophysiological tests predict clinical onset and survival in SOD1(G93A) ALS mice." Muscle Nerve 50(6): 943-949. Mancuso, R., E. Santos-Nogueira, R. Osta and X. Navarro (2011). "Electrophysiological analysis of a murine model of motoneuron disease." Clin Neurophysiol 122(8): 1660-1670. Martin, L. J., Z. Liu, K. Chen, A. C. Price, Y. Pan, J. A. Swaby and W. C. Golden (2007). "Motor neuron degeneration in amyotrophic lateral sclerosis mutant superoxide dismutase-1 transgenic mice: mechanisms of mitochondriopathy and cell death." J Comp Neurol 500(1): 20-46. Mason, S., J. Wardrope, G. Turpin and A. Rowlands (2002). "Outcomes after injury: a comparison of workplace and nonworkplace injury." J Trauma 53(1): 98-103. Mattiazzi, M., M. D'Aurelio, C. D. Gajewski, K. Martushova, M. Kiaei, M. F. Beal and G. Manfredi (2002). "Mutated human SOD1 causes dysfunction of oxidative phosphorylation in mitochondria of transgenic mice." J Biol Chem 277(33): 29626-29633. McCarty, D. M., P. E. Monahan and R. J. Samulski (2001). "Self-complementary recombinant adeno-associated virus (scAAV) vectors promote efficient transduction independently of DNA synthesis." Gene Ther 8(16): 1248- 1254. McComas, A. J., P. R. Fawcett, M. J. Campbell and R. E. Sica (1971). "Electrophysiological estimation of the number of motor units within a human muscle." J Neurol Neurosurg Psychiatry 34(2): 121-131. McCombe, P. A. and R. D. Henderson (2010). "Effects of gender in amyotrophic lateral sclerosis." Gend Med 7(6): 557-570. McCord, J. M. and I. Fridovich (1969). "Superoxide dismutase. An enzymic function for erythrocuprein (hemocuprein)." J Biol Chem 244(22): 6049- 6055.

180

McGovern, V. L., C. C. Iyer, W. D. Arnold, S. E. Gombash, P. G. Zaworski, A. J. Blatnik, 3rd, K. D. Foust and A. H. Burghes (2015). "SMN expression is required in motor neurons to rescue electrophysiological deficits in the SMNDelta7 mouse model of SMA." Hum Mol Genet 24(19): 5524-5541. Mears, S. C. and E. Frank (1997). "Formation of specific monosynaptic connections between muscle spindle afferents and motoneurons in the mouse." J Neurosci 17(9): 3128-3135. Mendell, J. R., S. Al-Zaidy, R. Shell, W. D. Arnold, L. R. Rodino-Klapac, T. W. Prior, L. Lowes, L. Alfano, K. Berry, K. Church, J. T. Kissel, S. Nagendran, J. L'Italien, D. M. Sproule, C. Wells, J. A. Cardenas, M. D. Heitzer, A. Kaspar, S. Corcoran, L. Braun, S. Likhite, C. Miranda, K. Meyer, K. D. Foust, A. H. M. Burghes and B. K. Kaspar (2017). "Single-Dose Gene- Replacement Therapy for Spinal Muscular Atrophy." N Engl J Med 377(18): 1713-1722. Menovsky, T. and J. F. Beek (2003). "Carbon dioxide laser-assisted nerve repair: effect of solder and suture material on nerve regeneration in rat sciatic nerve." Microsurgery 23(2): 109-116. Meyer, K., L. Ferraiuolo, L. Schmelzer, L. Braun, V. McGovern, S. Likhite, O. Michels, A. Govoni, J. Fitzgerald, P. Morales, K. D. Foust, J. R. Mendell, A. H. Burghes and B. K. Kaspar (2015). "Improving Single Injection CSF Delivery of AAV9-mediated Gene Therapy for SMA: A Dose-response Study in Mice and Nonhuman Primates." Mol Ther 23(3): 477-487. Milani, P., S. Gagliardi, E. Cova and C. Cereda (2011). "SOD1 Transcriptional and Posttranscriptional Regulation and Its Potential Implications in ALS." Neurol Res Int 2011: 458427. Miledi, R. and C. R. Slater (1970). "On the degeneration of rat neuromuscular junctions after nerve section." J Physiol 207(2): 507-528. Miller, B. R., C. Press, R. W. Daniels, Y. Sasaki, J. Milbrandt and A. DiAntonio (2009). "A dual leucine kinase-dependent axon self-destruction program promotes Wallerian degeneration." Nat Neurosci 12(4): 387-389. Miller, T. M., S. H. Kim, K. Yamanaka, M. Hester, P. Umapathi, H. Arnson, L. Rizo, J. R. Mendell, F. H. Gage, D. W. Cleveland and B. K. Kaspar (2006). "Gene transfer demonstrates that muscle is not a primary target for non- cell-autonomous toxicity in familial amyotrophic lateral sclerosis." Proc Natl Acad Sci U S A 103(51): 19546-19551. Mintz, E. L., J. A. Passipieri, D. Y. Lovell and G. J. Christ (2016). "Applications of In Vivo Functional Testing of the Rat Tibialis Anterior for Evaluating Tissue Engineered Skeletal Muscle Repair." J Vis Exp(116). Mori, A., S. Yamashita, M. Nakajima, H. Hori, A. Tawara, Y. Matsuo, Y. Misumi and Y. Ando (2016). "CMAP decrement as a potential diagnostic marker for ALS." Acta Neurol Scand 134(1): 49-53. Morrison, B. M., A. Tsingalia, S. Vidensky, Y. Lee, L. Jin, M. H. Farah, S. Lengacher, P. J. Magistretti, L. Pellerin and J. D. Rothstein (2015). "Deficiency in monocarboxylate transporter 1 (MCT1) in mice delays regeneration of peripheral nerves following sciatic nerve crush." Exp Neurol 263: 325-338.

181

Mukaka, M. M. (2012). "Statistics corner: A guide to appropriate use of correlation coefficient in medical research." Malawi Med J 24(3): 69-71. Murakami, T., I. Nagano, T. Hayashi, Y. Manabe, M. Shoji, Y. Setoguchi and K. Abe (2001). "Impaired retrograde axonal transport of adenovirus-mediated E. coli LacZ gene in the mice carrying mutant SOD1 gene." Neurosci Lett 308(3): 149-152. Naumenko, N., E. Pollari, A. Kurronen, R. Giniatullina, A. Shakirzyanova, J. Magga, J. Koistinaho and R. Giniatullin (2011). "Gender-Specific Mechanism of Synaptic Impairment and Its Prevention by GCSF in a Mouse Model of ALS." Front Cell Neurosci 5: 26. Navarro, X. (2016). "Functional evaluation of peripheral nerve regeneration and target reinnervation in animal models: a critical overview." Eur J Neurosci 43(3): 271-286. Neuwirth, C., S. Nandedkar, E. Stalberg, P. E. Barkhaus, M. Carvalho, J. Furtula, J. P. Dijk, R. Baldinger, J. Castro, J. Costa, M. Otto, A. Sandberg and M. Weber (2011). "Motor Unit Number Index (MUNIX): a novel neurophysiological marker for neuromuscular disorders; test-retest reliability in healthy volunteers." Clin Neurophysiol 122(9): 1867-1872. Ngo, S. T., F. Baumann, P. G. Ridall, A. N. Pettitt, R. D. Henderson, M. C. Bellingham and P. A. McCombe (2012). "The relationship between Bayesian motor unit number estimation and histological measurements of motor neurons in wild-type and SOD1(G93A) mice." Clin Neurophysiol 123(10): 2080-2091. Nicolopoulos-Stournaras, S. and J. F. Iles (1983). "Motor neuron columns in the lumbar spinal cord of the rat." J Comp Neurol 217(1): 75-85. Ojha, J., G. Masilamoni, D. Dunlap, R. A. Udoff and A. G. Cashikar (2011). "Sequestration of toxic oligomers by HspB1 as a cytoprotective mechanism." Mol Cell Biol 31(15): 3146-3157. Olivan, S., A. C. Calvo, A. Rando, M. J. Munoz, P. Zaragoza and R. Osta (2015). "Comparative study of behavioural tests in the SOD1G93A mouse model of amyotrophic lateral sclerosis." Exp Anim 64(2): 147-153. Palispis, W. A. and R. Gupta (2017). "Surgical repair in humans after traumatic nerve injury provides limited functional neural regeneration in adults." Exp Neurol 290: 106-114. Patel, Y. J., M. D. Payne Smith, J. de Belleroche and D. S. Latchman (2005). "Hsp27 and Hsp70 administered in combination have a potent protective effect against FALS-associated SOD1-mutant-induced cell death in mammalian neuronal cells." Brain Res Mol Brain Res 134(2): 256-274. Perlson, E., G. B. Jeong, J. L. Ross, R. Dixit, K. E. Wallace, R. G. Kalb and E. L. Holzbaur (2009). "A switch in retrograde signaling from survival to stress in rapid-onset neurodegeneration." J Neurosci 29(31): 9903-9917. Perry, J. J., D. S. Shin, E. D. Getzoff and J. A. Tainer (2010). "The structural biochemistry of the superoxide dismutases." Biochim Biophys Acta 1804(2): 245-262.

182

Petrov, D., C. Mansfield, A. Moussy and O. Hermine (2017). "ALS Clinical Trials Review: 20 Years of Failure. Are We Any Closer to Registering a New Treatment?" Front Aging Neurosci 9: 68. Pette, D., H. Peuker and R. S. Staron (1999). "The impact of biochemical methods for single muscle fibre analysis." Acta Physiol Scand 166(4): 261- 277. Philips, T. and J. D. Rothstein (2015). "Rodent Models of Amyotrophic Lateral Sclerosis." Curr Protoc Pharmacol 69: 5 67 61-21. Plomp, J. J., M. G. M. Huijbers and J. Verschuuren (2018). "Neuromuscular synapse electrophysiology in myasthenia gravis animal models." Ann N Y Acad Sci 1412(1): 146-153. Plomp, J. J., G. T. Van Kempen, M. B. De Baets, Y. M. Graus, J. B. Kuks and P. C. Molenaar (1995). "Acetylcholine release in myasthenia gravis: regulation at single end-plate level." Ann Neurol 37(5): 627-636. Polkey, M. I., R. A. Lyall, J. Moxham and P. N. Leigh (1999). "Respiratory aspects of neurological disease." J Neurol Neurosurg Psychiatry 66(1): 5- 15. Porensky, P. N., C. Mitrpant, V. L. McGovern, A. K. Bevan, K. D. Foust, B. K. Kaspar, S. D. Wilton and A. H. Burghes (2012). "A single administration of morpholino antisense oligomer rescues spinal muscular atrophy in mouse." Hum Mol Genet 21(7): 1625-1638. Powers, S. K., L. L. Ji, A. N. Kavazis and M. J. Jackson (2011). "Reactive oxygen species: impact on skeletal muscle." Compr Physiol 1(2): 941-969. Pugdahl, K., A. Fuglsang-Frederiksen, M. de Carvalho, B. Johnsen, P. R. Fawcett, A. Labarre-Vila, R. Liguori, W. A. Nix and I. S. Schofield (2007). "Generalised sensory system abnormalities in amyotrophic lateral sclerosis: a European multicentre study." J Neurol Neurosurg Psychiatry 78(7): 746-749. Pun, S., A. F. Santos, S. Saxena, L. Xu and P. Caroni (2006). "Selective vulnerability and pruning of phasic motoneuron axons in motoneuron disease alleviated by CNTF." Nat Neurosci 9(3): 408-419. Qaisar, R., S. Bhaskaran, P. Premkumar, R. Ranjit, K. S. Natarajan, B. Ahn, K. Riddle, D. R. Claflin, A. Richardson, S. V. Brooks and H. Van Remmen (2018). "Oxidative stress-induced dysregulation of excitation-contraction coupling contributes to muscle weakness." J Cachexia Sarcopenia Muscle. Raez, M. B., M. S. Hussain and F. Mohd-Yasin (2006). "Techniques of EMG signal analysis: detection, processing, classification and applications." Biol Proced Online 8: 11-35. Rasminsky, M. and T. A. Sears (1972). "Internodal conduction in undissected demyelinated nerve fibres." J Physiol 227(2): 323-350. Rathbone, C. R., J. C. Wenke, G. L. Warren and R. B. Armstrong (2003). "Importance of satellite cells in the strength recovery after eccentric contraction-induced muscle injury." Am J Physiol Regul Integr Comp Physiol 285(6): R1490-1495.

183

Reaume, A. G., J. L. Elliott, E. K. Hoffman, N. W. Kowall, R. J. Ferrante, D. F. Siwek, H. M. Wilcox, D. G. Flood, M. F. Beal, R. H. Brown, Jr., R. W. Scott and W. D. Snider (1996). "Motor neurons in Cu/Zn superoxide dismutase- deficient mice develop normally but exhibit enhanced cell death after axonal injury." Nat Genet 13(1): 43-47. Reichert, F., A. Saada and S. Rotshenker (1994). "Peripheral nerve injury induces Schwann cells to express two macrophage phenotypes: phagocytosis and the galactose-specific lectin MAC-2." J Neurosci 14(5 Pt 2): 3231-3245. Renkawek, K., G. J. Stege and G. J. Bosman (1999). "Dementia, gliosis and expression of the small heat shock proteins hsp27 and alpha B-crystallin in Parkinson's disease." Neuroreport 10(11): 2273-2276. Rexed, B. (1954). "A cytoarchitectonic atlas of the spinal cord in the cat." J Comp Neurol 100(2): 297-379. Reynolds, M. L. and C. J. Woolf (1992). "Terminal Schwann cells elaborate extensive processes following denervation of the motor endplate." J Neurocytol 21(1): 50-66. Riet-Correa, G., C. G. Fernandes, L. A. Pereira and D. L. Graca (2002). "Ethidium bromide-induced demyelination of the sciatic nerve of adult Wistar rats." Braz J Med Biol Res 35(1): 99-104. Rios, E. and G. Brum (1987). "Involvement of dihydropyridine receptors in excitation-contraction coupling in skeletal muscle." Nature 325(6106): 717- 720. Rizzuto, R., A. W. Simpson, M. Brini and T. Pozzan (1992). "Rapid changes of mitochondrial Ca2+ revealed by specifically targeted recombinant aequorin." Nature 358(6384): 325-327. Rogers, K. L., S. Picaud, E. Roncali, R. Boisgard, C. Colasante, J. Stinnakre, B. Tavitian and P. Brulet (2007). "Non-invasive in vivo imaging of calcium signaling in mice." PLoS One 2(10): e974. Romeo-Guitart, D., J. Fores, X. Navarro and C. Casas (2017). "Boosted Regeneration and Reduced Denervated Muscle Atrophy by NeuroHeal in a Pre-clinical Model of Lumbar Root Avulsion with Delayed Reimplantation." Sci Rep 7(1): 12028. Rosen, D. R., T. Siddique, D. Patterson, D. A. Figlewicz, P. Sapp, A. Hentati, D. Donaldson, J. Goto, J. P. O'Regan, H. X. Deng and et al. (1993). "Mutations in Cu/Zn superoxide dismutase gene are associated with familial amyotrophic lateral sclerosis." Nature 362(6415): 59-62. Rossor, A. M., G. L. Davidson, J. Blake, J. M. Polke, S. M. Murphy, H. Houlden, A. Innes, B. Kalmar, L. Greensmith and M. M. Reilly (2012). "A novel p.Gln175X [corrected] premature stop mutation in the C-terminal end of HSP27 is a cause of CMT2." J Peripher Nerv Syst 17(2): 201-205. Rothstein, J. D. (1995). "Excitotoxicity and neurodegeneration in amyotrophic lateral sclerosis." Clin Neurosci 3(6): 348-359. Ruijs, A. C., J. B. Jaquet, S. Kalmijn, H. Giele and S. E. Hovius (2005). "Median and ulnar nerve injuries: a meta-analysis of predictors of motor and

184

sensory recovery after modern microsurgical nerve repair." Plast Reconstr Surg 116(2): 484-494; discussion 495-486. Rushton, W. A. (1951). "A theory of the effects of fibre size in medullated nerve." J Physiol 115(1): 101-122. Rutkove, S. B. (2015). "Clinical Measures of Disease Progression in Amyotrophic Lateral Sclerosis." Neurotherapeutics 12(2): 384-393. Rutkove, S. B., J. B. Caress, M. S. Cartwright, T. M. Burns, J. Warder, W. S. David, N. Goyal, N. J. Maragakis, L. Clawson, M. Benatar, S. Usher, K. R. Sharma, S. Gautam, P. Narayanaswami, E. M. Raynor, M. L. Watson and J. M. Shefner (2012). "Electrical impedance myography as a biomarker to assess ALS progression." Amyotroph Lateral Scler 13(5): 439-445. Sabado, J., A. Casanovas, O. Tarabal, M. Hereu, L. Piedrafita, J. Caldero and J. E. Esquerda (2014). "Accumulation of misfolded SOD1 in dorsal root ganglion degenerating proprioceptive sensory neurons of transgenic mice with amyotrophic lateral sclerosis." Biomed Res Int 2014: 852163. Sahenk, Z., G. Galloway, C. Edwards, V. Malik, B. K. Kaspar, A. Eagle, B. Yetter, A. Forgie, D. Tsao and J. C. Lin (2010). "TrkB and TrkC agonist antibodies improve function, electrophysiologic and pathologic features in Trembler J mice." Exp Neurol 224(2): 495-506. Sakuma, M., G. Gorski, S. H. Sheu, S. Lee, L. B. Barrett, B. Singh, T. Omura, A. Latremoliere and C. J. Woolf (2016). "Lack of motor recovery after prolonged denervation of the neuromuscular junction is not due to regenerative failure." Eur J Neurosci 43(3): 451-462. Salmons, S. and G. Vrbova (1969). "The influence of activity on some contractile characteristics of mammalian fast and slow muscles." J Physiol 201(3): 535-549. Schneider, M. F. and W. K. Chandler (1973). "Voltage dependent charge movement of skeletal muscle: a possible step in excitation-contraction coupling." Nature 242(5395): 244-246. Schrank, B., R. Gotz, J. M. Gunnersen, J. M. Ure, K. V. Toyka, A. G. Smith and M. Sendtner (1997). "Inactivation of the survival motor neuron gene, a candidate gene for human spinal muscular atrophy, leads to massive cell death in early mouse embryos." Proc Natl Acad Sci U S A 94(18): 9920- 9925. Schreyer, D. J. and E. G. Jones (1982). "Growth and target finding by axons of the corticospinal tract in prenatal and postnatal rats." Neuroscience 7(8): 1837-1853. Scott, W., J. Stevens and S. A. Binder-Macleod (2001). "Human skeletal muscle fiber type classifications." Phys Ther 81(11): 1810-1816. Seddon, H. J., P. B. Medawar and H. Smith (1943). "Rate of regeneration of peripheral nerves in man." J Physiol 102(2): 191-215. Shakhbazau, A., C. Mohanty, D. Shcharbin, M. Bryszewska, A. M. Caminade, J. P. Majoral, J. Alant and R. Midha (2013). "Doxycycline-regulated GDNF expression promotes axonal regeneration and functional recovery in transected peripheral nerve." J Control Release 172(3): 841-851.

185

Sharp, P., M. Krishnan, O. Pullar, R. Navarrete, D. Wells and J. de Belleroche (2006). "Heat shock protein 27 rescues motor neurons following nerve injury and preserves muscle function." Exp Neurol 198(2): 511-518. Sharp, P. S., M. T. Akbar, S. Bouri, A. Senda, K. Joshi, H. J. Chen, D. S. Latchman, D. J. Wells and J. de Belleroche (2008). "Protective effects of heat shock protein 27 in a model of ALS occur in the early stages of disease progression." Neurobiol Dis 30(1): 42-55. Shefner, J. M., R. H. Brown, Jr., D. Cole, P. Chaturvedi, D. Schoenfeld, K. Pastuszak, R. Matthews, M. Upton-Rice and M. E. Cudkowicz (2001). "Effect of neurophilin ligands on motor units in mice with SOD1 ALS mutations." Neurology 57(10): 1857-1861. Shefner, J. M., M. Cudkowicz and R. H. Brown, Jr. (2006). "Motor unit number estimation predicts disease onset and survival in a transgenic mouse model of amyotrophic lateral sclerosis." Muscle Nerve 34(5): 603-607. Shefner, J. M., M. E. Cudkowicz and R. H. Brown, Jr. (2002). "Comparison of incremental with multipoint MUNE methods in transgenic ALS mice." Muscle Nerve 25(1): 39-42. Shefner, J. M., D. Liu, M. L. Leitner, D. Schoenfeld, D. R. Johns, T. Ferguson and M. Cudkowicz (2016). "Quantitative strength testing in ALS clinical trials." Neurology 87(6): 617-624. Shefner, J. M., M. L. Watson, L. Simionescu, J. B. Caress, T. M. Burns, N. J. Maragakis, M. Benatar, W. S. David, K. R. Sharma and S. B. Rutkove (2011). "Multipoint incremental motor unit number estimation as an outcome measure in ALS." Neurology 77(3): 235-241. Sheth, K. A., C. C. Iyer, C. G. Wier, A. E. Crum, A. Bratasz, S. J. Kolb, B. C. Clark, A. H. M. Burghes and W. D. Arnold (2018). "Muscle strength and size are associated with motor unit connectivity in aged mice." Neurobiol Aging 67: 128-136. Shichinohe, H., S. Kuroda, H. Yasuda, T. Ishikawa, M. Iwai, M. Horiuchi and Y. Iwasaki (2004). "Neuroprotective effects of the free radical scavenger Edaravone (MCI-186) in mice permanent focal brain ischemia." Brain Res 1029(2): 200-206. Shimada, Y., R. Tanaka, H. Shimura, K. Yamashiro, T. Urabe and N. Hattori (2014). "Phosphorylation enhances recombinant HSP27 neuroprotection against focal cerebral ischemia in mice." Neuroscience 278: 113-121. Sieck, G. C. and Y. S. Prakash (1997). "Morphological adaptations of neuromuscular junctions depend on fiber type." Can J Appl Physiol 22(3): 197-230. Siemionow, M. and G. Brzezicki (2009). "Chapter 8: Current techniques and concepts in peripheral nerve repair." Int Rev Neurobiol 87: 141-172. Sievers, C., N. Platt, V. H. Perry, M. P. Coleman and L. Conforti (2003). "Neurites undergoing Wallerian degeneration show an apoptotic-like process with Annexin V positive staining and loss of mitochondrial membrane potential." Neurosci Res 46(2): 161-169.

186

Simon, N. G., M. R. Turner, S. Vucic, A. Al-Chalabi, J. Shefner, C. Lomen-Hoerth and M. C. Kiernan (2014). "Quantifying disease progression in amyotrophic lateral sclerosis." Ann Neurol 76(5): 643-657. Sleivert, G. G. and H. A. Wenger (1994). "Reliability of measuring isometric and isokinetic peak torque, rate of torque development, integrated electromyography, and tibial nerve conduction velocity." Arch Phys Med Rehabil 75(12): 1315-1321. Smith, J. S., T. Imagawa, J. Ma, M. Fill, K. P. Campbell and R. Coronado (1988). "Purified ryanodine receptor from rabbit skeletal muscle is the calcium- release channel of sarcoplasmic reticulum." J Gen Physiol 92(1): 1-26. Song, S., C. J. Miranda, L. Braun, K. Meyer, A. E. Frakes, L. Ferraiuolo, S. Likhite, A. K. Bevan, K. D. Foust, M. J. McConnell, C. M. Walker and B. K. Kaspar (2016). "Major histocompatibility complex class I molecules protect motor neurons from astrocyte-induced toxicity in amyotrophic lateral sclerosis." Nat Med 22(4): 397-403. Spillane, M., A. Ketschek, S. L. Jones, F. Korobova, B. Marsick, L. Lanier, T. Svitkina and G. Gallo (2011). "The actin nucleating Arp2/3 complex contributes to the formation of axonal filopodia and branches through the regulation of actin patch precursors to filopodia." Dev Neurobiol 71(9): 747-758. Srivastava, A. K., S. R. Renusch, N. E. Naiman, S. Gu, A. Sneh, W. D. Arnold, Z. Sahenk and S. J. Kolb (2012). "Mutant HSPB1 overexpression in neurons is sufficient to cause age-related motor neuronopathy in mice." Neurobiol Dis 47(2): 163-173. Stam, F. J., H. D. MacGillavry, N. J. Armstrong, M. C. de Gunst, Y. Zhang, R. E. van Kesteren, A. B. Smit and J. Verhaagen (2007). "Identification of candidate transcriptional modulators involved in successful regeneration after nerve injury." Eur J Neurosci 25(12): 3629-3637. Staprans, I., H. Takahashi, M. P. Russell and S. Watanabe (1972). "Skeletal and cardiac troponins and their components." J Biochem 72(3): 723-735. Staron, R. S. (1997). "Human skeletal muscle fiber types: delineation, development, and distribution." Can J Appl Physiol 22(4): 307-327. Sterniczuk, R., M. C. Antle, F. M. Laferla and R. H. Dyck (2010). "Characterization of the 3xTg-AD mouse model of Alzheimer's disease: part 2. Behavioral and cognitive changes." Brain Res 1348: 149-155. Stetler, R. A., G. Cao, Y. Gao, F. Zhang, S. Wang, Z. Weng, P. Vosler, L. Zhang, A. Signore, S. H. Graham and J. Chen (2008). "Hsp27 protects against ischemic brain injury via attenuation of a novel stress-response cascade upstream of mitochondrial cell death signaling." J Neurosci 28(49): 13038- 13055. Stifani, N. (2014). "Motor neurons and the generation of spinal motor neuron diversity." Front Cell Neurosci 8: 293. Sudhof, T. C. (2013). "Neurotransmitter release: the last millisecond in the life of a synaptic vesicle." Neuron 80(3): 675-690. Sugarman, E. A., N. Nagan, H. Zhu, V. R. Akmaev, Z. Zhou, E. M. Rohlfs, K. Flynn, B. C. Hendrickson, T. Scholl, D. A. Sirko-Osadsa and B. A. Allitto

187

(2012). "Pan-ethnic carrier screening and prenatal diagnosis for spinal muscular atrophy: clinical laboratory analysis of >72,400 specimens." Eur J Hum Genet 20(1): 27-32. Sun, J. Y., V. Anand-Jawa, S. Chatterjee and K. K. Wong (2003). "Immune responses to adeno-associated virus and its recombinant vectors." Gene Ther 10(11): 964-976. Sunderland, S. (1947). "Rate of regeneration in human peripheral nerves; analysis of the interval between injury and onset of recovery." Arch Neurol Psychiatry 58(3): 251-295. Sunderland, S. (1951). "A classification of peripheral nerve injuries producing loss of function." Brain 74(4): 491-516. Sweeney, H. L. and D. W. Hammers (2018). "Muscle Contraction." Cold Spring Harb Perspect Biol 10(2). Swoboda, K. J., C. B. Scott, S. P. Reyna, T. W. Prior, B. LaSalle, S. L. Sorenson, J. Wood, G. Acsadi, T. O. Crawford, J. T. Kissel, K. J. Krosschell, G. D'Anjou, M. B. Bromberg, M. K. Schroth, G. M. Chan, B. Elsheikh and L. R. Simard (2009). "Phase II open label study of valproic acid in spinal muscular atrophy." PLoS One 4(5): e5268. Takeshima, H., S. Nishimura, T. Matsumoto, H. Ishida, K. Kangawa, N. Minamino, H. Matsuo, M. Ueda, M. Hanaoka, T. Hirose and et al. (1989). "Primary structure and expression from complementary DNA of skeletal muscle ryanodine receptor." Nature 339(6224): 439-445. Talbot, K. (2016). "Clinical tool for predicting survival in ALS: do we need one?" J Neurol Neurosurg Psychiatry 87(12): 1275. Tannemaat, M. R., R. Eggers, W. T. Hendriks, G. C. de Ruiter, J. J. van Heerikhuize, C. W. Pool, M. J. Malessy, G. J. Boer and J. Verhaagen (2008). "Differential effects of lentiviral vector-mediated overexpression of nerve growth factor and glial cell line-derived neurotrophic factor on regenerating sensory and motor axons in the transected peripheral nerve." Eur J Neurosci 28(8): 1467-1479. Taylor, C. A., D. Braza, J. B. Rice and T. Dillingham (2008). "The incidence of peripheral nerve injury in extremity trauma." Am J Phys Med Rehabil 87(5): 381-385. Taylor, J. P., R. H. Brown, Jr. and D. W. Cleveland (2016). "Decoding ALS: from genes to mechanism." Nature 539(7628): 197-206. Towne, C., C. Raoul, B. L. Schneider and P. Aebischer (2008). "Systemic AAV6 delivery mediating RNA interference against SOD1: neuromuscular transduction does not alter disease progression in fALS mice." Mol Ther 16(6): 1018-1025. Toyoshima, C. and G. Inesi (2004). "Structural basis of ion pumping by Ca2+- ATPase of the sarcoplasmic reticulum." Annu Rev Biochem 73: 269-292. Uchitel, O. D., F. Scornik, D. A. Protti, C. G. Fumberg, V. Alvarez and S. H. Appel (1992). "Long-term neuromuscular dysfunction produced by passive transfer of amyotrophic lateral sclerosis immunoglobulins." Neurology 42(11): 2175-2180.

188

Valentine, J. S., P. A. Doucette and S. Zittin Potter (2005). "Copper-zinc superoxide dismutase and amyotrophic lateral sclerosis." Annu Rev Biochem 74: 563-593. Van Damme, P., G. Callewaert, J. Eggermont, W. Robberecht and L. Van Den Bosch (2003). "Chloride influx aggravates Ca2+-dependent AMPA receptor-mediated motoneuron death." J Neurosci 23(12): 4942-4950. Van Damme, P., W. Robberecht and L. Van Den Bosch (2017). "Modelling amyotrophic lateral sclerosis: progress and possibilities." Dis Model Mech 10(5): 537-549. Van Den Bosch, L., P. Van Damme, E. Bogaert and W. Robberecht (2006). "The role of excitotoxicity in the pathogenesis of amyotrophic lateral sclerosis." Biochim Biophys Acta 1762(11-12): 1068-1082. van der Pijl, E. M., M. van Putten, E. H. Niks, J. J. Verschuuren, A. Aartsma-Rus and J. J. Plomp (2016). "Characterization of neuromuscular synapse function abnormalities in multiple Duchenne muscular dystrophy mouse models." Eur J Neurosci 43(12): 1623-1635. van der Weerd, L., M. Tariq Akbar, R. Aron Badin, L. M. Valentim, D. L. Thomas, D. J. Wells, D. S. Latchman, D. G. Gadian, M. F. Lythgoe and J. S. de Belleroche (2010). "Overexpression of heat shock protein 27 reduces cortical damage after cerebral ischemia." J Cereb Blood Flow Metab 30(4): 849-856. van Eijk, R. P. A., M. J. C. Eijkemans, T. A. Ferguson, S. Nikolakopoulos, J. H. Veldink and L. H. van den Berg (2018). "Monitoring disease progression with plasma creatinine in amyotrophic lateral sclerosis clinical trials." J Neurol Neurosurg Psychiatry 89(2): 156-161. Vanrell, L., M. Di Scala, L. Blanco, I. Otano, I. Gil-Farina, V. Baldim, A. Paneda, P. Berraondo, S. G. Beattie, A. Chtarto, L. Tenenbaum, J. Prieto and G. Gonzalez-Aseguinolaza (2011). "Development of a liver-specific Tet-on inducible system for AAV vectors and its application in the treatment of liver cancer." Mol Ther 19(7): 1245-1253. Veldink, J. H., P. R. Bar, E. A. Joosten, M. Otten, J. H. Wokke and L. H. van den Berg (2003). "Sexual differences in onset of disease and response to exercise in a transgenic model of ALS." Neuromuscul Disord 13(9): 737- 743. Vinsant, S., C. Mansfield, R. Jimenez-Moreno, V. Del Gaizo Moore, M. Yoshikawa, T. G. Hampton, D. Prevette, J. Caress, R. W. Oppenheim and C. Milligan (2013). "Characterization of early pathogenesis in the SOD1(G93A) mouse model of ALS: part I, background and methods." Brain Behav 3(4): 335-350. Vinsant, S., C. Mansfield, R. Jimenez-Moreno, V. Del Gaizo Moore, M. Yoshikawa, T. G. Hampton, D. Prevette, J. Caress, R. W. Oppenheim and C. Milligan (2013). "Characterization of early pathogenesis in the SOD1(G93A) mouse model of ALS: part II, results and discussion." Brain Behav 3(4): 431-457.

189

Vleminckx, V., P. Van Damme, K. Goffin, H. Delye, L. Van Den Bosch and W. Robberecht (2002). "Upregulation of HSP27 in a transgenic model of ALS." J Neuropathol Exp Neurol 61(11): 968-974. von Jonquieres, G., N. Mersmann, C. B. Klugmann, A. E. Harasta, B. Lutz, O. Teahan, G. D. Housley, D. Frohlich, E. M. Kramer-Albers and M. Klugmann (2013). "Glial promoter selectivity following AAV-delivery to the immature brain." PLoS One 8(6): e65646. Wang, H., E. J. Sorenson, R. J. Spinner and A. J. Windebank (2008). "Electrophysiologic findings and grip strength after nerve injuries in the rat forelimb." Muscle Nerve 38(4): 1254-1265. Wang, J. T., Z. A. Medress and B. A. Barres (2012). "Axon degeneration: molecular mechanisms of a self-destruction pathway." J Cell Biol 196(1): 7-18. Wang, L., K. Sharma, G. Grisotti and R. P. Roos (2009). "The effect of mutant SOD1 dismutase activity on non-cell autonomous degeneration in familial amyotrophic lateral sclerosis." Neurobiol Dis 35(2): 234-240. Watanabe, H., N. Atsuta, A. Hirakawa, R. Nakamura, M. Nakatochi, S. Ishigaki, A. Iida, S. Ikegawa, M. Kubo, D. Yokoi, H. Watanabe, M. Ito, M. Katsuno, Y. Izumi, M. Morita, K. Kanai, A. Taniguchi, I. Aiba, K. Abe, K. Mizoguchi, M. Oda, O. Kano, K. Okamoto, S. Kuwabara, K. Hasegawa, T. Imai, A. Kawata, M. Aoki, S. Tsuji, K. Nakashima, R. Kaji and G. Sobue (2016). "A rapid functional decline type of amyotrophic lateral sclerosis is linked to low expression of TTN." J Neurol Neurosurg Psychiatry 87(8): 851-858. Waxman, S. G. and R. E. Foster (1980). "Development of the axon membrane during differentiation of myelinated fibres in spinal nerve roots." Proc R Soc Lond B Biol Sci 209(1176): 441-446. Webb, A. A., N. D. Jeffery, N. J. Olby and G. D. Muir (2004). "Behavioural analysis of the efficacy of treatments for injuries to the spinal cord in animals." Vet Rec 155(8): 225-230. Welniarz, Q., I. Dusart and E. Roze (2017). "The corticospinal tract: Evolution, development, and human disorders." Dev Neurobiol 77(7): 810-829. Weydt, P., S. Y. Hong, M. Kliot and T. Moller (2003). "Assessing disease onset and progression in the SOD1 mouse model of ALS." Neuroreport 14(7): 1051-1054. Wijesekera, L. C. and P. N. Leigh (2009). "Amyotrophic lateral sclerosis." Orphanet J Rare Dis 4: 3. Williamson, T. L. and D. W. Cleveland (1999). "Slowing of axonal transport is a very early event in the toxicity of ALS-linked SOD1 mutants to motor neurons." Nat Neurosci 2(1): 50-56. Wong, M. and L. J. Martin (2010). "Skeletal muscle-restricted expression of human SOD1 causes motor neuron degeneration in transgenic mice." Hum Mol Genet 19(11): 2284-2302. Wood, M. D., S. W. Kemp, C. Weber, G. H. Borschel and T. Gordon (2011). "Outcome measures of peripheral nerve regeneration." Ann Anat 193(4): 321-333.

190

Xu, C., R. Craig, L. Tobacman, R. Horowitz and W. Lehman (1999). "Tropomyosin positions in regulated thin filaments revealed by cryoelectron microscopy." Biophys J 77(2): 985-992. Xue, Y. Q., B. F. Ma, L. R. Zhao, J. B. Tatom, B. Li, L. X. Jiang, R. L. Klein and W. M. Duan (2010). "AAV9-mediated erythropoietin gene delivery into the brain protects nigral dopaminergic neurons in a rat model of Parkinson's disease." Gene Ther 17(1): 83-94. Yamada, T., T. Mishima, M. Sakamoto, M. Sugiyama, S. Matsunaga and M. Wada (2006). "Oxidation of myosin heavy chain and reduction in force production in hyperthyroid rat soleus." J Appl Physiol (1985) 100(5): 1520- 1526. Yamanaka, K., S. Boillee, E. A. Roberts, M. L. Garcia, M. McAlonis-Downes, O. R. Mikse, D. W. Cleveland and L. S. Goldstein (2008). "Mutant SOD1 in cell types other than motor neurons and oligodendrocytes accelerates onset of disease in ALS mice." Proc Natl Acad Sci U S A 105(21): 7594- 7599. Yi, J., C. Ma, Y. Li, N. Weisleder, E. Rios, J. Ma and J. Zhou (2011). "Mitochondrial calcium uptake regulates rapid calcium transients in skeletal muscle during excitation-contraction (E-C) coupling." J Biol Chem 286(37): 32436-32443. Yuen, E. C. and R. K. Olney (1997). "Longitudinal study of fiber density and motor unit number estimate in patients with amyotrophic lateral sclerosis." Neurology 49(2): 573-578. Zajac, F. E. and J. S. Faden (1985). "Relationship among recruitment order, axonal conduction velocity, and muscle-unit properties of type-identified motor units in cat plantaris muscle." J Neurophysiol 53(5): 1303-1322. Zarei, S., K. Carr, L. Reiley, K. Diaz, O. Guerra, P. F. Altamirano, W. Pagani, D. Lodin, G. Orozco and A. Chinea (2015). "A comprehensive review of amyotrophic lateral sclerosis." Surg Neurol Int 6: 171. Zhang, W., M. Narayanan and R. M. Friedlander (2003). "Additive neuroprotective effects of minocycline with creatine in a mouse model of ALS." Ann Neurol 53(2): 267-270. Zhou, C., C. P. Zhao, C. Zhang, G. Y. Wu, F. Xiong and C. Zhang (2007). "A method comparison in monitoring disease progression of G93A mouse model of ALS." Amyotroph Lateral Scler 8(6): 366-372. Zou, Z. Y., Z. R. Zhou, C. H. Che, C. Y. Liu, R. L. He and H. P. Huang (2017). "Genetic epidemiology of amyotrophic lateral sclerosis: a systematic review and meta-analysis." J Neurol Neurosurg Psychiatry 88(7): 540-549.

191

Appendix A: Figures

Figure A.1 Muscle Physiology Rig Set-up

(A) Anesthetized mouse under isoflurane (e) in position for muscle physiology recording. Knee held in place with the knee lock (a) and stimulating electrodes (b) placed over the tibial nerve. Hindpaw is taped into the paw pedal (c) with motor (d). (B) Side-view of measured hindpaw demonstrating paw positioning. Knee lock position in relation to tibial nerve (asterisk) shown in inset.

192

Figure A.2 AAV9-GFP vector map

GFP was expressed under the cytomegalovirus (CMV) enhancer/chicken beta- actin (CBA) promoter.

193

Figure A.3 AAV9-SMN vector map

Human survival motor neuron protein (SMN) was expressed under the CMV enhancer/CBA promoter.

194

Figure A.4 AAV9-HSPB1 vector map

Expression of human wildtype heat shock protein B1 (HSPB1) was under the CMV enhancer/CBA promoter.

195

Appendix B: Tables

Table B.1 Longitudinal outcome measurements in SOD1(G93A) male and wildtype male mice.

SOD1(G93A) male mice (shaded, n = 10); wildtype male mice (unshaded, n = 10). Mean outcome measurements with nominal 95% confidence interval (CI) in brackets. Significant differences between SOD1(G93A) and wildtype male mice indicated in bold. Twitch and tetanic outcome measurements were normalized to mouse body mass. Adjusted p-values: * = p<0.05, ** = p<0.01, *** = p<0.001. Abbreviations: Compound muscle action potential (CMAP), motor unit number estimation (MUNE), and single motor unit potential (SMUP).

Continued

196

Normalized Normalized SMUP Age CMAP (mV) MUNE Twitch (mN- Tetantic (mN- (µV) (PND) [95% CI] [95% CI] m/g) m/g) [95% CI] [95% CI] [95% CI] 375.2 33.8 215.8 [195.3, 0.087 [0.073, 0.48 [305.9, [27.2, 41.0] 237.2] 0.101] ** [0.42, 0.54] * 451.6] 35 419.7 47.2 236.6 [215.2, 0.117 [0.103, 0.59 [346.2, [39.3, 55.7] 259.0] 0.131] ** [0.53, 0.65] * 500.3] 308.0 31.0 222.8 [203.8, 0.083 [0.070, 0.43 [252.4, [25.4, 37.2] * 242.6] 0.095] ** [0.38, 0.48] ** 369.2] 42 390.4 44.6 233.0 [213.6, 0.119 [0.106, 0.57 [327.5, [37.8, 51.9] * 253.1] 0.131] ** [0.52, 0.62] ** 458.9] 251.8 28.3 229.9 [212.3, 0.078 [0.067, 0.39 [204.9, [23.3, 33.8] * 248.3] 0.090] *** [0.34, 0.43] *** 303.5] * 49 367.5 42.5 229.3 [211.8, 0.120 [0.108, 0.55 [310.4, [36.3, 49.2] * 247.6] 0.132] *** [0.51, 0.60] *** 429.5] * 205.2 25.6 237.2 [220.6, 0.074 [0.062, 0.35 [164.1, [21.0, 30.7] ** 254.3] 0.085] *** [0.30, 0.39] *** 250.8] ** 56 350.4 40.9 225.8 [209.6, 0.121 [0.110, 0.54 [296.0, [35.0, 47.3] ** 242.5] 0.133] *** [0.50, 0.59] *** 409.4] ** 166.9 22.9 244.5 [228.8, 0.069 [0.058, 0.31 [130.1, [18.5, 27.7] ** 260.9] 0.080] *** [0.26, 0.35] *** 208.2] *** 63 338.8 39.7 222.2 [207.2, 0.123 [0.111, 0.54 [285.4, [33.9, 46.0] ** 237.8] 0.134] *** [0.49, 0.58] *** 396.8] ***

Continued

197

135.7 20.2 252.0 [236.6, 0.064 [0.053, 0.28 [102.6, [16.1, 24.8] *** 267.9] * 0.076] *** [0.23, 0.32] *** 173.4] *** 70 332.4 39.0 218.7 [204.3, 0.124 [0.112, 0.53 [279.2, [33.2, 45.2] *** 233.5] * 0.135] *** [0.49, 0.58] *** 390.2] *** 17.7 110.7 [80.9, 259.6 [244.1, 0.060 [0.048, 0.25 [13.8, 22.0] *** 145.0] *** 275.6] ** 0.071] *** [0.20, 0.29] *** 331.0 77 38.6 215.2 [201.2, 0.125 [0.113, 0.53 [277.8, [32.9, 44.9] *** 229.7] ** 0.136] *** [0.49, 0.58] *** 388.9] *** 15.2 90.8 [64.1, 267.3 [251.3, 0.055 [0.043, 0.22 [11.6, 19.2] *** 122.1] *** 283.8] *** 0.066] *** [0.18, 0.27] *** 334.7 84 38.7 211.7 [197.6, 0.126 [0.114, 0.54 [281.4, [33.0, 45.0] *** 226.3] *** 0.137] *** [0.49, 0.58] *** 392.6] *** 12.8 75.2 [51.2, 275.1 [258.2, 0.050 [0.038, 0.20 [9.6, 16.4] *** 103.8] *** 292.6] *** 0.061] *** [0.16, 0.25] *** 343.5 91 39.2 [33.4, 208.3 [193.7, 0.126 [0.115, 0.54 [289.8, 45.5] *** 223.3] *** 0.138] *** [0.50, 0.59] *** 401.8] *** 10.5 63.3 [41.4, 283.1 [264.7, 0.045 [0.033, 0.19 [7.6, 13.9] *** 90.0] *** 302.0] *** 0.056] *** [0.14, 0.23] *** 357.6 98 40.1 204.9 [189.5, 0.127 [0.116, 0.56 [302.6, [34.3, 46.5] *** 220.8] *** 0.139] *** [0.51, 0.60] *** 417.1] *** 8.4 54.5 [33.7, 291.1 [270.9, 0.039 [0.028, 0.18 [5.7, 11.5] *** 80.3] *** 312.0] *** 0.051] *** [0.13, 0.22] *** 377.3 105 41.5 201.5 [185.0, 0.128 [0.116, 0.57 [319.4, [35.4, 48.1] *** 218.6] *** 0.140] *** [0.52, 0.62] *** 440.1] *** 6.5 [4.0, 9.5] 48.2 [27.4, 299.2 [276.8, 0.034 [0.021, 0.17 *** 74.8] *** 322.5] *** 0.047] *** [0.11, 0.22] *** 403.1 112 43.3 198.1 [180.3, 0.128 [0.116, 0.59 [338.9, [36.6, 50.5] *** 216.8] *** 0.141] *** [0.54, 0.64] *** 472.9] *** 4.7 [2.4, 7.8] 44.1 [22.0, 307.5 [282.6, 0.029 [0.014, 0.16 *** 73.8] *** 333.5] *** 0.043] *** [0.10, 0.22] *** 435.5 119 45.6 194.8 [175.5, 0.129 [0.115, 0.62 [360.0, [37.9, 54.0] *** 215.1] *** 0.143] *** [0.55, 0.68] *** 518.2] ***

197 Table B.2 Longitudinal outcome measurements in SOD1(G93A) female and wildtype female mice.

SOD1(G93A) female mice (shaded, n = 10); wildtype female mice (unshaded, n = 10). Mean outcome measurements with nominal 95% confidence interval (CI) in brackets. Significant differences between SOD1(G93A) and wildtype female mice indicated in bold. Twitch and tetanic outcome measurements were normalized to mouse body mass. Adjusted p-values: * = p<0.05, ** = p<0.01, *** = p<0.001. Abbreviations: Compound muscle action potential (CMAP), motor unit number estimation (MUNE), and single motor unit potential (SMUP).

Continued

198 Normalized Normalized SMUP Age CMAP (mV) MUNE Twitch (mN- Tetantic (mN- (µV) (PND) [95% CI] [95% CI] m/g) m/g) [95% CI] [95% CI] [95% CI] 45.1 361.7 [293.7, 215.8 [195.3, 0.099 [0.085, 0.45 [37.4, 53.4] 436.8] 237.2] 0.113] [0.39, 0.51] 35 39.4 259.9 [202.8, 236.6 [215.2, 0.097 [0.083, 0.46 [32.2, 47.2] 324.0] 259.0] 0.111] [0.40, 0.52] 42.9 334.4 [276.3, 222.8 [203.8, 0.100 [0.088, 0.46 [36.2, 50.1] 398.1] 242.6] 0.113] [0.40, 0.51] 42 37.9 271.6 [219.5, 233.0 [213.6, 0.105 [0.092, 0.49 [31.7, 44.7] 329.3] 253.1] 0.117] [0.44, 0.54] 40.6 307.5 [255.4, 229.9 [212.3, 0.101 [0.089, 0.46 [34.6, 47.1] 364.4] 248.3] 0.113] [0.41, 0.50] 49 36.9 283.0 [233.1, 229.3 [211.8, 0.111 [0.099, 0.51 [31.1, 43.1] 337.7] 247.6] 0.123] [0.47, 0.56] 38.3 281.0 [232.6, 237.2 [220.6, 0.101 [0.089, 0.45 [32.5, 44.4] 334.1] 254.3] 0.112] [0.41, 0.50] * 56 36.3 293.9 [244.3, 225.8 [209.6, 0.117 [0.105, 0.54 [30.7, 42.3] 348.1] 242.5] 0.128] [0.49, 0.58] * 35.8 255.1 [209.1, 244.5 [228.8, 0.099 [0.088, 0.44 [30.3, 41.8] 305.6] 260.9] 0.111] * [0.40, 0.49] ** 63 36.1 304.3 [253.9, 222.2 [207.2, 0.121 [0.110, 0.56 [30.5, 42.1] 359.3] 237.8] 0.132] * [0.51, 0.60] ** 33.3 229.8 [186.1, 252.0 [236.6, 0.097 [0.086, 0.43 [28.0, 39.2] 278.1] 267.9] * 0.108] ** [0.38, 0.47] *** 70 36.2 314.1 [262.6, 218.7 [204.3, 0.125 [0.113, 0.57 [30.7, 42.3] 370.3] 233.5] * 0.136] ** [0.53, 0.62] *** 30.8 205.2 [163.9, 259.6 [244.1, 0.094 [0.082, 0.41 [25.7, 36.4] 251.2] * 275.6] ** 0.105] ** [0.37, 0.46] *** 77 36.8 323.4 [271.0, 215.2 [201.2, 0.127 [0.116, 0.59 [31.2, 42.9] 380.5] * 229.7] ** 0.139] ** [0.54, 0.63] *** 28.3 181.5 [142.9, 267.3 [251.3, 0.090 [0.078, 0.39 [23.4, 33.6] 224.8] ** 283.8] *** 0.101] *** [0.35, 0.44] *** 84 37.8 332.1 [279.1, 211.7 [197.6, 0.129 [0.117, 0.59 [32.1, 44.0] 389.8] ** 226.3] *** 0.140] *** [0.55, 0.64] *** 25.7 158.8 [123.0, 275.1 [258.2, 0.084 [0.073, 0.37 [21.1, 30.8] * 199.1] *** 292.6] *** 0.096] *** [0.32, 0.41] *** 91 39.2 340.2 [286.7, 208.3 [193.7, 0.130 [0.118, 0.60 [33.4, 45.5] * 398.2] *** 223.3] *** 0.141] *** [0.55, 0.64] *** 23.2 137.1 [103.9, 283.1 [264.7, 0.078 [0.067, 0.34 [18.8, 28.0] ** 174.8] *** 302.0] *** 0.090] *** [0.30, 0.39] *** 98 41.1 347.6 [293.4, 204.9 [189.5, 0.129 [0.118, 0.60 [35.2, 47.5] ** 406.3] *** 220.8] *** 0.141] *** [0.55, 0.65] ***

Continued

199 20.6 116.5 [85.4, 291.1 [270.9, 0.071 [0.060, 0.31 [16.4, 25.4] *** 152.4] *** 312.0] *** 0.083] *** [0.26, 0.36] *** 105 43.4 354.3 [298.2, 201.5 [185.0, 0.128 [0.116, 0.60 [37.2, 50.2] *** 415.2] *** 218.6] *** 0.140] *** [0.55, 0.64] *** 18.2 97.2 [67.2, 299.2 [276.8, 0.063 [0.051, 0.28 [13.9, 23.0] *** 132.8] *** 322.5] *** 0.076] *** [0.22, 0.33] *** 112 46.3 360.3 [299.9, 198.1 [180.3, 0.126 [0.113, 0.59 [39.4, 53.8] *** 426.2] *** 216.8] *** 0.138] *** [0.54, 0.64] *** 15.7 79.3 [49.3, 307.5 [282.6, 0.054 [0.040, 0.24 [11.3, 20.9] *** 116.3] *** 333.5] *** 0.068] *** [0.18, 0.30] *** 119 49.7 365.6 [297.2, 194.8 [175.5, 0.123 [0.108, 0.58 [41.6, 58.5] *** 441.0] *** 215.1] *** 0.137] *** [0.52, 0.64] ***

200 Table B.3 Electrophysiological, contractile and behavioral correlations

Pearson’s correlation coefficients between electrophysiological, contractile and behavioral measurements. Abbreviations: compound muscle action potential (CMAP), motor unit number estimate (MUNE). “Norm” denotes normalized measurements (to mouse body mass). * = p<0.05, ** = p<0.01, ***= p<0.001. Significance set at p<0.05.

Norm. Norm. CMAP Twitch Tetanic Grip MUNE Twitch Tetanic CMAP - 0.84** 0.76* 0.82** 0.73** 0.68* 0.65* Twitch 0.84** - 0.98*** 0.66* 0.92*** 0.4 0.41 Tetanic 0.76* 0.98*** - 0.58 0.92*** 0.25 0.28 Grip 0.82** 0.66* 0.58 - 0.71* 0.51 0.52 MUNE 0.73** 0.92*** 0.92*** 0.71* - 0.41 0.32 Norm. 0.68* 0.4 0.25 0.51 0.41 - 0.98*** Twitch Norm. 0.65* 0.41 0.28 0.52 0.32 0.98*** - Tetanic

201 Table B.4 Counts of motor neurons transduced

Motor neuron transduction percentage was calculated by dividing the number of motor neurons co-labeling HSPB1 and ChAT (HSPB1/ChAT) by the total number of motor neurons quantified (ChAT). n=3 mice, L3-L4 spinal cord segments selected as they contain motor pools that innervate muscles assessed using behavioral and electrophysiological measurements (Nicolopoulos-Stournaras and Iles 1983).

PND50 ChAT HSPB1/ChAT % (mean ± stdv) (mean ± stdv) (mean ± stdv) L3-L4 segments 267 ± 21 108 ± 3 40.6 ± 4.3

202