Glycans and glycoproteins as biomarkers and treatments for muscular dystrophies and

genetic myopathies

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy

in the Graduate School of The Ohio State University

By

Kelly Eileen Crowe

Graduate Program in Molecular Cellular and Developmental Biology

The Ohio State University

2018

Dissertation Committee

Paul T. Martin, Advisor

Kevin M. Flanigan

Louise R. Rodino-Klapac

Noah Weisleder

1 Copyrighted by

Kelly Eileen Crowe

2018

2 Abstract

Neuromuscular diseases are a devastating group of disorders that either directly or indirectly impair muscle function. While much work has been done to address the diagnosis and treatment of these diseases, significant gaps in our understanding remain.

In this work, we have utilized glycans and glycoproteins as diagnostic tools and treatments for neuromuscular diseases. First, we have identified a new serum glycoprotein biomarker that should aid in therapeutic monitoring for Duchenne Muscular

Dystrophy (DMD), a devastating childhood disease characterized by progressive muscle wasting and premature death. We have also created a new model that allows direct visualization of glycan therapies for GNE myopathy, an adult-onset myopathy that commonly results in loss of ambulation.

DMD is caused by in the encoding , a of the dystrophin-associated glycoprotein (DAG) complex which connects the muscle cell to its surrounding extracellular matrix. Without dystrophin, levels of the other DAG complex are also decreased, making the muscle membrane weak and more prone to damage. Many therapies have been proposed to replace dystrophin and restore the DAG complex; however, clinical trials to test these therapies suffer from a lack of reliable biomarker measures. To address this, we have developed an ELISA-based assay to measure serum levels of αDG-N, a DAG complex component that is constitutively

ii cleaved and released into serum. We have shown thatαDG-N is significantly decreased in DMD patient serum as compared to age-matched control serum, a finding that could potentially be exploited in the future as a biomarker in clinical trials for DMD.

GNE myopathy is caused by recessive, hypomorphic mutations in glucosamine

(UDP-N-acetyl)-2-epimerase/N-acetylmannosamine kinase (GNE), which catalyzes the committed step in the biosynthesis of sialic acid (SA), an essential glycan. GNE myopathy pathophysiology is poorly understood, and the only therapy currently being pursued utilizes large doses of oral SA or its precursors given daily. Visualization of cellular changes in muscle SA via GNE or oral SA therapy, however, is not possible by immunostaining due to common presence of SA in all tissues. To address this issue, we have created a novel mouse model of GNE myopathy with a humanized sialylglycome which allows one to visualize tissue incorporation of non-human SAs, in the form of N-glycolylneuraminic acid, using immunohistochemistry. We have used this new animal model of GNE myopathy to compare oral SA therapy, which has been used in clinical trials, and GNE gene therapy, which should permanently prevent disease in this recessive, monogenic disorder. Surprisingly, GNE gene therapy targeting the liver was more effective in increasing levels of SA in the muscle than was muscle-directed

GNE gene therapy. In addition, we have shown that large doses of free SA therapy is indeed is taken up by muscle and incorporated into muscle sarcolemmal membranes, while SA taken up from diet cannot do so. Taken together, these studies address critical deficits in the diagnosis and treatment of neuromuscular diseases through the lens of

iii glycobiology that allow one to better assess therapeutic outcomes in DMD and GNE myopathy.

iv Dedication

I would like to dedicate this work to my husband, Alex Hardin, who is always and forever the grand marshal of my parade.

v Acknowledgments

I am extremely fortunate to have been afforded incredible opportunities and support throughout my graduate career. I would first and foremost like to thank my advisor, Dr. Paul Martin, for his patience and encouragement during my time in graduate school and for his support of my growth as a scientist and an educator. I would also like to thank my current and former committee members, Dr. Kevin Flanigan, Dr. Louise

Rodino-Klapac, Dr. Noah Weisleder, Dr. Brian Kaspar, and Dr. Douglas McCarty, for their advice and mentorship.

I would like to thank current and former members of the Martin lab, including Dr.

Rui Xu, Deb Zygmunt, Megan Cramer, Ying Jia, Ben Hood, Haley Guggenheim, Paul

Thomas, Guohong Shao, Paul Thomas, Beth Golden, and Neha Singhal. I have loved getting to know each of you, and am forever grateful for your support and friendship over the past five years. Thanks to Dr. Shibi Likhite, Dr. Nico Wein, and everyone in the

Harper lab for all their help and encouragement. And a special thanks to Gennie Faber, my first lab mentee, who set a high bar for future students.

I am incredibly grateful for the many people who have invested in my growth and development off the bench. Thanks to the members of the executive board of Sigma Xi, including Dr. Mark Peeples, Dr. Lynette Rogers, Dr. Larry Feth, and especially Dr. Noah

Weisleder, for providing me with one of the most transformative opportunities in my

vi graduate career. Thanks also to Kathy Wallace, Dr. Heather Rhodes, Dr. Amy Santas, members of RITA, and the whole team behind Mechanisms of Human Health and

Disease for shaping me as an educator.

Finally, I would like to thank my friends and family, including my husband, mom, dad, Dayla Kratzer, and Kaitlin Snider, who never missed an opportunity to brighten my days.

vii Vita

May 2007 New Covenant Academy

May 2011 B.A. Biology, B.A. Psychology, Drury University

2013 to present Graduate Research Fellow, Molecular Cellular and Developmental

Biology Program, The Ohio State University

Publications

Zygmunt, D., Singhal, N., Kim, M. L., Cramer, M. L., Crowe, K. E., Xu, R., Jia, Y.,

Adair, J., Martinex-Pena Y Valenzuela, I., Akaaboune, M., White, P., Janssen, Pl. M., &

Martin, P. T. (2017). Deletion of Pofut1 in mouse skeletal myofibers induces muscle aging-related phenotypes in cis and in trans. Molecular and Cellular Biology, 37(10), e00426-16. doi: 10.1128/MCB.00426-16

Zygmunt, D., Crowe, K. E., Flanigan, K. M., & Martin, P. T. (2017). Comparison of serum rAAV serotype-specific antibodies in patients with Duchenne Muscular

Dystrophy, Becker Muscular Dystrophy, Inclusion Body Myositis or GNE myopathy.

Human Gene Therapy, 28(9), 737-746. doi:10.1089/hum.2016.141

viii Crowe, K. E., Shao, G., Flanigan, K. M., & Martin, P. T. (2016). N-terminalα

Dystroglycan(αDG-N): A Potential Serum Biomarker for Duchenne Muscular

Dystrophy. JND Journal of Neuromuscular Diseases, 3(2), 247-260. doi:10.3233/jnd-

150127

Fields of Study

Major Field: Molecular Cellular and Developmental Biology

ix Table of Contents

Abstract...... ii Dedication...... v Acknowledgments...... v Vita...... viii Table of Contents...... x List of Tables ...... xiii List of Figures...... xiv Chapter 1. Introduction ...... 1 1.1 Introduction to neuromuscular disorders ...... 1 1.2 Muscular Dystrophies ...... 1 1.2.1 Duchenne Muscular Dystrophy (DMD) and Becker Muscular Dystrophy (BMD)...... 1 1.2.2 Animal models of DMD ...... 2 1.2.3 Treatments for DMD...... 4 1.2.4 Need for biomarkers for DMD ...... 6 1.2.5 Current biomarkers for DMD ...... 7 1.2.6 αDG-N as a potential biomarker for DMD...... 8 1.3 Serum antibodies as a barrier to gene therapy in DMD...... 9 1.3.1 Adeno-associated virus (AAV)...... 9 1.3.2 Immune response to AAV ...... 10 1.4 The role of CMAH in DMD...... 12 1.4.1 Inflammation in the muscular dystrophies...... 12 1.4.2 The effect of CMAH on the sialylglycome across species...... 12 1.4.3 Altered immune response in Cmah-/- mdx mice ...... 13 1.5 Genetic myopathies...... 14 1.5.1 GNE myopathy ...... 14 x 1.5.2 Mouse models of GNE myopathy...... 17 1.5.3 Possible mechanisms of muscle damage in GNE myopathy...... 19 1.5.4 Treatments strategies for GNE myopathy...... 21 Chapter 2. Serum studies in the muscular dystrophies ...... 24 2.1 Abstract...... 24 2.2 Introduction...... 25 2.3 Materials and Methods...... 28 2.3.1αDG-N protein production and purification ...... 28 2.3.2 Protein biotinylation...... 28 2.3.3 Serum αDG-N ELISA assays ...... 29 2.3.4 Western blots ...... 33 2.3.5 Immunofluorescence staining ...... 34 2.3.6 Statistics for αDG-N ...... 35 2.3.7 Human serum samples for rAAV antibody titer...... 35 2.3.8 ELISA to identify serum rAAV antibodies ...... 36 2.3.9 Statistics for rAAV antibody titer...... 37 2.3.10 Glycan arrays ...... 37 2.4 Results...... 38 2.4.1DevelopmentofanαDG-N ELISA assay...... 38 2.4.2αDG-N is decreased in the serum of patients with DMD...... 41 2.4.3SerumlevelsofαDG-N do not differ significantly based on age, gender, or ambulation status ...... 44 2.4.4 mdx Utrn-/- mice, but not mdx mice, show decreased expression of serum αDG-N compared to normal mice ...... 44 2.4.5GRMDdogsshowatrendtowardsdecreaseinexpressionofserumαDG-N compared to GR dogs...... 47 2.4.6 Characterization of human serum rAAV antibodies...... 48 2.4.7 Assessment of antibody titer to Neu5Gc-containing glycans in DMD patients ...... 52 2.5 Discussion...... 53 2.6 Acknowledgements...... 59 Chapter 3. Glycan-based treatments for genetic myopathies...... 89 3.1 Abstract...... 89

xi 3.2 Introduction...... 90 3.3 Materials and Methods...... 93 3.3.1 Mice ...... 93 3.3.2 Grip strength and treadmill ambulation ...... 93 3.3.3 Muscle physiology...... 94 3.3.4 Histology...... 95 3.3.5 Immunofluorescence staining ...... 95 3.3.6 Neu5Gc serum titering...... 96 3.3.7 Cloning and production of rAAV constructs...... 96 3.3.8 In vitro transfection...... 98 3.3.9 Infection with rAAV...... 98 3.3.10 Quantitative polymerase chain reaction (qPCR)...... 99 3.3.11 Semi-quantitative real-time polymerase chain reaction (qRT-PCR)...... 101 3.3.12 Western blotting and immunoprecipitation ...... 102 3.3.13 Luciferase assay...... 102 3.3.14 Statistics...... 103 3.4 Results...... 103 3.4.1 Characterization of the Gne(−/−)hGNED176V-Tg and Gne(−/−)hGNED176V- TgCmah(−/−)models of GNE myopathy ...... 103 3.4.2 High dose gavage of free SA, but not long-term SA-enriched food, can be incorporated into muscles and organs in the Gne(−/−)hGNED176V-TgCmah(−/−) model...... 106 3.4.3 Liver-specific expression of GNE gene therapy is more effective than muscle- specific GNE gene therapy at increasing muscle SA levels, but may lead to liver pathology...... 109 3.4.4 GNE overexpression does not recapitulate GNE myopathy histopathology . 115 3.4.5 Gne knockdown may recapitulate some GNE myopathy histopathology ..... 117 3.5 Discussion...... 119 3.6 Acknowledgements...... 123 Chapter 4. Conclusions and future directions...... 179 Bibliography ...... 187

xii List of Tables

Table 1. Summary of relevant patient information...... 64 Table 2. Antibody titers to rAAV serotypes in young and adult normal human subjects 77 Table 3. Antibody titers to rAAV serotypes in patients with Duchenne Muscular Dystrophy (DMD) and Becker Muscular Dystrophy (BMD)...... 79 Table 4. Antibody titers to rAAV serotypes in patients with Inclusion Body Myositis (IBM) and GNE Myopathy (GNE)...... 81 Table 5. Injection information for expression of tissue-specific GNE in vivo...... 146

xiii List of Figures

Figure1.Standardcurvesof2A3and3B4antibodybindingtoαDG-N protein using different ELISA assay methods ...... 61 Figure 2. Total αDG-N ELISA signal and calculated serum concentration are decreased in serum from patients with DMD compared to normal...... 63 Figure 3. Western blot analysis of αDG-N expression in normal human and DMD serum ...... 66 Figure 4. Signals from serum αDG-N ELISA are independent of age in otherwise normal, BMD and DMD patients, and are independent of gender for normal patients ...... 67 Figure 5. Signals from serum αDG-N ELISA are independent of ambulation status in BMD and DMD patients...... 68 Figure 6. Serum αDG-N signal is decreased in mdx Utrn-/- mice relative to wild type and mdx mice...... 69 Figure 7. Reduction in serum αDG-N signal in mdx Utrn-/- mice as compared to mdx mice is restored in mct mdx Utrn-/- mice ...... 71 Figure 8. Immunostaining for αDG-N reveals little or no expression in adult mouse skeletal muscle but positive staining in intramuscular peripheral nerves...... 73 Figure 9. Western blotting for αDG-N reveals little or no expression in adult mouse skeletal muscle...... 74 Figure 10. Serum αDG-N signal is not significantly decreased in GRMD relative to GR dogs...... 76 Figure 11. Antibody titers to rAAV serotypes in young and adult normal human subjects ...... 82 Figure 12. Antibody titers to rAAV serotypes in patients with Duchenne or Becker Muscular Dystrophy, Inclusion Body Myositis, or GNE Myopathy...... 83 Figure 13. Average reciprocal dilution factor for positive signal and propensity of positive signals in young and adult normal patients compared to DMD, BMD, IBM and GNE patients...... 84 Figure 14. Antibody titer to Neu5Gc-containing glycans in otherwise normal and DMD patients ...... 85 Figure 15. Analysis of Neu5Ac and Neu5Gc serum titers by sialic acid type, linkage type, and age...... 86 Figure 16. Schematic of aDG-N processing and potential changes in DMD ...... 87 Figure 17. Gne(−/−)hGNED176V-Tg mice largely do not differ from Gne(−/−)hGNED176V-TgCmah(−/−)mice in lifespan, body weight, muscle weight, or organ weight up to 15 months of age...... 124 Figure 18. GNE myopathy mouse model gross motor function study timeline ...... 126 xiv Figure 19. Gne(−/−)hGNED176V-Tg mice do not differ from Gne(−/−)hGNED176V- TgCmah(−/−) mice in gross motor ability up to 15 months of age ...... 127 Figure 20. H&E and Congo Red staining of gastrocnemius muscle in Gne(−/−)hGNED176V-Tg mice, Gne(−/−)hGNED176V-TgCmah(−/−) mice, and littermate controls...... 129 Figure 21. Electrophysiology measures in Gne(−/−)hGNED176V-Tg mice and littermate controls...... 131 Figure 22. Several Gne(−/−)hGNED176V-TgCmah(−/−) but not Gne(−/−)hGNED176V-Tg mice have increased Neu5Gc antibody serum titers at 15 months of age ...... 133 Figure 23. Endogenous Neu5Gc is absent in Gne(−/−)hGNED176V-TgCmah(−/−) muscles, but overall SA loss is equivalent in Gne(−/−)hGNED176V-Tg and Gne(−/−)hGNED176V- TgCmah(−/−)muscles ...... 134 Figure 24. Method to study short-term, high-dose exogenous sialic acid incorporation in a GNE myopathy model ...... 135 Figure 25. Neu5Gc is incorporated into Gne(−/−)hGNED176V-TgCmah(−/−) muscle tissues after oral gavage...... 136 Figure 26. Neu5Gc is incorporated into several Gne(−/−)hGNED176V-TgCmah(−/−) organs after oral gavage...... 138 Figure 27. Method to study long-term, low-dose exogenous sialic acid incorporation in a GNE myopathy model ...... 139 Figure 28. Neu5Gc is not incorporated into Gne(−/−)hGNED176V-TgCmah(−/−) muscle tissues after feeding of PSM food...... 140 Figure 29. Neu5Gc is not incorporated into Gne(−/−)hGNED176V-TgCmah(−/−) organs after feeding of PSM food...... 142 Figure 30. Constructs for tissue-specific expression of GNE in vivo ...... 144 Figure 31. The LSP promoter is expressed in liver cells in vitro ...... 145 Figure 32. Method to study muscle- or liver-specific expression of GNE in Gne(−/−)hGNED176V-Tg and Gne(−/−)hGNED176V-TgCmah(−/−) mice ...... 147 Figure 33. qPCR and qRT-PCR analysis for rAAV injection of GNE with tissue-specific promoters ...... 148 Figure 34. Western blot analysis of GNE protein expression in muscle and liver extracted with varying detergents...... 150 Figure 35. The LSP promoter expresses GNE in liver in vivo...... 151 Figure 36. Normalized muscle and organ weights of Gne(−/−)hGNED176V- TgCmah(−/−)mice injected with GNE targeted to muscle or liver ...... 152 Figure 37. Intraperitoneal injection of GNE under a liver-specific promoter increases muscle and liver sialic acid in Gne(−/−)hGNED176V-Tg mice ...... 154 Figure 38. Intraperitoneal injection of GNE under a liver-specific promoter increases liver sialic acid...... 156 Figure 39. Intraperitoneal injection of GNE under a liver-specific promoter increases muscle sialic acid ...... 158 Figure 40. H&E staining of liver and skeletal muscle and Congo Red staining of skeletal muscle in virus-injected Gne(−/−)hGNED176V-TgCmah(−/−)mice...... 160

xv Figure 41. H&E staining of liver in virus-injected Gne(−/−)hGNED176V-TgCmah(−/−) mice...... 161 Figure 42. Immunoprecipitation and western blot analysis of wild-type or mutant GNE delivered under tissue-specific promoters ...... 162 Figure 43. Wild type and mutant GNE under a muscle-specific promoter are expressed in vivo...... 163 Figure 44. H&E staining of gastrocnemius in mice injected with wild-type or mutant GNE...... 165 Figure 45. Congo Red staining of gastrocnemius in mice injected with wild-type or mutant GNE ...... 167 Figure 46. Artificial miR shuttle constructs for knockdown of Gne expression ...... 169 Figure 47. Luciferase expression and qRT-PCR analysis for GNE miR shuttle constructs in vitro...... 170 Figure 48. In vivo miRNA shuttle study construct map and timeline ...... 171 Figure 49. rAAV.U6.GNEmiR4.CMV.GFP is expressed in vivo...... 172 Figure 50. qPCR and qRT-PCR analysis for rAAV delivery of GNE miR shuttle...... 173 Figure 51. Normalized muscle weights of Gne(+/−)mice injected with rAAV.U6.GNEmiR4.CMV.GFP...... 175 Figure 52. H&E and Congo Red staining of tibialis anterior in mice injected with rAAV.U6.GNEmiR4.CMV.GFP ...... 176 Figure 53. Congo Red positivity of tibialis anterior in mice injected with rAAV.U6.GNEmiR4.CMV.GFP ...... 178

xvi Chapter 1. Introduction

1.1 Introduction to neuromuscular disorders

Neuromuscular disorders are a broad class of acquired or genetic ailments that affect muscle function. They include diseases that primarily impact the nervous system, such as Spinal Muscular Atrophy, Amyotrophic Lateral Sclerosis, and Charcot-Marie-

Tooth disease, and those that primarily impact the muscles themselves, such as Duchenne muscular dystrophy (DMD), Limb Girdle Muscular Dystrophies, Congenital Muscular

Dystrophies, and genetic myopathies such as GNE myopathy. Dystrophies are characterized by progressive cycles of muscle degeneration and regeneration, often leading to replacement of the muscle by fatty or fibrotic tissue. Genetic myopathies present with muscle weakness; however, the histopathology, while unique to each disorder, lacks certain hallmarks of muscular dystrophy such as muscle inflammation1.

1.2 Muscular Dystrophies

1.2.1 Duchenne Muscular Dystrophy (DMD) and Becker Muscular Dystrophy (BMD)

Duchenne muscular dystrophy (DMD) is a severe, progressive, X-linked muscle disease of childhood caused by mutations in the dystrophin (DMD) gene that lead to loss of expression of the dystrophin protein2, 3 Diagnosis of DMD typically occurs shortly after children begin to ambulate, with loss of ambulation occurring in the early teenage years and death typically ensuing in the third decade of life due to respiratory

1 insufficiency and/or cardiomyopathy4. Becker muscular dystrophy (BMD) is also caused by mutations in the DMD gene, but muscle cells in BMD patients are able to express mutated, partially functional dystrophin protein that allows for a generally less severe clinical phenotype5. Loss of dystrophin expression increases the frailty of the muscle cell membrane (also known as the sarcolemma), thereby leading to membrane rupture and loss of calcium homeostasis that ultimately causes muscle wasting6-9. This membrane frailty in DMD results, in part, from the failure of dystrophin protein to properly anchor components of the dystrophin-associated glycoprotein (DAG) complex in the sarcolemmal membrane, which in turn weakens connections between the extracellular matrix, the sarcolemmal membrane, and the intracellular F-actin cytoskeleton10-14.

1.2.2 Animal models of DMD

Spontaneous dystrophin mutations in several animals allowed for the development of DMD animal models. These spontaneous mutant models in animals such as mice and dogs have informed much of subsequent research in DMD, even after human

DMD gene cloning was achieved in 19873.

The most common mouse model of DMD is the mdx mouse. The mdx mouse genome contains a spontaneous nonsense in 23 of the DMD gene, resulting in loss of the encoded full-length dystrophin protein. Despite this loss of dystrophin, the mdx mouse model displays a milder phenotype than patients with DMD.

Myofiber necrosis and inflammatory cell infiltration are evident in mdx muscles, peaking around one month of age and then largely stabilizing until the mouse reaches nearly two years of age. Despite showing relevant histopathology, the mdx mouse has limitations to 2 its use as a model for DMD due to its almost normal lifespan and its relatively mild clinical phenotypes15. Because of these shortcomings, several additional mouse models have been created to develop a more severe mouse model of DMD; these models typically combine the mdx mutation with additional genetic deletions to exacerbate the disease phenotype. One such model, the mdx Utrn-/- mouse, combines dystrophin loss with deletion of utrophin, a dystrophin homologue that can substitute for dystrophin in stabilizing the DAG complex. In healthy mammals, utrophin expression is generally restricted to the neuromuscular and myotendinous junctions in postnatal skeletal muscle.

Utrophin is also present, however, in adult regenerating muscle fibers, which are prevalent in the case of muscular dystrophy16, indicating that utrophin may play a role in modulating the muscular dystrophy phenotype. In fact, the mdx Utrn-/- model has a much more severe phenotype than its mdx counterpart, exhibiting reduced body weight, severe

DMD-like histopathology, and premature death17.

Another commonly used model for DMD is a result of a spontaneous splice site mutation in the golden retriever (GR) breed of dog, resulting in the GR muscular dystrophy (GRMD) model. The GRMD dog is generally severe clinically, exhibiting decreased body weight, weakness, and gait abnormalities before three months of age. In addition, these GRMD models show histopathology more akin to DMD muscle, including increased myofiber size variability, necrosis, and inflammation relative to the mdx mouse. Notably, the GRMD model has marked heterogeneity in disease severity; as a result, GRMD dogs are generally classified as either mildly or severely affected18. As

3 the GRMD model is not an inbred line, it is thought that this variance is a result of genetic disease modifiers such as SPP1 and Jagged119, 20.

1.2.3 Treatments for DMD

Several classes of therapies have been developed for DMD that seek to restore or replace dystrophin in the DAG complex. These include dystrophin or microdystrophin gene therapy, dystrophin gene , dystrophin therapies that stimulate read- through of nonsense mutations, and stem cell therapy.

Gene therapy allows for delivery of a gene to affected tissues through use of an attenuated viral vector such as recombinant adeno-associated virus (rAAV). Gene therapy can be utilized in DMD for either gene replacement or surrogate gene therapy. In gene replacement for DMD, a form of dystrophin gene is delivered to cells to restore the missing dystrophin protein and to thereby stabilize the DAG complex. However, this idea is complicated by the sheer size of the DMD gene; the full-length dystrophin cDNA is about 11kb, while the packaging capacity of rAAV is a mere 4.7kb. To circumvent this limitation, truncated versions of dystrophin have been developed, namely minidystrophin and microdystrophin, which are approximately one half and one third of the size of full- length dystrophin, respectively21. As such, they are more amenable to packaging in rAAV for gene replacement, with microdystrophin deliverable via a single rAAV vector and minidystrophin deliverable via dual rAAV vectors22. Versions of mini- and microdystrophin have been utilized in intramuscular gene therapy clinical trials, and are now beginning to be used in intravenous trials.

4 In addition to gene replacement, gene therapy allows for delivery of surrogate to treat DMD. One such surrogate gene is GALGT2 (B4GALNT2), which encodes a

β1-4-N-acetyl-D-galactosamine(βGalNAc)glycosyltransferasethatformsthecytotoxic

T-cell (CT) glycan. The CT glycan is normally localized to the adult neuromuscular junction, but GALGT2 overexpression results in ectopic expression of the CT glycan in the extrasynaptic muscle membrane. This ectopic expression, in turn, induces the ectopic overexpression of several dystroglycan-binding proteins that are normally restricted to the adult neuromuscular junction, including utrophin, laminins, agrin, and plectin.

Overexpressionoftheseproteins,incombinationwithstrengtheningofthebindingofα dystroglycan to the extracellular matrix, allows for an amelioration of the dystrophic phenotype, as evidenced by changes in physiology and histopathology. This amelioration has been shown through both muscle-specific transgenic GALGT2 overexpression in an mdx background (mCT mdx line) and through postnatal, muscle-specific delivery of

GALGT2 to mdx mice23, 24. Recently, this surrogate gene therapy has been delivered in a phase I through intravenous limb infusion (ILI).

A second class of therapies for DMD alters the translation of the dystrophin transcript to allow for dystrophin expression. In exon skipping, antisense

(AONs) bind the splice site of the exon containing the frame-shifting mutation, preventing inclusion of the mutant exon in the transcript. This allows for expression of the transcript, albeit with an internal truncation, rather than degradation of the transcript through nonsense-mediated decay. Thus, production of a truncated dystrophin transcript through AON therapy has the potential to shift the patient phenotype from a DMD-like

5 phenotype to a BMD-like phenotype25. One such exon-skipping drug, , recently became the first FDA-approved drug for DMD26. Exon-skipping therapies with chemical modifications, like and peptide nucleic acids, have also been developed to alter stability and immunogenicity27, 28. Similarly, in read-through therapy, pharmacological agents cause a conformational change in the dystrophin mRNA, allowing for substitution of a single amino acid in the presence of a premature . This approach leads to production of a dystrophin transcript with minimal alterations29.

Another class of therapies is stem-cell therapies, where dystrophic muscles are treated with dystrophin-containing progenitor cells. This method, in theory, both delivers dystrophin to the affected muscle and refreshes the limited regenerative capacity of muscle that is taxed in muscular dystrophy. These stem cells are typically derived either from healthy donors or from the patients themselves, which necessitates ex vivo correction of the DMD gene30. While several of these stem cell therapies are undergoing clinical trials, there are some limitations in their potential due to the limited efficiency of delivery to skeletal muscle, poor implantation and survival of transplanted cells, and stem cell-mediated increases in immunogenicity31.

1.2.4 Need for biomarkers for DMD

Although many therapies seeking to replace dystrophin and stabilize the DAG complex in DMD are in development, clinical trials suffer from a lack of reliable outcome measures32. Rather, a single muscle biopsy is often used to determine the extent of dystrophin restoration33. Unfortunately, muscle biopsies are not only invasive, but can 6 give an inaccurate picture of therapeutic efficacy, as dystrophin restoration can vary widely after therapy, even in the same muscle34. These difficulties have complicated clinical trials for DMD in practice. In fact, Eteplirsen, a drug for DMD which seeks to restore dystrophin to the muscle membrane by restoring the DMD frame, had difficulties in receiving FDA approval due in part to a lack of confidence in dystrophin quantification35.

1.2.5 Current biomarkers for DMD

The classic serum biomarker for DMD is increased activity of creatine kinase (CK), a muscle protein that is released into the serum as the result of membrane perforation36. Elevation of serum CK activity is evident in DMD patients even at birth, and serum CK levels are typically ten to one hundred-fold higher than those found in otherwise normal patients37, 38. Serum CK activity, however, can also be elevated by exercise in non-dystrophic patients and by a variety of other muscle insults, such as viral infections. As such, elevated serum CK activity is not a specific marker for DMD and elevations can be highly variable even within individual DMD patients39, 40. Similarly, increased levels of serum cardiac troponin can be found in DMD patients, many of whom develop dilated cardiomyopathy, but this again can be found in a variety of other cardiac events not specific to DMD41.

Because of the need to identify markers of global DMD disease that would not require a muscle biopsy, a number of proteomics studies have been initiated to identify further serum markers that change in DMD. While certainly not an exhaustive list, additional protein markers identified as elevated in DMD serum include fibronectin42, 7 titin43, myomesin 343, filamin C43, actin43, phosphoglycerate mutase 243, myoglobin43, 44, fibrinogen43, 44, matrix metalloproteinase 945, tissue inhibitor of matrix metalloproteinase

145, osteopontin46 and follistatin47. In addition, some urine proteins, including titin48, and several serum micro-, including miR-149, 50, miR-13349, 50, and miR-20649, 50, are elevated in DMD patients. This list suggests that a myriad of molecular changes can arise as muscle damage occurs in DMD patients. Some of these markers are further suggested to have altered expression as muscle pathology and clinical findings progress39.

1.2.6 αDG-N as a potential biomarker for DMD

α dystroglycan can be cleaved in muscle cells to generate a secreted N-terminal fragment, αDG-N, that Matsumura and colleagues first showed can be observed in human serum as a glycosylated protein51. αDG-N is normally cleaved by furin in the Golgi apparatus as dystroglycan is being secreted to the cell surface, and αDG-N is removed from α dystroglycan in skeletal muscle13, 52, 53. Thus, αDG-N is released by muscle, but is also likely released by non-muscle cells where dystroglycan is also normally expressed.

Because dystroglycan protein is expressed in many tissues, serum αDG-N likely reflects a collection of cleaved dystroglycan proteins emanating from various tissues throughout the body, however, skeletal muscle comprises a significant fraction of this tissue. In addition to serum, αDG-N has also been identified in the cerebrospinal fluid, lachrymal fluid and urine fluid54, 55. αDG-N expression in the cerebrospinal fluid is elevated in patients with Lyme neuroborrelosis, suggesting that αDG-N expression may be altered in certain disease states54, an idea proposed by Brancaccio and colleagues for the muscular dystrophies56. 8 1.3 Serum antibodies as a barrier to gene therapy in DMD

1.3.1 Adeno-associated virus (AAV)

The adeno-associated virus (AAV) is a small, single-stranded DNA virus that has been widely used for gene therapy. It has been promising as a viral vector because it is able to infect dividing and post-mitotic cells, leading to persistent transgene expression and with less activation of an immune response than found with other viral vectors57.

Wild type AAV is inserted into the human genome on the long arm of chromosome 19, and is also maintained outside the host genome as an episomal plasmid58. Wild-type AAV requires a secondary host virus such as adenovirus or herpes simplex virus to complete its replication cycle. When such a virus is also present, AAV can complete its life cycle, packaging and releasing virus59. To accomplish this, AAV has a genome with two open reading frames, rep and cap, enclosed in a capsid. rep is vital for viral replication, whereas cap is necessary for self-assembly of the capsid. However, when used for therapy, rep and cap are removed, rendering the virus replication- incompetent and preventing integration60. These genes are then added in trans along with ageneencodingthenecessary“helper”genesfromtheAdenovirusgenometo produce

AAV in cultured cells61.

Various serotypes of AAV have been isolated, which are characterized by the inability to cross-react with antibodies of other serotypes57. These serotypes show differing tissue distributions due to changes in the capsid protein; AAV serotypes 1-9 share approximately 45% amino acid homology in the capsid protein62. AAV2 is the most widely studied AAV serotype, but other types have been widely used to transduce 9 specific tissues; for instance, AAV8 shows higher expression in the heart and AAV9 shows stronger expression in the lung62. These sorts of differences have been exploited in gene therapy for diseases that would be benefitted by tissue-specific therapeutic expression.

1.3.2 Immune response to AAV

As a virus, AAV can activate both the innate and adaptive immune system responses, causing cytokine release, antibody production, and killing by cytotoxic T- cells63. In fact, an immune response can be seen in 40-70% of patients treated with AAV gene therapy64.

Both humoral and cellular immune responses to virally-introduced transgenic proteins have been problematic for the use of AAV in gene therapy. This immune response is a concern because of both patient safety and for the necessity for long-term, stable gene expression. It is also an issue for potential re-administration of gene therapies; the humoral response can produce neutralizing antibodies against the capsid, making re- administration difficult, and the cytotoxic T-cell response can potentially target and destroy transduced cells65.

1.3.3 Immune response as a barrier to AAV gene therapy treatment in neuromuscular diseases

One of the main impediments to clinical use of rAAV vectors is the suppression of tissue transduction by host humoral immune response to the viral capsid protein66. This is the case not only for neutralizing rAAV antibodies, which can block rAAV infection of cells in vitro and/or in vivo, but also for non-neutralizing antibodies, which can impact 10 tissue transduction through Fc-mediated capsid uptake into dendritic cells and macrophages and through increasing tissue inflammation66. Several strategies have been tested to attempt to bypass the block of tissue transduction by pre-existing rAAV serum antibodies. These include immune suppression through the use of the B cell inhibitor rituximab, plasmapheresis to remove serum antibodies, and direct tissue injection to bypass access of serum antibodies to delivered rAAV66. In an isolated limb perfusion study where rAAV was delivered intra-arterially, both rAAVrh74.MCK.GALGT2 and rAAVrh74.MCK.µDystrophin vectors were able to yield significant (near 50%) transduction of skeletal myofibers in the targeted gastrocnemius muscles if macaques had total serum antibody levels to rAAVrh74 that were positive below a 1:800 dilution67. In macaques where serum rAAVrh74 titers were positive at a dilution of 1:800 or higher prior to treatment, muscle transduction rarely exceeded 10% of total muscle cells. Use of plasmapheresis prior to treatment in rAAV antibody-positive macaques, however, lowered serum antibody levels at the time of treatment to below the1:800 threshold and allowed for muscle transduction at levels that were equivalent to those in macaques with no pre-existing serum rAAVrh74 antibodies. Thus, the presence of low levels of pre- existing rAAV serum antibody were not inhibitory to muscle tissue transduction, while higher levels of serum antibody were inhibitory.

A number of studies have shown humans exposed to AAV develop serum rAAV antibodies that react with multiple of rAAV serotypes, including rAAV1, rAAV2, rAAV5, rAAV6, rAAV8, rAAV9, rAAVrh1066. In general, these findings suggest that neutralizing antibodies to rAAV1 and rAAV2 are present in between 30 and 60% of all

11 humans, while such antibodies are only found in 15 to 30% of subjects for rAAV7, 8 and

9. Serum rAAV antibodies for individual serotypes can also vary by geographic location and by age. It is unclear the extent to which chronic inflammatory conditions, for example the muscular dystrophies, might influence such propensities.

1.4 The role of CMAH in DMD

1.4.1 Inflammation in the muscular dystrophies

Inflammation and infiltration of immune cells in the muscle tissue are hallmarks of the muscular dystrophies. In DMD, lack of dystrophin destabilizes the DAG complex, causingthemusclemembranetobecomeleaky.Thisleakinessreleasesthemyofiber’s contents into the extracellular space68. Many of the release contents can be potent activators of innate immunity, causing cytokine release and inflammation. Meanwhile, necrosis of myofibers leads to sequential recruitment of neutrophils, M1 macrophages, and M2 macrophages to clear the damaged cells. However, as DMD myofibers are in a chronic, asynchronous state of degeneration and regeneration, this necrosis is never fully resolved. This leads to the chronic presence of inflammatory infiltrates characteristic of

DMD biopsies69. These inflammatory and immune phenotypes are recapitulated, at least in part, in models of muscular dystrophy, including the GRMD dog model and the mdx mouse model18, 70.

1.4.2 The effect of CMAH on the sialylglycome across species

Sialic acids (SAs) are an essential class of glycans that are typically found on the terminal end of glycan chains on glycoproteins and glycolipids. Due in part to their

12 localization and negative charge, they play a critical role in processes such as cell-cell signaling, immune recognition, ion transport, and stability of secreted glycoproteins71. In most mammals, there are two predominant forms of SA: N-acetylneuraminic acid

(Neu5Ac) and N-glycolylneuraminic acid (Neu5Gc), which differ by only a single hydroxyl group. In the SA biosynthetic pathway, approximately half of Neu5Ac is converted to Neu5Gc through cytidine monophosphate (CMP)-N-acetylneuraminic acid hydroxylase (Cmah), so most non-human animals express around a 50:50 mixture of

Neu5Ac and Neu5Gc forms of SA72. However, due to in inactivating mutation in the

CMAH gene, humans are unable to convert Neu5Ac to Neu5Gc, and thus predominately express the Neu5Ac form of SA while having no discernable Neu5Gc73.

1.4.3 Altered immune response in Cmah-/- mdx mice

The unique SA repertoire of humans has proven to be a modifier of several pathologies, including cancer, infectious diseases, and muscular dystrophies74. One such study found that the mdx model of muscular dystrophy was exacerbated by loss of Cmah expression75. Mice, like most non-human mammals, express a roughly equal mix of

Neu5Ac and Neu5Gc. This study found that when an mdx model was combined with a

Cmah knockout model, this novel DMD model with a humanized SA repertoire exhibited increases in mortality, muscle weakness, and histopathology. It was found that these

Cmah-/- mdx had reduced levels of DAG complex proteins, and that α dystroglycan glycosylated with Neu5Ac-capped sugar chains had decreased binding to extracellular matrix proteins as compared to Neu5Gc-glycosylated α dystroglycan75.

13 In addition to this decrease in abundance and strength of DAG complex- extracellular matrix interactions, the study also found alterations in Cmah-/- mdx immune responses. As Neu5Gc is not endogenously expressed in Cmah-/- mice (as is the case in humans), exogenously derived Neu5Gc from dietary sources can integrate into tissues and act as a xeno-autoantigen, provoking an immune response. This study found that

Cmah-/- mdx, but not Cmah-/- or mdx mice, developed an antibody titer to Neu5Gc. As newly regenerating fibers in Cmah-/- mdx muscles seemed to incorporate dietary Neu5Gc, it seems likely that lack of endogenous Neu5Gc expression coupled with a large population of regenerating fibers triggered this antibody production. These Neu5Gc antibodies, in turn, seemed to be activating complement-mediated killing of Neu5Gc- containing, regenerating myofibers, exacerbating histopathology in this model75.

1.5 Genetic myopathies

1.5.1 GNE myopathy

GNE myopathy (previously known as Hereditary Inclusion Body Myopathy 2

(HIBM), Distal Myopathy with Rimmed Vacuole (DMRV), Nonaka Distal Myopathy, and Quadriceps Sparing Myopathy (QSM)) is a rare, adult-onset muscle disease characterized by muscle weakness and wasting. It has an autosomal recessive inheritance pattern, and patients have homozygous or compound heterozygous mutations in the affected gene. GNE myopathy is globally estimated to affect 1 in 1,000,000 people76.

Due to the presence of founder mutations, there is an overrepresentation of this disease in

Japanese and Iranian Jewish populations, although the disease can manifest in other ethnic groups as well77-79. 14 Patients typically present after age 30 with foot drop, indicating distal weakness of the tibialis anterior (TA). The disease progresses slowly towards proximal muscles over the next several decades, with patients requiring use of a wheelchair 20 to 30 years after diagnosis. However, in the most typical presentation, the quadriceps muscles are notably spared, even as other muscle groups show marked deterioration76. This quadriceps sparing is evident in MRI scans of patients as well; MRI of patients with advanced disease reveals fatty replacement of muscle, but notable sparing of the quadriceps muscles. Patients typically have a modest increase in serum CK and myopathic motor unit potentials as shown by electromyogram, but have normal nerve conduction velocity. In addition, heart function, respiratory function, speech, and cognition are not known to be affected by the disease80. Characteristic histopathology of affected muscles in GNE myopathy shows both amyloid-beta(Aβ)-positive inclusions and rimmed vacuoles characteristic of autophagy within the myofiber81.

This disease was first described in a Japanese population by Nonaka et al. in

198182. It was next described in an Iranian Jewish population by Argov et al. in 198483, and in additional ethnic groups in 199684. By 2001, the causative gene of the disease was discovered; GNE myopathy was found to be caused by mutations in glucosamine (UDP-

N-acetyl)-2-epimerase/N-acetylmannosamine kinase (GNE), a bifunctional enzyme which catalyzes two steps in SA biosynthesis, including the committed step85-87.

Eight isoforms of GNE are created by . GNE1, the predominant isoform in humans, is found in many organs including skeletal muscle and liver. GNE2 has an extended N-terminal region, and has strong placental expression, as

15 well as low levels of expression in organs such as liver and kidney. The third isoform,

GNE3, has a truncated N-terminal region, and is found at low levels in kidney, liver, placenta and colon88. Five additional, tissue-specific isoforms are predicted based on mRNA expression, but are poorly understood89. X-ray crystallography of GNE1 reveals formation of a homodimer due to interactions of the kinase domain, and suggests formation of tetramers or hexamers in vivo90-92.

In the SA biosynthetic pathway, glucosamine 6-phosphate is converted to

Neu5Ac and shuttled into the nucleus, where it is then converted to a sialyltransferase substrate by addition of cytidine monophosphate (CMP). Upon activation, CMP-Neu5Ac is shuttled back into the cytoplasm, where it can be used by sialyltrasnferases in the Golgi to modify SA-containing glycoproteins and glycolipids. GNE is a bifunctional enzyme that catalyzes two key steps in SA biosynthesis: the committed step of conversion of

Uridine diphosphate N-acetylglucosamine (UDP-GlcNAc) to N-Acetylmannosamine

(ManNAc) through the (UDP-N-acetyl)-2-epimerase domain, and conversion of

ManNAc to ManNAc-6-phosphate (ManNAc 6-p) through the ManNAc kinase domain.

SA biosynthesis is tightly regulated through feedback inhibition in this pathway; CMP-

Neu5Ac, the final product of the SA biosynthetic pathway, can bind and inhibit the epimerase domain of GNE92. Interestingly, mutations in this CMP-Neu5Ac binding site can abrogate this feedback loop and lead to the rare disease sialuria, in which excessive

SA production causes jaundice, hepatosplenomegaly, and hypotonia93, 94.

GNE protein expression and cytoplasmic localization remain unchanged in GNE myopathy95. Instead, mutations found in the disease generally reduce the enzymatic

16 efficiency of either domain of GNE. Enzyme assays of recombinant GNE show that the

GNE D176V epimerase mutation reduces activity by 80%, while the M712T and V572L mutations, which are found in the kinase domain, reduces efficiency by 30% and 90%, respectively96, 97.

Despite the well-documented hypomorphic nature of the mutations in GNE myopathy, actual hyposialylation in the disease has been a matter of much debate. While some groups have reported an overall decrease in muscle sialylation97-99, one group reported no such change when studying sialylation of B lymphoblastoid cells derived from GNE myopathy patients96. Another group saw an overall decrease in SA that was more pronounced when measuring only free SA that was not bound to glycoproteins or glycolipids100. Some found that presence or absence of hyposialylation was dependent on the specific patient mutation101, 102. Finally, it has been suggested that hyposialylation in

GNE myopathy is restricted to either a single protein, like αDG103 or NCAM104, or a class of proteins, like O-linked proteins105. Importantly, subsequent studies showed that αDG hyposialylation, if present, does not seem to affect binding to the extracellular matrix, calling into question its importance in GNE myopathy pathophysiology106. Overall, there seems to be evidence of hyposialylation in the muscles of GNE myopathy patient, although the extent and exact nature of the hyposialylation is under debate.

1.5.2 Mouse models of GNE myopathy

To better understand the pathophysiology of GNE myopathy and develop treatment strategies, it became important to develop a mouse model of the disease.

However, this venture is complicated by the embryonic lethality of a mouse Gne 17 knockout107. To circumvent this, several groups have employed knockout/transgenic or knock-in approaches to model GNE myopathy, with varying success in recapitulating the patient phenotype.

In 2007, Malicdan et al. created a mouse model of GNE myopathy that mimicked much of the human disease pathology. To do this, they knocked out both copies of mouse

Gne; by itself, this deletion is embryonic lethal, but pups were viable with the addition of a human GNE transgene containing the human patient-specific D176V mutation. The

Gne(−/−)hGNED176V-Tg mice show many similarities to human GNE myopathy patients.

First, as expected, these mice display hyposialylation in muscle, liver, and other organs due to the hypomorphic transgenic GNE expression. Interestingly, the SA reduction in liver (~85% reduction) and serum (~95% reduction) was much more striking than the SA deficit in muscle (~30% reduction), the affected tissue in GNE myopathy. Additionally, this model was reported to show gross motor deficits, decreased body weight, and muscle atrophy beginning at 30 weeks. Measures of gross motor function in these mice showed a decline over time, with deficits of approximately 30% reported in mice over 40 weeks old. Finally, histopathology with amyloid-positive inclusions was seen beginning at 32 weeks, with atrophic fibers evident at 40 weeks and autophagic rimmed vacuoles appearing at 42 weeks108.

A second model was made to mimic the GNE M712T mutation, a GNE myopathy founder mutation in the Iranian Jewish population. This mouse model had a

GneM712T/M712T knock-in, allowing homozygous expression of mutant mouse Gne under the control of its endogenous promoter. These model mice initially died 72 hours after

18 birth from glomerular proteinuria, exhibiting no muscle phenotype109. Supplementation with the SA precursor ManNAc allowed survival of the mouse model until weaning age, but this increased lifespan was again not accompanied by a muscle phenotype. This model did, however, show significant deficits in free, but not bound, SA in the muscle as compared to levels in wild type (WT) mice. This free SA comprised approximately 7% of the total SA in muscle, with the remaining 93% representing SA bound to glycoproteins or glycolipids. In addition, free SA in serum was lower than in WT mice100.

A third model of the disease was created to mimic the GNE myopathy V572L founder mutation, developing a knock-in mouse with this mutation. Similarly to the

GneM712T/M712T knock-in model, these GneV572L / V572L knock-in mice displayed severe albuminuria leading to renal failure with no muscle phenotype. However, these GneV572L /

V572L mice had a longer lifespan than the previously created GneM712T/M712T model, with a median survival time of 500 days without intervention. Administration of Neu5Ac to this model in the drinking water improved renal histopathology, but treatment effects on lifespan were not assessed82.

1.5.3 Possible mechanisms of muscle damage in GNE myopathy

GNE myopathy patient muscles show significant atrophy110, and necrosis is not a significant histopathological finding. Rather, these biopsies show signs of apoptosis.

Specifically, TUNEL-positive myonuclei have been observed111. In addition, cultured fibers from GNE myopathy muscle biopsies show a significant propensity for activation of caspases 3 and 9 as compared to normal biopsies in response to the apoptotic inducer

19 staurosporine112. Two mechanisms have been proposed to lead to this muscle degeneration: amyloid aggregation and oxidative stress.

Muscles affected by GNE myopathy often have amyloid-beta(Aβ)-positive inclusions113.AβisacleavageproductofAPP (amyloid precursor protein, which can be processedsequentiallybyβ- andγ-secretase to form the cytotoxic, aggregation-proneAβ fragment114. In 2016, Bosch-Morató et al. showed that, in fact, GNE myopathy fibers in vitro have increased clathrin-dependent endocytosis of the cytotoxic Aβ, and that hyposialylation is both necessary and sufficient for this response115. To support this hypothesis, GNE myopathy muscles show signs of cell stress related to protein aggregation, including colocalization of molecular chaperones with β-positive inclusions and upregulation of αB-crystallin, a part of the unfolded protein response that protects against the inappropriate accumulation of proteins116, 117.

Both the Gne(−/−)hGNED176V-Tg mouse model and patient myotubes have been shown to have increased levels of intracellular reactive oxygen species (ROS), which can be reduced by treatment with SA. GNE myopathy patient muscles also displayed increased S-nitrosylation, a modification that occurs in conditions of excess intracellular

ROS. Additionally, treatment of Gne(−/−)hGNED176V-Tg mice with a pharmacological antioxidant ameliorated muscle pathology, indicating a link between ROS and pathology in the disease118. In line with this idea, transcriptome analysis of GNE myopathy patient muscle and in vitro studies of patient mutations indicated deficits in mitochondrial function115, 119.

20 While the exact pathophysiology of GNE myopathy is not yet fully elucidated,

Aβ-mediated and ROS-mediated stress represent non-mutually exclusive potential mechanisms for muscle degradation in the disease.

1.5.4 Treatments strategies for GNE myopathy

Although the causative mutation of GNE myopathy was identified more than 15 years ago, the pathophysiology of the disease remains largely enigmatic. As such, treatment strategies being pursued rely almost entirely on the knowledge of the mutation; clinical trials are underway using oral preparations of SAs and their metabolic precursors.

The first treatment tested in clinical trials for GNE myopathy was intravenous immunoglobulin (IVIG) in 2005, as IgG contains some SA (~8μmolSA/g IgG). This treatment provided a total of 7.9mg Neu5Ac/kg over the course of 1 month76. While patients in this study experienced some increase in muscle strength during the month that this trial took place, these gains were lost shortly after administration ceased. In addition, no increase in sialylation of αDG or NCAM was detected as a result of this treatment120.

Some gene therapy approaches have also been preliminarily tested for GNE myopathy. In 2011, Nemunaitis et al. published the results of a phase I trial of GNE gene therapy in a single patient. The patient received a daily dose in GNE delivered via lipoplex for seven days. The patient showed GNE gene expression, stabilization of muscle function, and modest improvement in muscle sialylation after administration of thegenetherapy.However,thedosageofasinglepatientcombinedwithpatient’ssevere pathology prior to treatment administration precluded a more thorough analysis of treatment efficacy121. In addition, proof-of-concept studies have also been conducted for 21 the delivery of trans-splicing therapy via AAV to replace the mutated exon in GNE with the wild type exon, but these therapies have not yet been tested in vivo122. Finally, systemic GNE gene replacement via AAV has been tested in wild-type mice, but has not yet been shown in a mouse model of GNE myopathy123.

Most therapies in development for GNE myopathy rely on oral supplementation of SA or its precursors. These have included ManNAc, a SA precursor formed after

GNE’sactionintheSA biosynthetic pathway, and extended release SA (SA-ER). After

ManNAc was shown to rescue hyposialylation in vitro107, 124 and to extend the shortened lifespan of the GneM712T/M712T knock-in mouse125, it was tested in the

Gne(−/−)hGNED176V-Tg mouse model at 20, 200, or 2000 mg/kg/day in drinking water.

All doses tested restored SA levels in skeletal muscle, and precluded formation of histopathology, atrophy, and muscle weakness with continuous administration over approximately one year126. Phase 1 clinical trials of ManNAc showed that single doses of

3, 6, or 10g of ManNAc were well tolerated and increased serum Neu5Ac127. Phase 2 trials are now underway, with patients receiving 12g/day of ManNac over 30 months.

SA-ER has also been tested, with phase 2 clinical trials showing increases of serum SA and a preservation of patient upper limb strength with a dose of 6g/day128. Unfortunately, phase 3 clinical trials failed when no functional primary or secondary endpoints, including muscle strength and mobility, were met following SA-ER administration of

6g/day for 48 weeks129.

22 Although some of the therapies being developed for GNE myopathy show promise, critical evaluation of current therapies and development of new therapeutic strategies will be important for the effective treatment of this disease.

23 Chapter 2. Serum studies in the muscular dystrophies

2.1 Abstract

Duchenne muscular dystrophy (DMD) is a severe, progressive, neuromuscular disorder of childhood. DMD typically arises from frameshift mutations in the dystrophin

(DMD) gene that lead to loss of the dystrophin protein, which normally anchors members of the dystrophin-associated glycoprotein (DAG) complex in the muscle membrane. The lack of dystrophin in DMD leads to a substantial decrease in muscle DAG complex proteins, while in-frame dystrophin deletions, like those found in Becker muscular dystrophy (BMD), lead to a more moderate decrease. The goal of several therapies is to restore dystrophin, and thus the DAG complex, in DMD patients; however, it is difficult to measure the molecular outcomes of these therapies at the whole-patient level. Instead, analysis of molecular changes is often based on a single muscle biopsy. Here, we have developed an ELISA-based assay to measure the relative serum expression of a constitutively cleaved and secreted component of the DAG complex, N-terminal alpha dystroglycan(αDG-N).WefoundthatαDG-N levels were significantly reduced in DMD patient serum (54.5±2.6µg/ml, n=9) relative to both normal patient serum

(182.2±15.8µg/ml, n=38, p<0.01) and BMD patient serum

(163.4±12.7µg/ml, n=11, p<0.05).Asexpected,αDG-N levels did not differ in the serum of patients with a myopathy with no apparent DAG complex perturbation, inclusion body myositis (171.6±8.8µg/ml, n=8, p=0.97),relativetonormalpatientserum.αDG-N levels did not change with age or ambulation status. Inaddition,αDG-N serum levels were significantly reduced in utrophin-deficient mdx mice (6.9±0.5µg/ml, n=4) as compared to

24 either mdx mice (10.1±0.4µg/ml, n=17, p< 0.01) or wild type mice (11.0±0.6µg/ml, n=11, p<0.001). These data identify a potential serum biomarker for DMD that is actually a component of the DAG complex, the affected complex in this disease. Such a measure could be further developed as a molecular readout of global dystrophin levels and/or

DAG complex stability to reflect the efficacy of certain DMD therapies. We have additionally addressed questions related to two categories of serum antibodies that may modulate treatment strategies and pathology in muscular dystrophy. First, the knowledge of the frequency of pre-existing antibody titers to rAAV capsid is important to the development of gene therapies for muscular dystrophies, as these titers can preclude effective treatment. To address this, we have assessed the frequency of serum antibodies to rAAV in normal populations as well as DMD, BMD, IBM, and GNE myopathy patient populations. Finally, we have assessed DMD patient antibody titer to a specific glycan,

N-glycolyl neuraminic acid (Neu5Gc), that may act as a xeno-autoantigen in the disease and exacerbate pathology.

2.2 Introduction

Duchenne muscular dystrophy (DMD) is a degenerative childhood disease characterized by progressive muscle damage and early death130. DMD is caused by mutations, typically frameshift mutations, in the dystrophin gene (DMD). This gene encodes the dystrophin protein, an essential component of the dystrophin-associated glycoprotein (DAG) complex which connects the muscle cell to its surrounding matrix131.

Without dystrophin, levels of the other DAG complex proteins are also decreased, making the muscle membrane weak and damage-prone132.

25 Many therapies seeking to replace dystrophin and stabilize the DAG complex in

DMD are in development, but clinical trials suffer from a lack of reliable molecular outcome measures for dystrophin protein recovery32. In fact, currently, molecular effects of these therapies are often determined by dystrophin quantification in a single muscle biopsy33. In addition to being an invasive procedure for these patients, muscle biopsies provide very limited information regarding therapeutic effects, as dystrophin restoration from these therapies varies greatly, even within the same muscle34. These difficulties in

DMD therapeutic assessment are exemplified by the case of Eteplirsen, a drug that seeks to restore the reading frame, and thus expression, of dystrophin through exon skipping.

Thistherapy’spathtoFDAapprovalwaslengthyandcontroversialdueinparttoalack of confidence in outcome measures; specifically, quantification of dystrophin via muscle biopsy and correlation of these levels to functional outcome measures was highly problematic35. Therefore, a robust DMD biomarker that would accurately reflect global dystrophin presence would be key for assessing clinical trials for drugs such as Eteplirsen and developing an effective treatment for this devastating disease.

Itwasfoundin2004thatalphadystroglycan(αDG),amemberoftheDAG complex, is posttranslationally processed via furin cleavage to generate an N-terminal fragment(αDG-N)52, 133. This fragment is constitutively cleaved and secreted into serum in both normal and dystrophic conditions51; so, its presence in serum could reflect DAG complex stability in the muscle membrane. Here we have developed a serum ELISA to assess the relative expression of serum αDG-N in patients with DMD relative to patients with BMD, IBM (a myopathy in which DAG complex expression is not typically

26 altered), orotherwisenormalcontrols.ChangesinαDG-N levels as a result of muscular dystrophy might be exploited with future development to aid in DMD diagnosis or in the assessment of certain DMD therapies.

In addition, many treatments for neuromuscular disorders are being developed utilizing AAV for gene transfer, including DMD, BMD, IBM and GNE myopathy.

Follistatin (rAAV1.CMV.FS344) gene therapy to increase muscle mass has been tried in both BMD and IBM patients134, gene replacement with partial dystrophin cDNAs, such as micro- or mini-dystrophin, has been tried in DMD patients135, and GNE gene replacement has been tested in a mouse model of GNE myopathy123. However, preexisting antibody titers to AAV can preclude appropriate viral transduction, which is necessary for therapeutic delivery66. Given the impediment of human serum rAAV antibodies to systemic gene therapy, we studied the repertoire and propensity of anti- rAAV serotype antibodies in these patient populations.

Finally, we chose to study the DMD patient immune response to Neu5Gc- containing glycans. Humans differ from all other animals in sialic acid (SA) repertoire due to an inactivating mutation in the CMAH gene, which encodes an enzyme that converts the Neu5Ac form of SA to the Neu5Gc form. Thus, while most animals have an approximately 50:50 mixture of Neu5Ac and Neu5Gc forms of SA, humans predominantly express the Neu5Ac form and do not express Neu5Gc136. This unique SA repertoire has shown to be a modifier of disease; specifically, the mdx mouse model is worsened by Cmah deletion, which is due in part to the antibody-mediated immune response to dietary Neu5Gc75. As Neu5Gc is present in most non-human mammals,

27 humans, unless they are vegans, eat large amounts of Neu5Gc in their diet as the result of meat and dairy consumption. Given that DMD patients, like Cmah-/- mdx mice, lack endogenous Neu5Gc and exhibit incorporation of Neu5Gc into regenerating myofibers75, it is possible that Neu5Gc as a xeno-autoantigen may play a role in exacerbating DMD pathology in humans as well. We therefore chose to study DMD serum antibody titers to

Neu5Gc-containing glycans.

2.3 Materials and Methods

2.3.1αDG-N protein production and purification

A cDNA encoding an N-terminal FLAG-tagged α dystroglycan protein, αDG-N

(MSALLILALVGAAVADYKDDDDKLAAANSHWPSEPSEAVRDWENQLEASMHS

VLSDLHEALPTVVGIPDGTAVVGRSFRVTIPTDLIGSSGEVIKVSTAGKEVLPSWL

HWDPQSHTLEGLPLDTDKGVHYISVSAAQLDANGSHIPQTSSVFSIEVYPEDHSEP

QSVRAASPDLGEAAASACAAEEPVTVLTVILDADLTKMTPKQRIDLLHRMQSFS

EVELHNMKLVPVVNNRLFDMSAFMAGPGNAKKVVENGALLSWKLGCSLNQNS

VPDIRGVEAPAREGTMSAQLGYPVVGWHIANKKPPLPKRIR), with the pre- protrypsin signal peptide from the pCMV1-FLAG expression vector, was cloned into the pFLAG-CMV-1 vector using EcoRI and XbaI sites to yield a secreted αDG-N (30-312)14 amino acid sequence with an N-terminal FLAG tag. The plasmid was transfected into

HEK293T cells, and αDG-N was purified from the supernatant via anti-FLAG (M2) affinity chromatography as previously described137.

2.3.2 Protein biotinylation

28 Purified, recombinant αDG-N and 3B4, a monoclonal antibody to αDG-N

(Creative Diagnostics; Shirley, NY) were biotinylated using EZ-Link™Sulfo-NHS-

Biotin (Thermo Scientific; Waltham, MA). The labeled proteins were subsequently desaltedusingZeba™DesaltSpinColumns(ThermoScientific) and protein concentrations measured using a modified Bradford assay, as before75.

2.3.3 Serum αDG-N ELISA assays

Human serum was obtained from subjects identified within the neuromuscular clinicatNationwideChildren’sHospital under an Institutional Review Board approved protocol (IRB13-00190). Clinical classification of DMD versus BMD was made based upontheconceptof“bestclinicaldiagnosis”,usinganexpertclinicaldiagnosisthat combines available information regarding clinical presentation features, family history, and (when available) protein expression. As described elsewhere, this approach takes into account but is not solely based upon mutation class or predicted reading frame138.

Patients who have lost ambulation at younger than age 12 are classified as DMD; those walking at age 15 are classified as BMD, and those who have lost ambulation between ages 12 and 15 are classified as intermediate muscular dystrophy (not included in this study). Mouse studies were conducted under the approved IACUC protocol AR07-00033 at Nationwide Children’sHospital. All mice ranged in age from 3 weeks old to 3 months old. To obtain mouse serum, blood was collection via facial vein and allowed to clot for 1 hour in non-heparinized tubes, then spun at 2,000 g for 10 min at 4°C to collect serum.

Before each experiment, a dilution curve was performed with normal, DMD,

BMD, and IBM human serum samples, or with wild type, mdx or mdx Utrn-/- mouse 29 serum samples, to determine at what dilution factor the serum samples would fall within the standard curve. This dilution factor ranged from 1:5,000 to 1:80,000 for different experiments, with most assays with human serum done at 1:80,000 and most assays with mouse serum done at 1:5,000. After the appropriate dilution factor was empirically determined, all serum samples were identically diluted using that dilution factor in 50mM bicarbonate buffer, pH 9.4, and a 100mL volume incubated overnight on 96-well pre- coated microtiter plate (Thermo-Fisher Scientific; Waltham, MA), in some cases with

0.5,1.0or1.5ngofαDG-N added. A standard curve using differing amounts of purified

αDG-N, ranging from pgs to 10ng of protein, was also immobilized overnight on every plate in 100mL of 50mM bicarbonate buffer, pH 9.4. Wells incubated overnight with sodium bicarbonate buffer alone (with no αDG-N) were used as controls for background signal and subtracted from sample values, and these values typically did not exceed 10% oftotalpositive(αDG-N) signal. Plates were blocked with 1% bovine serum albumin in

Tris-buffered saline with 0.1% Tween 20 (TBST), washed, and incubated with 2A3, a monoclonal mouse anti- αDG-N-specific antibody (WH0001605M1; Sigma; St. Louis,

MO) at a 1:1000 dilution in blocking buffer. Wells were subsequently washed and incubated with horseradish peroxidase (HRP)-conjugated goat anti-mouseFcγsubclass

2A-specific IgG (Jackson ImmunoResearch; West Grove, PA) at a 1:4000 dilution in blocking buffer. Wells were washed and developed using the Substrate Reagent Pack

(R&D Systems; Minneapolis, MN). The reaction was stopped by 2N sulfuric acid after

20 minutes. Plates were measured for absorbance at 450 nm (OD450) on a SpectraMax

M2 plate reader (Molecular Devices, Sunnyvale, CA). After background signal was

30 subtracted for all values, replicates were averaged and compared to the averaged OD450 for normal serum on each plate, yielding a fold-change from normal average for each sample that was independent of signal variability between experiments. In addition, concentrations for all samples were calculated in reference to an immobilized αDG-N standard curve generated for each plate.

For the competition ELISA, 25ng per well of 2A3 was immobilized on 96-well pre-coated microtiter plates overnight in 50 mM sodium bicarbonate buffer, pH 9.4.

Subsequently, plates were blocked with 1% BSA in TBST, washed, and incubated with a standard curve of purified αDG-N that was premixed with biotinylatedαDG-N at an empirically determined saturating concentration (5ng/well). Serum was diluted 1:100 and added in the presence of 5ng/well biotinylatedαDG-N, sometimes with 1.5ng or 2.5ng of non-biotinylatedαDG-N added as a spike-in. Wells were then washed and incubated with streptavidin-HRP (Jackson ImmunoResearch) at a dilution of 1:1000 in blocking buffer, washed, and developed in a manner identical to the serum-immobilized ELISA assay above. Wells coated with 2A3, but where no αDG-N or serum was added, followed by developed as above, were used as controls for background signal and subtracted from sample values. Background signals for the competition ELISA were generally higher than for the immobilized αDG-N ELISA, ranging between 26% and 41% of total positive

(αDG-N) signal.

The sandwich ELISA was done in a manner almost identical to that previously described to measure αDG-N in human uterine fluid139. Here, 50 ng of 2A3 per well was immobilized on 96-well pre-coated microtiter plates overnight in 50 mM sodium

31 bicarbonate buffer, pH 9.4. Subsequently, plates were blocked with 1% BSA in TBST, washed, and incubated with differing amounts of either purified full-length, native, αDG-

N, purified as described above, or a partial αDG-N fragment made as a fusion protein with GST in E. coli consisting of amino acids 31-141 (αDG-GST, H00001605-Q01-25ug;

Novus Biologicals; Littleton, CO). Wells were then washed and incubated with 1ug/ml of a second, biotinylated monoclonal antibody to αDG-N, 3B4 (Creative Diagnostics), washed again and incubated with streptavidin-HRP (Jackson ImmunoResearch) at a dilution of 1:1000 in blocking buffer. After final washes, the assay was developed for

HRP activity as described for the serum-immobilized ELISA assay. Wells coated with

2A3 but where no serum or αDG-N was added were developed and background signal subtracted from sample values. Background signal for the sandwich ELISA were very high,sometimesreaching75%oftotalpositive(αDG-N) signal.

To determine whether 2A3 and 3B4 competed for binding to αDG-N, ELISA plates were coated overnight with 25ng per well of either 2A3 or 3B4 diluted in 50mM sodium bicarbonate buffer, pH 9.4. Wells were blocked in 1% BSA in TBST. Next, recombinant αDG-N was added in differing amounts ranging from 1pg to 10ng. For the

5ng incubation amount, some samples were first mixed with 1ug/ml of 3B4 (for 2A3- coated plates) or 2A3 (for 3B4-coated plates). Plates were washed and incubated with streptavidin-HRP (Jackson ImmunoResearch) at a dilution of 1:1000 in blocking buffer, washed again and developed as described above.

CV values were determined by the ratio of the standard deviation to the mean for replicates on the same plate (for intra-assay CV) or for the same samples on different

32 plates (for inter-assay CV). Recovery precision values were determined by first subtracting the unspiked result from the spiked result to ascertain the actual spike recovery, which was then compared to the expected spike recovery to determine the recovery yield. Upper limit of quantification (ULOQ) and lower limit of quantification

(LLOQ) were determined by the highest and lowest values respectively with a curve backfit of 80-120% and an inter-assay CV of <30%.

2.3.4 Western blots

For serum Westerns, total serum proteins from each sample were diluted (by identical amounts) in SDS denaturing buffer and separated on 4-12% gradient SDS-

PAGE gels and then transferred to nitrocellulose. 1uL of serum was denatured and run per lane. After transfer, blots were blocked in TBST with 5% non-fat dry milk (NFDM), then incubated with primary antibody, either anti-αDG-N (2A3; Sigma; St. Louis, MO) or anti-fetuin (orb27630, Biobyt; Cambridge, UK), washed in TBST, incubated with appropriate horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG2A secondary antibody (115-035-206, Jackson ImmunoResearch; Seattle, WA) or goat anti-rabbit IgG

(111-035-144, Jackson ImmunoResearch), washed again, and developed using an ECL developing kit (Amersham; Piscataway, NJ), much as previously described140. To remove glycans, recombinant αDG-N purified from transfected HEK293 cell lysate or supernatant, or whole human serum samples, were enzymatically deglycosylated using a protein deglycosylation mix (P6039S, New England Biolabs; Ipswich, MA) to remove both N- and O-linked glycans. Deglycosylated or untreated proteins were then compared by Western blot using 2A3 to probe for αDG-N or an anti-fetuin antibody as above. 33 For mouse muscle westerns, 40 µg of muscle lysates were diluted in SDS denaturing buffer, separated on 4-12% gradient SDS-PAGE gels, and then transferred to nitrocellulose. After transfer, blots were blocked in TBST with 5% NFDM, and then incubated with primary antibody, either anti-αDG-N (2A3, Sigma), anti-hDAG (AF6868,

R&D Systems; Minneapolis, MN), anti-αDG (IIH6C4, Upstate Cell Signaling Solutions;

Lake Placid, NY), or anti-actin (MAB1501, Chemicon International; Temecula, CA), washed in TBST, incubated with appropriate incubated with appropriate horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG2A secondary antibody (115-035-206,

Jackson ImmunoResearch; Seattle, WA), rabbit anti-sheep IgG (313-035-045, Jackson

ImmunoResearch), goat anti-mouse IgM secondary antibody (115-035-075, Jackson

ImmunoResearch) or goat anti-rabbit IgG (111-035-144, Jackson ImmunoResearch) secondary antibody, washed, and again developed using an ECL developing kit

(Amersham).

2.3.5 Immunofluorescence staining

WT mouse skeletal muscles were snap frozen in liquid nitrogen-cooled isopentane, mounted on corkboard, and cut in 10µm cross-sections on a cryostat. Slides being stained with the mouse monoclonal antibody were blocked with Mouse-on-Mouse blocking reagent (Vector Laboratories, MKB-2213, Burlingame, CA). All slides were then blocked with phosphobuffered saline (PBS) containing 5% bovine serum albumin

(BSA) for 1 hour each. Sections were incubated with anti-laminin-a2 (Sigma, L0663) and anti-DAG1 antibody (WH0001605M1, mAb clone 2A3 made to DAG1 amino acids

31-141; Sigma; St. Louis, MO) or AP1528 (kindly provided by Dr. Kiichiro Matsumura, 34 made to DAG1 amino acids 141-170) in blocking solution (5% BSA) overnight at 4°C, washed in PBS, and incubated with donkey anti-rat IgG conjugated to Cy3 (Jackson

ImmunoResearch, 712-165-153, West Grove PA) and goat anti-mouse IgG conjugated to

FITC (Jackson ImmunoResearch, 115-095-071, West Grove PA) or goat anti-rabbit IgG conjugated to Cy2 (Jackson ImmunoResearch, 111-225-144, West Grove PA) for 2A3 and AP1528 primaries, respectively. All sections were washed and placed in mounting medium containing DAPI. Imaging was done on a Zeiss Axiophot epifluorescence microscope using AxioVision LE 4.1 imaging.

2.3.6 Statistics for αDG-N

For analysis of human samples using the serum-immobilized ELISA assay, linear regression with post-hoc Tukey’spairwise comparison was used to assess significance, adjusting for age and gender. For comparison of mouse serum samples using the serum- immobilized ELISA, significance was determined by ANOVA with post-hoc Tukey's pairwise comparison. R square values were determined by linear regression or non-linear regression where appropriate. Statistics were analyzed using GraphPad Prism Version

6.03 (GraphPad Software Inc., La Jolla, CA), save human data, which was analyzed by the Biostatistics Core at Nationwide Children’sHospital.

2.3.7 Human serum samples for rAAV antibody titer

De-identified human serum samples were obtained through the Neuromuscular clinicatNationwideChildren’sHospitalfollowinginformedconsentobtainedundera protocol approved by the InstitutionalReviewBoardatNationwideChildren’sHospital.

35 A total of 16 samples were collected from patients with BMD, 22 samples were collected from patients with DMD, and 18 samples were collected from patients with IBM. Young normal serum samples (<18 years of age) were also collected through the Neuromuscular clinic. De-identified adult normal human serum samples (>18 years of age) were purchased from Bioreclamation IVT (Westbury, NY) and de-identified GNE myopathy patient serum samples were obtained from Dr. Yadira Valles (HIBM Research Group,

Chatsworth, CA).

2.3.8 ELISA to identify serum rAAV antibodies

All rAAV viral vectors were obtained from the Viral Vector Core facility at

NationwideChildren’sHospital. rAAV vectors were produced by the triple transfection method in HEK293 cells and highly purified using density centrifugation and anion exchange chromatography. The identity of various serotypes was confirmed using serotype-specific monoclonal antibodies. Plates were coated overnight at 4°C in coating solution with or without rAAV particles (0.2M bicarbonate buffer, pH 9.4 with or without or 2x109 viral particles/well of either rAAVrh74, rAAV8, rAAV1, rAAV2, rAAV6 or rAAV9). Plates were blocked for 2 - 3 hours at 37°C in blocking buffer (5% milk, 1% goat serum in PBS). Initial screening of samples was done by diluting the serum 1:50 in blocking buffer and adding 100µl to each of 4 wells. Each sample was done in duplicate in wells coated with viral particles and with bicarbonate buffer alone to adjust for background. Samples were incubated at 37°C for one hour. Each plate was washed 5 times with wash buffer (PBS with 0.05% Tween-20). HRP-conjugated secondary antibody was then added at a 1:10,000 dilution (Human IgG-Fc fragment antibody, 36 Bethyl Laboratories, Montgomery, TX) in blocking buffer and incubated at room temperature for 30 minutes in the dark. Each plate was washed 5 times again with wash buffer. Substrate reagent (R&D Systems, Minneapolis, MN) was added to each well and allowed to develop in the dark for 15 minutes. The reaction was stopped with 1N sulfuric acid and the optical density of each well was read on a Synergy2 Plate Reader (BioTek,

Winooski, VT) at 450 nm. Any samples that tested positive at 1:50 were retested by

ELISA with a dilution series from 1:100 until a dilution was identified where at least a two-fold increase in signal was no longer obtained. For each sample, the mean optical density (OD) was calculated by subtracting the background value (-viral particle wells) from the sample value (+viral particle wells) and determining the signal/background ratio. The serum sample was called AAV-positive if the signal/background ratio was greater than 2. Background signals remained constant throughout at an OD of 0.1 to 0.2, making for little or no variability in the quotient used to determine 2-fold or greater differences.

2.3.9 Statistics for rAAV antibody titer

Comparisons of significant differences in lowest positive signal dilution between groups was done using a Kruskal-Wallis test followed by a multiple comparisons test, as recommended previously for non-linear serum dilution measures. Measures with p<0.05 were considered significant.

2.3.10 Glycan arrays

37 Neu5Gc- and Neu5Ac-containing glycan arrays were obtained from our collaborator, Ajit Varki, at UC San Diego, and were assayed for human antibody titers much as previously described141.

2.4 Results

2.4.1DevelopmentofanαDG-N ELISA assay

WecomparedthreeapproachestoassayingserumαDG-N expression using an

ELISA assay (Fig. 1). In the first approach, we immobilized serum at high dilutions or immobilizedpurifiedαDG-N directly onto the ELISA plate and then probed amounts of

αDG-N using 2A3, a mouse monoclonal antibody specific to this region of the protein51.

2A3 binding was then indirectly visualized by binding of an anti-mouse IgG2a coupled to horseradish peroxidase (HRP), followed by a standard HRP color enzyme reaction and reading of absorbance at 450nm (OD450) in an ELISA plate reader. In the second and third assays, we tried an indirect competition ELISA assay and a sandwich ELISA assay to measureαDG-N levels in serum that was added to an ELISA plate immobilized with

2A3. FortheindirectELISA,wecombinedaconstantamountofbiotinylatedαDG-N withdifferingamountsofunlabeledαDG-N or serum and assessed loss of signal resulting from increasedcompetitivebindingofunlabeledαDG-N. After washing, streptavidin-HRP was added to develop the signal using standard color development for

HRP enzyme activity. For the sandwich method, we again immobilized 2A3 on the

ELISA plate. PurifiedαDG-N orserumwasthenadded,washed,andαDG-N binding visualizedbyadditionofbiotinylated3B4,asecondαDG-N antibody, followed by streptavidin-HRP and development as before. This sandwich assay is almost identical to 38 that used in a study recently published assay by Nie and colleagues to measure αDG-N in human uterine fluid139.

In Fig. 1, we show examples of standard curves for each type of assay using purified,recombinantαDG-N protein. WhenαDG-N was immobilized on the ELISA plate in different amounts, we found that 2A3 binding could be correlated with different amountsofimmobilizedαDG-N (Fig. 1A). For this serum-immobilized assay, the upper limit of quantification (ULOQ) was 5ng and the lower limit of quantification (LLOQ) was 0.16ng. Similarly, for the indirect ELISA, we could show a correlation between loss of signal using increasing amounts of non-biotinylatedαDG-N in a range from 2ng to

15ng (Fig. 1B). Surprisingly, we found no correlation in antibody binding from the sandwichELISAassayusingrecombinantαDG-N (Fig. 1C). We did, however, find a correlationwhenapartialαDG-N protein fragment linked to glutathione-S-transferase

(αDG-GST) was used (Fig. 1C). This was the protein previously described by Nie and colleagues in theirαDG-N sandwich ELISA assay139. αDG-GST has only amino acids

31-141oftheexpectedαDG-N sequence, which begins at amino acid 30 after the signal peptide and ends at amino acid 31253. Inaddition,thisshorterαDG-GST protein was made in E. coli and so would not be glycosylated. Bycontrast,therecombinantαDG-N protein we had made in transfected HEK293 cells corresponded to the entire amino acid

30-312 protein sequence expected for the furin-cleavedαDG-N fragment and was glycosylated, as has been previously reported by Matsumura and colleagues51. We found we could generate a standard curve using the ELISA assay with 2A3 (Fig. 1D) or biotinylated 3B4 (Fig. 1E) antibody using recombinant, full-lengthαDG-N, but in both

39 instances, pre-incubation with the other antibody in solution eliminated all such binding

(Fig. 1D and E). ThissuggeststhatthebindingsitefortheseantibodiesonnativeαDG-N are incompatible with use in a sandwich ELISA, as both antibodies cannot both simultaneously recognizenativeαDG-N. AsnootherαDG-N antibodies were available, we did not pursue the sandwich method further.

We next compared spike-insofknownamountsofpurifiedαDG-N protein, adding 0.5, 1.0 or 1.5ng (for serum-immobilized assay) or 1.25 or 2.5ngofαDG-N (for competition assay) from 2 normal human and 2 DMD patient sera samples to determine recoveryprecisionofaddedαDG-N. For the serum-immobilized assay, we measured a

56±4%recoveryyieldofαDG-N from normal human serum and a 55±7% recovery yield ofαDG-NfromDMDserum,andthisyieldwasroughlyequivalentatalladdedαDG-N amounts. In addition, these yields were not significantly different between normal human and DMD serum. For the competition ELISA assay, we measured a cumulative recovery yieldof189±37%forαDG-N from normal human serum and a recovery yield of 88±11% forαDG-N from DMD serum. These yields were in fact significantly different (p=0.03).

In other experiments, the degree of yield changes between DMD and normal was sometimes even more pronounced if serum amounts added reached the lower OD450 signal levels on the standard curve (not shown). Because the yield of spiked signal was beyond the expected signal for normal serum for the competition ELISA, and because

DMD and normal sera also showed significantly different responses, we did not pursue this assay further. WethereforeproceededtoinvestigatedifferentialαDG-N levels using the serum immobilization assay.

40 2.4.2αDG-N is decreased in the serum of patients with DMD

We performed serum-immobilizedELISAstomeasureαDG-N levels in human serum from patients with Duchenne muscular dystrophy (DMD), Becker muscular dystrophy (BMD), otherwise normal patient controls, and a myopathy unrelated to the

DAG complex, Inclusion Body Myositis (IBM) (Fig 2). A summary of relevant patient information is included in Table 1. This includes the fact that 8 of 9 DMD patients, 2 of

11 BMD patients, and 0 of 8 IBM patients had been treated with corticosteroids in the 3 months prior to the taking of the serum sample. There were no significant changes in comparing any disease group with regard to corticosteroid use (not shown). Serum dilutions sometimes had to be altered to maintain all signals within the linear range of the standard curve, but were generally done at or near a 1:80,000 per plate, which was the predominant dilution used for human samples. Assuming a serum concentration of

80mg/mL protein, a 1:80,000 serum dilution would result in 100ng of serum protein being immobilized per well of a 96-well ELISA plate. We ascribe the need to change serum dilution for some experiments to the quality of immobilization on various lots of

ELISA plates and to the relatively tight range of the αDG-N standard curve. In each experiment, however, all normal and patient serum samples were identically diluted. We analyzed the data in two different ways. First, we analyzed the absolute relative change in

OD450 signal between serum samples (Fig. 2A). To do this, we reported the fold-changes in each signal relative to the average signal for all otherwise normal samples in each experiment, set to a value of 1. 3-7 replicate experiments were done per sample, each with duplicate measures per assay. We found that patients with DMD showed a 41 significant reduction (27±3% decrease from normal, n=9)inOD450signalforαDG-N compared to otherwise normal patients (p ≤0.001,n=38). BMD patients showed an intermediatelevelofreductionαDG-N in the serum (14±2% decrease from normal, n=11), statistically differing from both otherwise normal patients (p ≤0.01)andDMD patients (p ≤0.05). In contrast, IBM patients, who show adult-onset progressive muscle wasting without major DAG complex alterations142, showed an insignificant change from normal patients (4±3% decrease from normal, p=0.99, n=8). Because DMD patients were younger than BMD patients (average age 10 versus 19, respectively), and because IBM patients were significantly older than both of these groups (average age 57), we performed linear regression adjusting for age and gender when comparing all disease groups to determine significance. In analyzing the variability of individual measures, we found that the serum-immobilized assay showed relatively robust reproducibility.

Considering all samples, the overall intra-assay coefficient of variation (CV) was 3.8% and the inter-assay CV was 16.3%.

WenextcomparedmeasuresofαDG-N concentrations derived from the standard curves done with each serum-immobilized assay (Fig. 2B). Because we had to occasionally use a different standard dilution for all of the samples to fit the OD450 signals within the range of the standard curve, we expected these values, which must take into account differing inter-assay dilution factors, to be more varied. While a few measuresweremorevariableduetochangeddilutionfactor,theserumαDG-N concentration was still significantly reduced in DMD serum compared to normal serum aswellastoBMDserum(Fig.2B),withtheoveralllevelofreductionofαDG-N

42 expression ranging from 65-70% for all such comparisons. ForcalculationsofαDG-N serum concentrations, the intra-assay CV remained low (4.5%), but because of the need to use different dilution factors for certain experiments to maintain linearity of all samples on the standard curve, the inter-assay CV was poor (73%).

WealsovalidatedchangedexpressionofαDG-N in DMD via Western blot (Fig.

3). WhilemorequalitativethananELISA,levelsofαDG-N were decreased in serum from patients with DMD as compared to age-matched, otherwise normal male patients

(Fig. 3A). Immunoblots for fetuin, an abundant serum protein whose expression should not be altered in DMD, were used as a control for serum protein loading and transfer. The

2A3 antibody recognized several protein bands in human serum migrating between the

39kDa and 51kDa molecular weight markers. Because loading of even 1µL of serum, as was done here, leads to warping of protein bands due to the large amounts of protein loaded, we could not discern the exact molecular weights of these species; however, these species did migrate in the same range as was found for 2A3 immunoblots of purified recombinantαDG-N purified from secreted HEK293 cells. We found that subjecting purifiedrecombinantαDG-N isolated from secreted HEK293 cells to enzymatic deglycosylation (of both N- and O-linked proteins) reduced the molecular weight of the

2A3-blotted protein from about 45kDa to 39kDa, such that it now equaled the molecular weight recognized by 2A3 in HEK293 cell lysate (Fig. 3B). A similar effect was found when normal and DMD serum were deglycosylated, and serum fetuin also showed reduced molecular weight after this treatment (Fig 3B). 37-45kDa is roughly the molecularweightofαDG-NproteinspeciespreviouslypublishedforαDG-N protein in

43 human serum and/or cerebrospinal fluid, with about a 37kDa protein identified after protein deglycosylation, much as was seen here51, 54, 55.

2.4.3SerumlevelsofαDG-N do not differ significantly based on age, gender, or ambulation status

DespitethefactthatourlinearregressionanalysisofhumanserumαDG-N

ELISAs took age and gender into account, we also plotted relativeserumlevelsofαDG-

N versus age for all normal patient samples and found that expression within this group did not significantly change with increasing age (r2=0.009) (Fig. 4A). αDG-N serum levels also did not change with increasing age in BMD (r2=0.147) or DMD (r2<0.001) samples (Fig. 4B and 4C, respectively). Additionally,αDG-N serum expression was not changed when otherwise normal patients were grouped by gender (females were 105±3% of males, p=0.2) (Fig. 4D). We next compared αDG-N serum levels of BMD and DMD patients based on ambulation status at the time the serum sample was taken (Fig. 5A and

5B, respectively), and again found no significant difference between ambulatory and non- ambulatory patients for BMD patients (non-ambulatory was 102±4% of ambulatory, p=0.75) and DMD patients (non-ambulatory was 107±7% of ambulatory, p=0.48). These datasuggestthatserumαDG-N expression is a marker of disease that is independent of patient age, gender, and ambulation status.

2.4.4 mdx Utrn-/- mice,butnotmdxmice,showdecreasedexpressionofserumαDG-N compared to normal mice

44 We next sought to replicate these findings in the mdx mouse model of DMD (Fig.

6). The mdx mouse, like DMD patients, lacks dystrophin protein expression in muscle cells143. Surprisingly,wefoundnosignificantdifferenceinαDG-N ELISA signal between wild type (WT) and mdx serum (p=0.66, n=10 for mdx and 11 for WT, Fig. 6A).

Because mdx mice show far less overall muscle pathology than DMD patients and also have upregulation of utrophin protein (made by the Utrn gene), a dystrophin paralog known to compensate for the loss of dystrophin in skeletal muscle by binding and stabilizing the DAG complex144, 145, we also assayed serum from mdx Utrn-/- mice. Note thatwhilethemouseandhumanαDG-N proteins are 92% identical in amino acid sequence, some interspecies differences may exist in comparing the human and mouse measures,asweusedαDG-N from the same species, rabbit14, to generate both sets of standard curves. TherabbitαDG-Nproteinsequenceis93%identicaltothehumanαDG-

Nsequenceand91%identicaltomouseαDG-N sequence. mdx Utrn-/- mice generally have far more severe disease pathology than do mdx animals due to the loss of both utrophin and dystrophin protein expression17, 146. In contrast to mdx mice, we found a robustdecreaseinserumαDG-N signal in mdx Utrn-/- mice as compared to WT or mdx mice (mdx Utrn-/- signal was reduced by 49±4% compared to WT, p<0.0001, n=4, Fig.

6A). The cumulative intra-assay CV for these serum measures was 4.0%, while the overall inter-assay CV was 13.1%. These reduced OD450 signals in mdx Utrn-/- mouse serum correlated with a reduced calculated serum concentration as well, with mdx Utrn-/-

αDG-N concentration reduced by 37±3% compared to WT, p<0.001) and 31±3% compared to mdx (p<0.01, Fig. 6B). ForαDG-N concentration measures, the inter-assay

45 CV was 4.8% and inter-assay CV was 34.4%. These experiments suggest that utrophin expressionmayimpactαDG-N expression in dystrophin-deficient mouse serum. By contrast,whencomparedtoWTmice,wesawnosuchdecreaseinαDG-N levels in the serum of mdx Cmah-/- mice, which also shows worsened pathology as compared to mdx mice (data not shown).

We next wanted to do a preliminary comparison of mdx Utrn-/- expression of serumαDG-N with that of the mct mdx Utrn-/-mouse model, which transgenically overexpresses the glycosyltransferase GALGT2 that has shown efficacy in ameliorating muscular dystrophy symptoms23.WefoundthatrelativeserumαDG-N signal in mct mdx

Utrn-/-were significantly higher than in mdx Utrn-/- mice (p <0.01, n=4 for mdx Utrn-/- and n=2 for mct mdx Utrn-/-, Fig. 7A). In fact, these levels in mct mdx Utrn-/-mice were indistinguishable from mdx levels (p=0.39, n=3 for mdx). These samples showed a reduction in calculated serum concentration as well (p <0.01 for mdx Utrn-/- and mct mdx

Utrn-/-, p=0.67 for mdx and mct mdx Utrn-/-, Fig. 7B). It should be noted, however, that these calculated serum concentrations differ from previously calculated levels, likely due to the poor inter-assay CV discussed earlier.

To check for the presence of αDG-N in skeletal muscle, mouse muscle gastrocnemius muscle was immunostained with the 2A3 monoclonal antibody, the same antibody used to recognize αDG-N in serum ELISAs and in Western blots, as well as the polyclonal, αDG-N-specific AP1528 antibody (Fig. 8). An antibody specific for laminin a2 was used in both cases as a counterstain to visualize the sarcolemmal membrane of skeletal myofibers. Although some αDG-N staining was observed in intramuscular

46 peripheral nerves, little or no αDG-N staining was evident in skeletal myofibers, suggesting that the majority of αDG-N is either secreted into the serum from muscle cells or is degraded and therefore not present in adult skeletal muscle, much as has been previously described10, 13, 14. Control staining with secondary antibody alone showed no staining of nerves or muscle fibers (data not shown). Notably, αDG-N staining with intramuscular peripheral nerves was also present in P3ProCreDag1lox/lox gastrocnemius muscle (data not shown), which lacks dystroglycan expression in skeletal muscle cells147.

This further suggests that αDG-N expression within peripheral nerves did not arise from secreted skeletal muscle protein, as skeletal muscle Dag1 expression is absent in these mice.

In addition, we also performed immunoblotting of WT, mdx, and mdx Utrn-/- mice gastrocnemius lysates using the 2A3 (αDG-N) antibody, the hDAG (α/β dystroglycan) antibody, the IIH6 (glycosylated α dystroglycan) antibody, and an antibody to actin (Fig. 9). We included purified, recombinant αDG-N as a positive control for the

2A3 antibody. We found that while all genotypes expressed dystroglycan to varying extents, none expressed detectable quantities of αDG-N, in agreement with our immunostaining data.

2.4.5GRMDdogsshowatrendtowardsdecreaseinexpressionofserumαDG-N compared to GR dogs

Finally,wecomparedserumαDG-N expression in golden retriever (GR) dogs and the GR muscular dystrophy (GRMD) model (Fig. 10). While there was a trend towards decreased expression in GRMD dogs as compared to GR dogs, this did not reach 47 statistical significance (p=0.098). However, the GRMD model is prone to significant variability in phenotype, and the extent to which this variability may impact levels of serum αDG-N is unclear.

2.4.6 Characterization of human serum rAAV antibodies

We next assessed antibody titers to rAAVrh74, rAAV8, rAAV1, rAAV2, rAAV6 and rAAV9 serotypes in human serum, as host immune response to the viral vector can significantly impede therapeutic delivery. A summary of relevant patient information is included in Table 2 (otherwise normal patients), Table 3 (DMD and BMD patients), and

Table 4 (IBM and GNE myopathy patients).

We first assessed these antibody titers in serum samples from otherwise normal patients. Because some of the diseases we wished to study occur in children, we subdivided normal serum samples into two groups based on age, young normal (<18 years-old) and adult normal (>18 years-old; by chance, there were no 18-year-old patients in this study). Because rAAV8 and rAAVrh74 are most similar with regard to capsid sequence (93% identical), we grouped these comparisons next to one another in each graph. Sera that were not elevated two-fold at a dilution of 1:50 were considered negative and given a value of zero. Sera with positive signals at 1:50 were then assayed at subsequent 1:2 serial dilutions to identify the dilution at which the last positive titer signal could be identified. In each instance, the reciprocal of that dilution factor is presented. In 19 young normal samples, we identified 5 with positive rAAV ELISA signals (Fig. 11A). In each instance, a positive signal at a 1:50 dilution was identified for all 6 rAAV serotypes tested in rAAV antibody-positive subjects. In addition, the 48 reciprocal dilution factor at which a positive signal was measured for rAAV2 was greater than or equal to the titer identified for all other serotypes. Of 21 adult normal samples, 7 samples showed positive antibody titers to all rAAV serotypes tested (Figure 11B). All rAAV serotypes except rAAV9 showed positive titers in 48% or more of these patient samples, with 76% having positive titers to rAAV2. There were 3 instances where no measurable titer to rAAV9 was observed but where measurable titers to all other rAAV serotypes were present. Adult normal samples were the only group where this was the case. The magnitude of dilution required for a positive signal was quite varied between antibody-positive individuals and between serotypes within the same subject. Dilution factors required for a negative rAAV2 signal were far and away the most variable, and positive titers for this serotype were the most common. Patient 7, for example, had a minimal positive signal for rAAV2 at a serum dilution of 1:205,000, while other patients were positive for rAAV2 only at a 1:50 dilution (Fig. 11A and 11B).

We next assessed the frequency of rAAV titers in patients with DMD (Fig. 12A) or BMD (Fig. 12B). In 22 DMD samples, we identified only 4 patients with positive titers to all rAAV serotypes tested, and 3 additional samples with positive titer only to rAAV2.

Here, the average dilution factor for a positive titer trended lower than what was seen in young normal subjects. Surprisingly, in 16 BMD samples, we identified 0 patients with antibody titers to rAAVrh74, rAAV8, rAAV1, rAAV6 or rAAV9, and only 1 patient with a measurable titer to rAAV2. By contrast to BMD and DMD samples, both of which had concentrated samples from young patients, we identified a far greater propensity to rAAV titers in IBM (Fig. 12C) and GNE patients (Fig. 12D), which have average ages of 60 and

49 32 respectively. In both instances, titer frequencies in these patient groups matched those found in adult normal subjects, with 9 of 18 IBM samples and 2 of 4 GNE samples showing positive titers to all rAAV serotypes. Three additional IBM patients showed a positive titer to rAAV2, with two of these also showing positive titer to rAAV1 (Fig.

12C).

We next compared the cumulative average dilution required for a positive antibody signal (Fig. 13A) and the frequency of positive signals (Fig. 13B) in each patient group. Because the serial antibody dilution assay we used was not linear with respect to signal, we performed a Kruskal-Wallis test followed by a multiple comparisons test to assess significant differences between average dilution factors for the various serotypes148. This analysis included all samples. In adult normal samples, positive titer signals for rAAV2 were significantly increased with respect to titers for rAAVrh74 and rAAV9 (p<0.05). In general, while average titers may have declined slightly in adult normal subjects relative to young normal subjects, adult normal subjects had a higher likelihood of having pre-existing serum antibody titers to rAAV serotypes. In addition, we found that antibody titers for rAAVrh74 were significantly higher for adult normal and IBM than for BMD, titers for rAAV8 and rAAV9 were significantly higher for IBM than for BMD, titers to rAAV1, rAAV2 and rAAV6 were significantly higher for adult normal and IBM than for BMD, and titers to rAAV2 were significantly higher for adult normal than for DMD (p<0.05 for all). None of these comparisons account for age, and so in some instances these are inappropriate comparisons. DMD ages, for example, only really match the young normal controls, while IBM patients in this analysis are older than

50 patients in all other groups. We also observed a general increase in the frequency of subjects with rAAV serum antibodies with age when assessing all groups in a single pool.

For example, young subjects (<18 years-old) had a 26% cumulative incidence of rAAV titers, while this increased in 20-39 year-olds to 58% and to 65% in 40-89 year-olds.

Additionally, there was a bias towards increased rAAV incidence in females vs. males

(61% vs. 36%) when the data was pooled independent of disease groups, but was due to the fact that DMD and BMD are X-linked diseases composed almost entirely of young male subjects, and incidence in younger subjects was lower than incidence in older subjects. Last, antibodies to rAAVrh74 appeared to be in line, both in terms of propensity of patients with positive titers and in terms of amplitude of those titers, with rAAV8, rAAV6 and rAAV9.

Previously, we had defined the serum dilution of 1:800 as being the dilution at which positive antibody signals to rAAVrh74 could significantly diminish muscle transgene transduction levels in rhesus macaques treated with either rAAVrh74.MCK.GALGT2 or rAAVrh74.MCK.µDystrophin67, 149. For these experiments, rAAV vector was delivered by intra-arterial delivery using an isolated limb perfusion method to treat the gastrocnemius muscle. We therefore also assessed the frequency of positive signals to all rAAV serotypes at a 1:800 dilution in addition to assessing the frequency at a 1:50 dilution (Fig. 13B). Young normal subjects showed no change in propensity of rAAV antibodies to any serotype between the 1:50 and 1:800 dilution, however, there were a few patients, particularly in the DMD group, that showed reduced frequencies of rAAV antibodies at 1:800 compared to 1:50. For example, the percentage

51 of DMD patients positive for antibodies to rAAVrh74 at 1:800 was only 5% as compared with the 20% at 1:50. Similarly, the frequency of DMD patients with positive titers to rAAV9 at 1:800 was only 10% compared to 20% at 1:50. By contrast, the frequency of positive antibody signals to rAAV1 and rAAV2 were identical at 1:50 and 1:800 for

DMD patients. Thus, most DMD patients have pre-existing antibody levels to rAAVrh74 that are below the amount that would be expected to block muscle tissue transduction with vascular delivery. This was also the case for BMD patients, who showed no measurable rAAVrh74 serum antibodies.

2.4.7 Assessment of antibody titer to Neu5Gc-containing glycans in DMD patients

Finally, we utilized a glycan array to assess antibody titers to Neu5Gc-containing glycans in otherwise normal and DMD patients. Antibody titer to 23 Neu5Gc-containing glycans trended higher than in DMD patients (n=22) as compared to otherwise normal patients (n=16), although this did not reach statistical significance in any case (Fig. 14).

As a whole, antibody titers to Neu5Gc-containing glycans in DMD patients trended higher than those in otherwise normal patients (although not statistically significant, p=0.18); antibodies to Neu5Ac-containing glycans did not exhibit this trend (p=0.95, Fig.

15A). Likewise, there was no difference between antibody titers to Neu5Gc-containing glycans in otherwise normal and DMD patients when split out by linkage to terminal

Neu5Gc (p=0.49 for 2,3 terminal linkage, p=.027 for 2,6 terminal linkage, and p=0.93 for

2,8 terminal linkage, Fig. 15B). We then assessed the relationship of antibody titers to

NeuAc- or Neu5Gc-containing glycans by age in otherwise normal and DMD patients

(Fig. 15C, 15D). We found that while serum samples from otherwise normal patients did 52 not vary by age in antibody titer to Neu5Ac-containing glycans (r2=0.046, p=0.42) or

Neu5Gc-containing glycans (r2=0.022, p=0.59), serum samples from DMD patients showed a decrease with age in antibody titer to Neu5Ac-containing glycans (r2=0.203, p=0.04) and Neu5Gc-containing glycans (r2=0.022, p=0.04).

2.5 Discussion

Here, we have developed an ELISA assay that utilizes immobilized diluted serum to measure levels of a normally cleaved N-terminal fragment of α dystroglycan, αDG-N.

In doing so, we have found that aDG-N expression in serum from patients with DMD is significantly reduced relative to serum from otherwise normal patients and to serum from

BMD patients. These findings were independent of age, suggesting that αDG-N reduction in DMD is more of a fixed marker of disease than a reflection of some ongoing disease process. There are a number of mechanisms that could give rise to the changed expression of the aDG-N protein fragment in DMD serum (Fig. 16). Reduced serum aDG-N levels may reflect reduced intracellular dystroglycan expression or stability in

DMD muscle, reduced aDG-N stability once cleaved in DMD muscle or serum, reduced aDG-N secretion from DMD muscle, or increased aDG-N scavenging in DMD serum.

As dystroglycan cleavage to liberate aDG-N in muscle appears to be complete in both normal and DMD muscle11, 13, 150, 151, it seems unlikely that reduced furin activity would account for changed aDG-N expression. Because aDG-N is immobilized for the ELISA measure done here, reduced aDG-N signals in DMD serum may also reflect increased masking of aDG-N due to increased binding of DMD serum proteins to αDG-N antibody-reactive epitopes. 53 While it is certainly possible that αDG-N expression in the serum reflects reduced dystrophin expression or reduced dystroglycan protein expression, both of which occur in

DMD2, 3, 151, the fact that the same findings could not be replicated in mdx mice, but were replicated in utrophin-deficient mdx mice, makes such a conclusion problematic. While certainly some differences in serum αDG-N expression could reflect human-mouse muscle differences, there is no doubt that most mdx muscles fail to express dystrophin2,

143, making it unlikely that serum αDG-N directly reflects dystrophin expression in skeletal muscle. While this reduction was observed in mdx Utrn-/- mice, it is also unclear how this might be explained by human-mouse differences. One explanation might be that utrophin is better at stabilizing dystroglycan expression in mouse muscle than it is in human muscle, but there is little evidence to support this possibility.

Preliminary studies also showed that overexpression of GALGT2, a therapeutic gene for muscular dystrophy, may restore levels of serum αDG-N in mdx Utrn-/- mice.

GALGT2 overexpression has been shown to ameliorate muscular dystrophy in mouse models through the ectopic expression of normally synaptic proteins such as utrophin throughout the muscle membrane, stabilizing the membrane23. While these results are far from conclusive, they may indicate that stabilization of the DAG complex by therapeutic intervention may be sufficient to restore levels of serum αDG-N.

If altered serum αDG-N expression were to indeed reflect altered dystrophin expression in DMD and BMD patients, then it could be exploited as a global marker of dystrophin protein recovery in therapies aimed at reintroducing a partially functional dystrophin protein to DMD patients. Such therapies include antisense- and -

54 based exon skipping strategies, such as drisapersen152, 153 and Eteplirsen154, 155, and also missense read-through therapies such as ataluren156. All of these types of therapies are plagued by the difficulty that analysis of dystrophin expression in single muscle biopsy does not typically reflect changed dystrophin protein expression in muscles throughout the entire body plan, which is the biomarker needed to best reflect overall drug efficacy33.

While additional work will be required to understand if such a finding can be exploited, it is possible that serum αDG-N, as a marker of muscle dystroglycan stability or expression, may be reduced in DMD patient serum because dystrophin is absent. The fact that BMD and DMD serum αDG-N signals differed from one another also suggests that this may be possible. Unfortunately, the level of decline in αDG-N serum signal in DMD vs. normal, while highly significant, is only 27% of total OD450 signal. While the calculated concentration difference is greater, this is a less robust measure due to the non-linear nature of the standard curves and issues with serum dilution. The lack of a greater overall change in αDG-N signal is likely is the result of the fact that dystroglycan is present in many tissues, for example skin, where dystrophin is not present and where dystroglycan can be stabilized by other dystrophin-like proteins such as plectin 1157, 158. Thus, there is likely a dystrophin-independent signal emanating from non-muscle tissues that contributes a significant fraction of serum αDG-N expression. Further work will be required to understand these and other issues that may affect changes in serum αDG-N expression, and whether there may be a unique muscle-specific modification of αDG-N that could be utilized to eliminate non-muscle background signal.

55 While the data presented here provide for a proof of concept that αDG-N expression is changed in the serum of DMD patients relative to otherwise normal patients, our results will benefit from the analysis of additional cohorts, and the assay we have used would need to be further optimized in order to exploit this measure for large- scale quantitative studies. Because we utilized an ELISA where serum was diluted and directly immobilized on the ELISA plate, we found that the serum dilution factor sometimes had to be altered between experiments in order for all signals to be below the saturation range of the standard curve. Although the serum samples within each plate were always diluted to the same degree, this increased inter-assay CV. Another recent study found that use of a second αDG-N monoclonal antibody, 3B4, in addition to the

2A3 antibody used here, allowed for development of a sandwich ELISA to measure

αDG-N levels in solution in human uterine fluid139. That assay used a non-native and smaller fragment of αDG-N protein to generate a standard curve, and we show here that use of native length and glycosylated αDG-N does not allow for such a sandwich assay using these antibodies; pre-incubation of the native full-length αDG-N with either 2A3 or

3B4 did not allow for recognition of αDG-N by the other antibody. It may be that the shorter protein fragment used previously oligomerized in such a way that more than one identical epitope was available for binding. If so, use of native αDG-N does not appear to allow for this to occur. Were a panel of monoclonal antibodies to be identified that could be mapped to specific protein domains of αDG-N and shown to recognize non- overlapping protein elements, a sandwich ELISA could be developed, and such an assay might obviate the need for immobilizing serum on the plate. Generation of such

56 antibodies would be very helpful for improving this assay, allowing for a more standardized measurement of αDG-N in the serum.

Additionally, in this study, we have shown that the majority of patients with

Duchenne muscular dystrophy (DMD) and Becker muscular dystrophy (BMD) would be amenable to treatment with rAAV gene therapy vectors, including rAAVrh74, rAAV8, rAAV1, rAAV2, rAAV6 and rAAV9, as only a minority of DMD and BMD patients have pre-existing serum antibodies to these serotypes of rAAV that might inhibit transduction of tissues when they are delivered via the blood. Serum antibody titers were lowest in DMD patients for antibodies to rAAVrh74 and rAAV9 when assayed at a 1:800 dilution, the dilution previously found to block rAAVrh74 muscle tissue transduction during vascular delivery67. The propensity of DMD and BMD patients to have such antibodies is equal to or lower than those found in age-matched normal subjects, in spite of the fact that these disorders are associated with increased tissue inflammation in skeletal muscles. In general, serum antibodies to rAAVrh74 appear to be on a par with antibodies to rAAV6, rAAV8 and rAAV9, which are three of the rAAV serotypes that commonly have the low incidence of serum antibodies in humans66. Such findings are significant as we hope to use rAAVrh74 in future DMD and BMD gene therapy studies, much as has recently been done in LGMD2D patients159.

While these findings in DMD and BMD patients were encouraging, we also found that the preponderance of pre-existing antibodies to all rAAV serotypes tested increased with the age of patients. Thus, IBM and GNE myopathy patients, all of whom are adults at the time of diagnosis, have only a 1in 2 chance of not possessing pre-existing rAAV

57 antibodies to any given serotype. This high incidence of subjects with pre-existing serum rAAV antibodies is similar to our findings in normal adult patients. The likelihood of adult subjects having pre-existing antibodies to rAAV2 was higher still, with 8 of every

10 adult normal subjects and 7 of every 10 IBM subjects showing a positive signal at a

1:50 dilution. Clearly, to effectively reach all IBM or GNE myopathy patients, some method to reduce pre-existing rAAV antibodies may be required, such as immunosuppression or plasmapheresis.

An additional finding of this study was that the majority of patients with positive rAAV antibody titers to a particular serotype were highly likely to display positive serum antibodies to all of the other serotypes tested. This is not an uncommon result in human studies67 and is not surprising as the capsid sequences of rAAV1, rAAV2, rAAV6, rAAV8, rAAV9 and rAAVrh74, the serotypes tested here, all share at least 80% sequence homology57. rAAV4 and rAAV5, by contrast, show less capsid homology to these other serotypes, and studies of serum rAAV4 antibodies in fact show that they are far less common in human subjects160. These results suggest that there would seem to be a significant chance of developing cross-reactive anti-capsid rAAV antibodies to related serotypes as the result of infection with a single rAAV vector, and this is also supported by numerous studies in humans, non-human primates and mice66, 161-164.

Finally, glycan arrays have shown a trend towards increased antibody titer to

Neu5Gc-containing glycoproteins in DMD patients as compared to otherwise normal patients. These antibody titers in DMD patients seem to decrease with age for both

Neu5Ac- and Neu5Gc-containing glycans, although the significance of this is unclear.

58 Previous studies have shown that deletion of Cmah in mice to mimic the human sialic acid repertoire exacerbates pathology in a mouse model of muscular dystrophy, partially due to antibody-mediated complement activation in Neu5Gc-containing myofibers75.

Given that DMD patients, like Cmah-/- mdx mice, lack endogenous Neu5Gc and exhibit incorporation of Neu5Gc into regenerating myofibers75, it is possible that Neu5Gc as a xeno-autoantigen may play a role in exacerbating DMD pathology in humans as well.

Further work will be needed to assess the role of the immune response to Neu5Gc as an autoantigen in DMD pathology.

2.6 Acknowledgements

For the αDG-N project, we would like to thank Guohong Shao for his assistance in the production of the αDG-N protein, Beth Golden for her immunostaining work,

Susan Gailey and Krista Kunkler for assistance with obtaining clinical serum samples, and Rui Xu for technical assistance with experiments. We would like to thank Han Yin and Igor Dvorchik in the Biostatistics Core at The Research Institute at Nationwide

Children’sHospitalfortheirassistancewithstatisticalanalysis. This work was funded in part by the Seilhamer fellowship from the Jeffrey J. Seilhamer Cancer Foundation and by the Grant-in-Aid of Research from the Ohio State University Chapter of Sigma Xi, the

Scientific Research Honor Society. This work is revised from a publication in the Journal of Neuromuscular Diseases. The final publication is available at IOS Press through http://dx.doi.org/10.3233/JND-150127165.

For the AAV serum titer antibody project, we would like to thank Debbie

Zygmunt for experimental work, Dr. Rui Xu for technical support and Dr. Scott Loiler

59 and the Viral Vector Core at Nationwide Children’sHospitalforproviding purified forms of the rAAV serotypes used. Thanks also to Darren Murrey (NCH) for assistance with the serum antibody ELISA protocol. This work was funded by NIH grant R01 AR049722 to

PTM. This work is revised from a publication in the Human Gene Therapy. The final publication is available at Mary Ann Liebert through https://www.liebertpub.com/doi/10.1089/hum.2016.141166.

We would finally like to thank Sandra Diaz in the Varki lab for performing the glycan array.

60 Figure 1. Standardcurvesof2A3and3B4antibodybindingtoαDG-N protein using different ELISA assay methods. A. Different amounts of purified recombinant αDG-N protein were immobilized on an ELISA plate and binding of αDG-N-specific antibody

2A3 was assayed using a serum-immobilized ELISA. B. 2A3 was immobilized on an

ELISA plate, and different amounts of unlabeled, recombinant αDG-N protein were

61 premixed with a standard concentration of biotinylated αDG-N protein using a competition ELISA. C. 2A3 was immobilized on an ELISA plate, and different amounts of purified recombinant native αDG-N protein (aa30-312) or a fragment of αDG-N made in E. coli (αDG -GST, aa31-141) were added and then detected using a sandwich ELISA with a second αDG-N antibody, 3B4. D. Different amounts of purified recombinant

αDG-N protein were immobilized and binding of 2A3 was assayed using a serum- immobilized ELISA. For the 5ng concentration of αDG-N, the sample was first incubated with or without 1ug/mL 3B4 prior to 2A3 addition. E. Different amounts of purified recombinant αDG-N protein were immobilized and binding of 3B4 was assayed using a serum-immobilized ELISA. For the 5ng concentration of αDG-N, the plate was first incubated with or without 1ug/mL 2A3 prior to 3B4 addition. For all curves, values are the average of duplicate readings. Points without error bars reflect error smaller than the size of the data point. Errors are SEM.

62 Figure 2. Total αDG-N ELISA signal and calculated serum concentration are decreased in serum from patients with DMD compared to normal. A. Otherwise normal adult (n=38), Duchenne muscular dystrophy (DMD, n=9), Becker muscular dystrophy (BMD, n=11), and Inclusion Body Myositis (IBM, n=8) patient serum samples were assayed for levels of αDG-N using the serum-immobilized ELISA. Relative serum

OD450 signal levels are expressed as fold change from the average of normal samples on the same plate. Each datum is the average of 3-7 independent measures per sample.

B. Serum ELISA signals in A were used to calculate serum αDG-N concentration using

αDG-N standard curves. * p <0.05, ** p <0.01, **** p <0.0001

63 DMD, Duchenne muscular dystrophy; BMD, Becker muscular dystrophy; IBM, Inclusion body myositis; f, female; m, male.

Table 1. Summary of relevant patient information.

64 Demographic and/or DMD gene mutation information is shown in 9 DMD patients, 11

BMD patients, 8 IBM patients and 38 otherwise normal patients.

65 Figure 3. Western blot analysis of αDG-N expression in normal human and DMD serum. A. αDG-N purified from cell lysate of transfected HEK293 cells, as well as 4 different DMD and 4 different otherwise normal age-matched human serum samples, were separated by SDS-PAGE and immunoblotted with the 2A3 (αDG-N) antibody or an antibody to fetuin. B. αDG-N purified from transfected HEK293 lysate and supernatant, along with human and DMD serum, were compared with or without deglycosylation to remove both N- and O-linked glycans. For both A and B, anti-fetuin blots were done as a control for serum protein loading and transfer.

66 Figure 4. Signals from serum αDG-N ELISA are independent of age in otherwise normal, BMD and DMD patients, and are independent of gender for normal patients. Serum samples from otherwise normal (n=38) (A), BMD (n=11) (B) or DMD

(n=9) (C) patients were assayed for expression of αDG-N via serum-immobilized ELISA.

D. Signals from αDG-N serum-immobilized ELISA assays of otherwise normal patients were compared by gender. Each datum is the average of 3-7 independent measures per sample.

67 Figure 5. Signals from serum αDG-N ELISA are independent of ambulation status in BMD and DMD patients. Signals from αDG-N serum-immobilized ELISA assays of samples from BMD (n=11) (A) or DMD (n=9) (B) patients were compared by ambulation status at the time of serum collection. Each datum is the average of 3-7 independent measures per sample.

68 Figure 6. Serum αDG-N signal is decreased in mdx Utrn-/- mice relative to wild type

(WT) and mdx mice. Serum samples from wild type (WT, n=11), mdx (n=10), and mdx

Utrn-/- (n=4), mice were assayed for expression of αDG-N using the serum-immobilized

ELISA.

69 A. Expression of relative ELISA signal is reported as fold change from the average of all

WT samples assayed on the same plate.

B. Calculated serum concentrations of αDG-N. Each datum is the average of 2-4 independent measures per sample. ** p <0.01, *** p <0.001, **** p <0.0001

70 Figure 7. Reduction in serum αDG-N signal in mdx Utrn-/- mice as compared to mdx mice is restored in mct mdx Utrn-/- mice. A. Expression of relative ELISA signal from serum samples from mdx (n=3), mdx Utrn-/- (n=4), and mct mdx Utrn-/- (n=2) mice is reported as fold change from the average of all WT samples assayed on the same plate.

71 B. Calculated serum concentrations of αDG-N. Each datum is the average of 2 independent measures per sample. ** p <0.01

72 Figure 8. Immunostaining for αDG-N reveals little or no expression in adult mouse skeletal muscle but positive staining in intramuscular peripheral nerves. The αDG-

N-specific 2A3 monoclonal antibody (top) or αDG-N-specific polyclonal AP1528 antibody (bottom) were used to stain gastrocnemius muscle from adult WT mice (green) and in a separate channel for laminin α2 to identify skeletal myofibers (red) and with

DAPI to stain nuclei (blue). Merge images were captured at 20x (left) and 40x (right).

Bar is 100mm for the left panel and 50mm for the right panel.

73 Figure 9. Western blotting for αDG-N reveals little or no expression in adult mouse skeletal muscle.

74 αDG-N purified from cell lysate of transfected HEK293 cells, as well as 3 WT, 3 mdx, and 3 mdx Utrn-/- mouse gastrocnemius samples were separated by SDS-PAGE and immunoblotted with the 2A3 (αDG-N) antibody, the hDAG (α/β dystroglycan) antibody, the IIH6 (glycosylated α dystroglycan) antibody, or an antibody to actin.

75 Figure 10. Serum αDG-N signal is not significantly decreased in GRMD relative to

GR dogs. Serum samples from GR (n=5) or GRMD (n=5) were assayed for expression of

αDG-N using the serum-immobilized ELISA. Levels are shown as calculated serum concentrations of αDG-N.

76 f, female; m, male; yrs, years; rAAV, recombinant Adeno-associated virus.

Table 2. Antibody titers to rAAV serotypes in young and adult normal human subjects.

77 Highest reciprocal dilutions allowing for positive signal for serum antibodies to rAAVrh74, rAAV8, rAAV1, rAAV2, rAAV6 or rAAV9 are shown in 19 young (<18 years old) and 21 adult (>18 years old) normal subjects.

78 f, female; m, male; yrs, years; rAAV, recombinant Adeno-associated virus; DMD,

Duchenne Muscular Dystrophy; BMD, Becker Muscular Dystrophy.

Table 3. Antibody titers to rAAV serotypes in patients with Duchenne Muscular

Dystrophy (DMD) and Becker Muscular Dystrophy (BMD).

79 Highest reciprocal dilutions allowing for positive signal for serum antibodies to rAAVrh74, rAAV8, rAAV1, rAAV2, rAAV6 or rAAV9 are shown in 22 DMD patients and 16 BMD patients.

80 f, female; m, male; yrs, years; rAAV, recombinant Adeno-associated virus; IBM,

Inclusion Body Myositis; GNE, GNE Myopathy.

Table 4. Antibody titers to rAAV serotypes in patients with Inclusion Body Myositis

(IBM) and GNE Myopathy (GNE). Highest reciprocal dilutions allowing for positive signal for serum antibodies to rAAVrh74, rAAV8, rAAV1, rAAV2, rAAV6 or rAAV9 are shown in 18 IBM patients and 4 GNE patients.

81 Figure 11. Antibody titers to rAAV serotypes in young and adult normal human subjects. Highest reciprocal dilutions allowing for positive signal for serum antibodies to rAAVrh74, rAAV8, rAAV1, rAAV2, rAAV6 or rAAV9 are shown in 19 young normal subjects (A) and 21 adult normal subjects (B).

82 Figure 12. Antibody titers to rAAV serotypes in patients with Duchenne or Becker

Muscular Dystrophy, Inclusion Body Myositis, or GNE Myopathy. Highest reciprocal dilutions required allowing for positive signal of serum antibodies to rAAVrh74, rAAV8, rAAV1, rAAV2, rAAV6 or rAAV9 are shown in patients with Duchenne Muscular

Dystrophy (A), Becker Muscular Dystrophy (B), Inclusion Body Myositis (C), or GNE myopathy (D).

83 Figure 13. Average reciprocal dilution factor for positive signal and propensity of positive signals in young and adult normal patients compared to DMD, BMD, IBM and GNE patients. Average reciprocal dilution factor required for a positive serum antibody signal to rAAVrh74, rAAV8, rAAV1, rAAV2, rAAV6 or rAAV9 (A) and the percentage of patients with positive signal at a 1:50 or 1:800 serum dilution (B) are shown. Errors in A are SEM for n=19 (<18 normal), 21 (>18 normal), 22 (DMD), 16

(BMD), 18 (IBM) or 4 (GNE) patients.

84 Figure 14. Antibody titer to Neu5Gc-containing glycans in otherwise normal and

DMD patients. Antibody titers (ng/µl) are shown for Neu5Gc-containing glycans in in otherwise normal (n=16) or DMD (n=22) patients.

85 Figure 15. Analysis of Neu5Ac and Neu5Gc serum titers by sialic acid type, linkage type, and age. Antibody titers (ng/µl) are shown for Neu5Ac-capped or Neu5Gc-capped glycans (A), for glycans with a 2,3, 2,6, or 2,8 terminal linkage to Neu5Gc (B), or by age for Neu5Ac-capped (C) or Neu5Gc-capped (D) in otherwise normal (n=16) or DMD

(n=22) patients.

86 Figure 16. Schematic of DG-N processing and potential changes in DMD.

Dystroglycan is normally present within the dystrophin-associated glycoprotein (DAG) complex, which includes dystrophin as a link between F-actin, syntrophins, dystroglycan

(DG), which contains both an aand a b chain, sarcoglycans, and other proteins. In the lumen of the Golgi, a dystroglycan is cleaved by furin proteases to liberate the N- terminal domain of the protein, aDG-N, which goes on to be secreted into the serum from muscle and other tissues. In DMD, dystrophin protein is not expressed, and serum aDG-N levels are reduced in the serum using a serum-immobilized ELISA assay.

87 This may reflect reduced intracellular DG expression or stability in DMD muscle, reduced aDG-N stability in DMD muscle or serum, reduced aDG-N secretion from

DMD muscle, or increased aDG-N scavenging in DMD serum. Because aDG-N is immobilized for the ELISA measure, reduced aDG-N in DMD serum may also reflect increased masking of aDG-N by other DMD proteins.

88 Chapter 3. Glycan-based treatments for genetic myopathies.

3.1 Abstract

GNE myopathy is a neuromuscular disorder characterized by progressive muscle atrophy and muscle pathology with inclusion bodies. GNE myopathy is caused by mutations in the GNE gene (UDP-GlcNAc epimerase/Man6-kinase), which encodes a requisite and committed enzymatic step in sialic acid (SA) biosynthesis. Mutations consequently show hypomorphic SA production. These GNE mutations cause a reduction of SA in all tissues, but such reductions are particularly harmful to skeletal muscles, causing muscle wasting and the formation of intracellular muscle inclusion bodies. The primary therapy being explored for GNE myopathy utilizes large doses of oral SA or its biosynthetic glycan precursors. However, it is difficult to understand the extent of therapeutic SA incorporation into skeletal muscle because all reagents that recognize SA recognize the presence of multiple SA forms, precluding visualization of specific SA forms via immunostaining. To address this issue, we have modified the mouse model of

GNE myopathy to endogenously express only a single form of SA, N-acetylneuraminic acid (Neu5Ac), which allows one to visualize the incorporation of a second form of SA,

N-glycolylneuraminic acid (Neu5Gc), in various tissues using immunohistochemistry when Neu5Gc is given as the SA therapeutic. We assessed muscle pathology in this model and the ability of oral SA and GNE gene therapy to rescue muscle hyposialylation.

89 Surprisingly, we found that GNE gene therapy targeting the liver is most effective in increasing levels of SA in the muscle, but may lead to liver pathology. Also, we found that levels of SA therapy equivalent to those being used in human clinical trials can be taken up by muscle and incorporated into the sarcolemmal membrane, while long-term feeding of SA conjugated to dietary glycoproteins is not significantly incorporated. We finally show that reducing, but not overexpressing, GNE in vivo may recapitulate some of the characteristic histopathology of the disease. In total, these findings can help to inform novel therapies for GNE myopathy moving forward.

3.2 Introduction

GNE myopathy is a severe autosomal recessive myopathy characterized by progressive muscle atrophy and weakness. The age of onset is typically in the third or fourth decade of life, beginning with weakness in the tibialis anterior (TA) and hamstring muscles and often rendering patients wheelchair bound by the second decade after diagnosis76. Muscle pathology typically shows rimmed vacuoles and inclusion bodies that contain various proteins, including theAβ1-42 peptide81. GNE myopathy is caused by mutations in the GNE gene, which encodes UDP-N-acetyl-2-epimerase/N- acetylmannosamine kinase87. GNE encodes a bifunctional enzyme required for synthesis of sialic acid (SA)85. The SA biosynthetic pathway culminates in the production of CMP-

Neu5Ac, which is utilized by sialyltransferases to transfer SA onto glycoproteins and glycolipids. CMP-Neu5Ac also acts as a negative feedback regulator of GNE through allosteric inhibition. SA addition is an essential post-translational modification of many glycoconjugates167. GNE mutations, which are more common in Japanese and Middle 90 Eastern populations due to founder mutations, lead to hypomorphic expression of GNE, and consequentially, lowered SA production76.

Cells deficient in GNE activity can be rescued by addition of SA or ManNAc, which can be converted to the end product of GNE activity through GlcNAc-6 kinase97,

107, 124. This result has led to two ideas about therapy for GNE myopathy: gene therapy to replace wild type GNE and glycan therapy to replace SA deficiencies. To date, some glycan therapies have shown some efficacy in the Gne(−/−)hGNED176V-Tg mouse, originally made by Noguchi, Nishino, and colleagues168. This mouse model has a deletion of the wild type mouse Gne coupled with transgenic human mutant GNE (D176V), a mutation common in Japanese GNE myopathy patients. Gne(−/−)hGNED176V-Tg is one of the few GNE myopathy models reported to show appropriate progressive muscle pathology and generalized muscle weakness. By 30 weeks, Gne(−/−)hGNED176V-Tg mice are reported to show significant lifespan reductions, reduced scores in rod climbing and constant speed treadmill walking, and modest elevation in serum CK activity and muscleproductionofAβ1-42 peptide. By 42 weeks, muscles exhibit rimmed vacuoles with congophilic inclusion bodies, as well as pathology in respiratory and cardiac muscles that are not found in human GNE myopathy patients108. However, these glycan therapies have shown minimal to no efficacy in human clinical trials128, 129, underscoring both the need for a better understanding of SA scavenging into muscle, the affected tissue in GNE myopathy, as well as the need for novel therapies for this disease.

There are multiple forms of SA in all species, and some SA forms show differences amongst species that likely have evolutionary relevance. For example, SA

91 composition in humans differs from that in all other mammals, including great apes, by one critical oxygen molecule due to an inactivating CMAH gene mutation that is present in all modern humans. CMAH encodes the CMP-Neu5Ac hydroxylase, which converts

CMP-Neu5Ac to CMP-Neu5Gc169. To mimic this human SA repertoire, Cmah(−/−) mice have been developed that only produce Neu5Ac170. In this model, Neu5Gc is not created endogenously, and therefore can be detected using highly specific purified Neu5Gc antibodies made in birds, which do not express a Cmah gene171.

This unique SA repertoire in humans has proven to be an important modifier for human disease, including bacterial infection, cancer, and neuromuscular disease74. To study the role of Cmah in muscle disease, this gene was deleted in the mdx mouse model of muscular dystrophy to create the mdx Cmah-/- mouse that would predominantly express the Neu5Ac form of SA, similarly to humans. These mdx Cmah-/- mice displayed a more severe disease phenotype than their mdx counterparts, showing worsened mortality, muscle weakness, and histopathology. Interestingly, these mdx Cmah-/- mice, but not mdx or Cmah-/- mice, developed antibody titers to Neu5Gc. As regenerating skeletal muscle fibers in mdx Cmah-/- mice incorporated dietary Neu5Gc, it is likely that this incorporated Neu5Gc acted as a xeno-autoantigen, triggering an immune response and exacerbating pathology. In fact, Neu5Gc-positive regenerating fibers in mdx Cmah-/- mice preferentially exhibited activated complement, supporting this idea75.

As GNE myopathy is a disease of SA, we chose to combine the

Gne(−/−)hGNED176V-Tg model and the Cmah(−/−) model to develop a model of GNE myopathy with a humanized sialylglycome. We used this new model to compare various

92 dosage strategies for oral SA therapy and gene therapy, which could correct the genetic deficit in the disease.

3.3 Materials and Methods

3.3.1 Mice

All animal experiments were conducted with approval from the Institutional

Animal Use and Care Committee (IACUC) at The Research Institute at Nationwide

Children’sHospital.MicelackingCmah (Cmah(−/−)) were obtained from Dr. Ajit

Varki172. Gne knockout mice transgenically expressing human GNE containing the

D176V mutation (Gne(−/−)hGNED176V-Tg) were obtained from Dr. Ichizo Nishino and colleagues108. Gne(−/−)hGNED176V-TgCmah(−/−) mice were obtained by interbreeding of

Cmah(−/−) mice with Gne(−/−)hGNED176V-Tg mice. Mice were fed and watered ad libitum with standard mouse chow, (Cat # 2919, Teklad Global Rodent diet, Harlan,

USA), which contains approximately 6µg of Neu5Gc per gram of chow75, or Neu5Gc- rich porcine submaxillary mucin (PSM), which contains 250µg of Neu5Gc per gram of chow (Dyets, Inc.; Bethlehem, PA) as indicated. For high-dose oral sialic acid (SA) studies, mice received a single dose of 2,000 mg/kg free N-acetylneuraminic acid (Sigma;

St. Louis, MO) or N-glycolylneuraminic acid (Sigma) diluted in drinking water and delivered via oral gavage.

3.3.2 Grip strength and treadmill ambulation

93 Forelimb and hindlimb grip strength were assessed in mice using a grip strength meter (Columbus Instruments). The highest three measurements were averaged from 10 forelimb repetitions and 10 hindlimb repetitions each day; mice were tested once per day for a week, and the daily measurements were averaged to produce a final forelimb and hindlimb measurement. To measure ambulation, the mice underwent a 75-min run on a treadmill with a 15° decline (Treadmill Simplex II; Columbus Instruments). The mice ran for 5 min at 5 m/min, which increased 1 m/min each minute until 15 m/min, and then remained at that speed for an additional 60 min. The time that the mice remained on the treadmill, up to a total time of 75 min, was recorded. After an initial week of training, mice were tested once per day for five days, and the daily measurements were averaged to produce a final measurement of treadmill running time before exhaustion. For all measures, Gne(−/−)hGNED176V-Tg and Gne(−/−)hGNED176V-Tg Cmah(−/−) mice were compared to littermate controls, which included Gne(+/−), Gne(+/−)hGNED176V-Tg,

Gne(+/−)Cmah(−/−), and Gne(+/−)hGNED176V-Tg Cmah(−/−) mice.

3.3.3 Muscle physiology

Tibialis anterior (TA) electrophysiology was performed as described by Hakim et al.173. In brief, mice were anesthetized with ketamine/xylazine, and hindlimb skin was removed to expose the TA muscle. The distal tendon was dissected out and sutured to a force transducer (Aurora Scientific, Aurora, ON). Muscle contractions were produced by stimulation of the sciatic nerve by bipolar platinum electrodes and the optimal length was determined by stretching the muscle until maximum twitch force was elicited. Maximum force was determined by successive stimulation at 50, 100, 150, and 200 Hz with a 1- 94 minute rest period between each stimulation, and specific force was ascertained by dividing the maximum force by the muscle cross sectional area. Finally, the TA underwent 10 cycles of eccentric contractions, each in which the TA is stimulated for

350ms total, the last 200ms of this while being stretched by 10%, before returning to optimum length. For these measurements, the maximum force generated in the first cycle prior to muscle lengthening is designated as 100%.

3.3.4 Histology

Mouse skeletal muscles were snap frozen in liquid nitrogen-cooled isopentane, mounted on corkboard, and cut in 10µm cross-sections on a cryostat. Hematoxylin and

Eosin (H&E) staining was done as before174. Congo Red staining was performed by fixing sections in 95% ethanol and rinsing briefly, before subsequent incubations in

Harris Hematoxylin with acetic acid, alkaline salt solution, and Congo Red solution (Poly

Scientific R&D Corporation; Bay Shore, NY). Finally, slides were dehydrated through graded alcohols, cleared in xylene, and sandwiched with mounting media under a glass cover slip.

3.3.5 Immunofluorescence staining

Mouse skeletal muscles were snap frozen in liquid nitrogen-cooled isopentane, mounted on corkboard, and cut in 10µm cross-sections on a cryostat. Slides being used for FLAG immunostaining were fixed in 4% paraformaldehyde, washed, permeabilized in 1% Triton, and washed again. All slides were blocked with phosphobuffered saline

(PBS) containing 10% goat serum (GS) (for FLAG or lectin staining) or 10% human

95 serum (HS) (for Neu5Gc staining) for 1 hour each. Sections were incubated with anti-

FLAG (F7425, Sigma) or anti-Neu5Gc (Poly21469, Biolegend; San Diego, CA) in blocking solution (10% HS or GS, as before) overnight at 4°C, and washed in PBS.

Slides with anti-FLAG primary were incubated with goat anti-mouse IgG conjugated to

Cy3 (111-165-144, Jackson ImmunoResearch; West Grove, PA), slides with anti-

Neu5Gc primary were incubated with donkey anti-chicken IgG conjugated to Cy3 (703-

165-155, Jackson ImmunoResearch), and slides for lectin staining were incubated with

FITC-labeled Maackia amurensis agglutinin (MAA) lectin (EY Laboratories; San Mateo,

CA), FITC-labeled Sambucus nigra agglutinin (SNA) lectin (EY Laboratories), or

Biotinylated peanut agglutinin (PNA) lectin (EY Laboratories) at room temperature for 1 hour. All sections were washed and cover-slipped with mounting medium containing

DAPI. For green fluorescent protein (GFP) visualization in mouse muscle sections, unprocessed sections were placed in ProLong™ Diamond Antifade Mountant with DAPI

(Thermo Fisher Scientific; Waltham, MA) and cover-slipped. All imaging was done on a

Zeiss Axiophot epifluorescence microscope using AxioVision LE 4.1 imaging software.

3.3.6 Neu5Gc serum titering

Neu5Gc titers were measured comparing mouse serum antibody binding to

Neu5Gc-rich and Neu5Gc-deficient glycoproteins by collaborators in the Varki lab at

UCSD.

3.3.7 Cloning and production of rAAV constructs

96 A cDNA encoding an C-terminal myc/FLAG-tagged wild-type GNE protein

(GNEWT) corresponding to the GNE1 isoform

(MEKNGNNRKLRVCVATCNRADYSKLAPIMFGIKTEPEFFELDVVVLGSHLIDDY

GNTYRMIEQDDFDINTRLHTIVRGEDEAAMVESVGLALVKLPDVLNRLKPDIMIV

HGDRFDALALATSAALMNIRILHIEGGEVSGTIDDSIRHAITKLAHYHVCCTRSAE

QHLISMCEDHDRILLAGCPSYDKLLSAKNKDYMSIIRMWLGDDVKSKDYIVALQ

HPVTTDIKHSIKMFELTLDALISFNKRTLVLFPNIDAGSKEMVRVMRKKGIEHHP

NFRAVKHVPFDQFIQLVAHAGCMIGNSSCGVREVGAFGTPVINLGTRQIGRETGE

NVLHVRDADTQDKILQALHLQFGKQYPCSKIYGDGNAVPRILKFLKSIDLQEPLQ

KKFCFPPVKENISQDIDHILETLSALAVDLGGTNLRVAIVSMKGEIVKKYTQFNPK

TYEERINLILQMCVEAAAEAVKLNCRILGVGISTGGRVNPREGIVLHSTKLIQEWN

SVDLRTPLSDTLHLPVWVDNDGNCAALAERKFGQGKGLENFVTLITGTGIGGGII

HQHELIHGSSFCAAELGHLVVSLDGPDCSCGSHGCIEAYASGMALQREAKKLHD

EDLLLVEGMSVPKDEAVGALHLIQAAKLGNAKAQSILRTAGTALGLGVVNILHT

MNPSLVILSGVLASHYIHIVKDVIRQQALSSVQDVDVVVSDLVDPALLGAASMV

LDYTTRRIY) (Origene; Rockville, MD) was acquired, and site-directed mutagenesis

(Agilent; Santa Clara, CA) was used to make D176V and M712T mutations in the GNE

ORF, generating GNEDV and GNEMT, respectively. GNEWT, GNEDV, and GNEMT were cloned into an rAAV.MCK vector and GNEWT was cloned into an rAAV.LSP vector using NotI sites. This yielded either wild-type or mutant GNE under the control of a muscle-specific promoter and wild-type GNE under a liver-specific promoter, both with

C-terminal myc and FLAG tags.

97 GNEmiR3 and GNEmiR4 shuttle constructs were designed as previously described175 to recognize both human GNE and mouse Gne, synthesized as complementary primers corresponding to each strand of the miRNA shuttle flanked by

Xho I and Spe I sites, and annealed together (Sequences available in Fig. 47). They were then cloned into the U6T6 vector (a kind gift from Dr. Scott Harper) using XhoI and

XbaI sites. GNEmiR4 was subsequently cloned into an rAAV.CMV.eGFP plasmid (from

Dr. Scott Harper) using EcoR1 sites.

rAAVrh74.MCK.GNEWT, rAAVrh74.MCK.GNEDV, rAAVrh74.MCK.GNEMT, and rAAVrh74.LSP.GNEWT, and rAAV.U6.GNEmiR.CMV.eGFP were produced by the

Viral Vector Core at Nationwide Children's Hospital using previously described methods67. The standard triple transfection method was used to produce rAAV in

HEK293 cells176, and packaged vector was purified using sucrose density centrifugation and anion exchange chromatography, as previously described177.

3.3.8 In vitro transfection

Constructs expressing GFP under the control of the liver-specific promoter (LSP) or the cytomegalovirus (CMV) promoter were transfected into HepG2 liver carcinoma cells using the Lipofectamine™2000TransfectionReagent (Thermo Fisher Scientific).

After 48 hours, green fluorescence was visualized using a Leica DM IRB inverted research microscope (Leica Microsystems; Wetzlar, Germany).

3.3.9 Infection with rAAV

98 For the liver- and muscle-specific expression of GNE, one- to two-month old

Gne(−/−)hGNED176V-Tg or Gne(−/−)hGNED176V-TgCmah(−/−) mice were injected via unilateral, intramuscular (IM) injection of 1E+11 vector genomes (vg)/TA and 5E+11 vg/gastrocnemius of rAAVrh74.MCK.GNEWT or unilateral, IM injection of 1E+11 vg/TA

(−/−) rAAVrh74.LSP.GNEWT as a control. Additionally, Gne hGNED176V-Tg and

Gne(−/−)hGNED176V-TgCmah(−/−) mice were injected via intraperitoneal (IP) injection with 5E+11 vg/mouse (approximately 2.5E+13vg/kg) of rAAVrh74.LSP.GNEWT or rAAVrh74.MCK.GNEWT as a control. Mice were sacrificed 10 months post-injection for analysis.

For expression of wild-type and mutant GNE in skeletal muscle, one month-old

Gne(+/−) and Gne(−/−)hGNED176V-TgCmah(−/−) mice were injected via unilateral, intramuscular (IM) injection with 1E+11 vg/TA, 5E+11 vg/gastrocnemius, and 5E+11 vg/quadriceps of rAAVrh74.MCK.GNEWT, rAAVrh74.MCK.GNEDV, or rAAVrh74.MCK.GNEMT. Mice were sacrificed 1, 3, or 6 months post-injection for analysis.

For knockdown of endogenous Gne in skeletal muscle, one month-old Gne(+/−) mice were injected via unilateral, intramuscular (IM) injection with 2E+11 vg/TA of rAAV.U6.GNEmiR.CMV.eGFP. Mice were sacrificed 1, 3, or 6 months post-injection for analysis.

3.3.10 Quantitative polymerase chain reaction (qPCR)

TaqMan quantitative polymerase chain reaction (qPCR) was used to quantify rAAV vector genome (vg) copies in rAAV-treated muscles. Genomic DNA was extracted 99 as previously described67 from treated and untreated muscles or liver. DNA purity and quantity were measured using an ND-1000 spectrophotometer (NanoDrop; Thermo

Fisher Scientific). Vector-specific primer/probe sets were used to amplify a portion of the vectorDNAencompassingthe3′endoftheMCK promoterandthe5′endoftheGNE cDNA (forward:5′-CGGAAGTGTTACTTCTGCTCTAA, probe:5′-/56-

FAM/CCGCCACCA/ZEN/TGGAGAAGAATGGAA/3IABkFQ/, andreverse:5′-

CGGTTACAAGTAGCAACACAAAC) for rAAVrh74.MCK.GNE, encompassingthe3′ end of the LSP promoterandthe5′ end of the GNE cDNA (forward:5′-

TTTCTCAGGATAACAAGAACGAAAC, probe:5′-/56-

FAM/AGCCACCGC/ZEN/TAGCAAGAATTCGAT/3IABkFQ/, andreverse:5′-

CCGCAGCTTTCGGTTATTTC) for rAAVrh74.LSP.GNE, and encompassing the 3′end of the CMV promoterandthe5′endoftheeGFP sequence (forward:5′-

TGTACGGTGGGAGGTCTAT,probe:5′-/56-

FAM/AGCAGAGCT/ZEN/GGTTTAGTGAACCGT/3IABkFQ/,andreverse:5′-

CTGAACTTGTGGCCGTTTAC) for rAAV.U6.GNEmiR.CMV.eGFP. The plasmids used to make virus were linearized with BamHI for rAAV.U6.GNEmiR.CMV.eGFP or

MluI for rAAVrh74.MCK.GNE and rAAVrh74.LSP.GNE, re-purified, and utilized to generate a standard curve from 50 to 5 million copies in log increments. The correlation coefficient of the standard curve equaled or exceeded 0.99 for all runs. Samples were measured in duplicate, and data are reported as vgs per microgram of genomic DNA assayed. Values for each group were averaged, and differences between the groups were analyzed using an ordinary one-way ANOVA and Tukey's multiple comparisons test.

100 3.3.11 Semi-quantitative real-time polymerase chain reaction (qRT-PCR)

Total RNA was isolated from frozen blocks of skeletal muscles or liver by use of

TRIzol reagent (Invitrogen, Carlsbad, CA). RNA was purified on a silica gel-based membrane-containing column (RNeasy; Qiagen, Germantown, MD). Relative levels were assessed by semiquantitative real-time PCR (qRT-PCR) using

178 theΔCT method, with 18S rRNA as an internal reference . A high-capacity cDNA archive kit (Applied Biosystems, Foster City, CA) was utilized for the reverse transcription RNA as per the manufacturer's guidelines. Samples underwent real-time

PCR in duplicate using a TaqMan ABI 7500 sequence detection system (Applied

Biosystems). Primer/probe sets for 18S (assay ID: Hs99999901_s1), human GNE (Assay

ID: Hs01103400_m1) and mouse Gne (assay ID: Mm00450174_m1) were purchased from Applied Biosystems (Thermo Fisher Scientific). A primer/probe set to recognize the myc/FLAG tag region common to both rAAVrh74.MCK.GNE and rAAVrh74.LSP.GNE was designed using PrimerQuest DNA software and synthesized by Integrated DNA

Technologies (Coralville, IA)- 5’-CTACACAACACGCAGGATCT(forward),5’-56-

FAM/TCTCAGAAG/Zen/AGGATCTGGCAGCAA/3IABkFQ/(probe),and5’-

ATCGTCGTCATCCTTGTAATCC (reverse). In all cases, relative mRNA levels were averaged for each group. For the tissue-specific GNE constructs, data are presented as

ΔCT for GNE tag expression, GNE gene expression relative to uninjected

Gne(−/−)hGNED176V-TgCmah(−/−) tissue for liver, and GNE tag expression relative to

Gne expression in uninjected wild type (WT) mice for skeletal muscle. For the GNE miR shuttle, data are presented as relative expression to wild-type Gne levels. Differences 101 between the groups were analyzed using an ordinary one-way ANOVA and Tukey's multiple comparisons test.

3.3.12 Western blotting and immunoprecipitation

Mouse skeletal muscle was lysed in buffers containing NP-40 (NP), sodium dodecyl sulfate (SDS), or SDS and 4% urea (Ur) where indicated. Where no buffer is specified, NP buffer was used. Where indicated, muscle lysates were immunoprecipitated using ANTI-FLAG® M2 Affinity Gel (Sigma)accordingtomanufacturer’sinstructions.

For Western blotting, 40µg of protein from muscle lysate or 500µg of immunoprecipitated protein were diluted in SDS denaturing buffer and separated on 4-

12% gradient SDS-PAGE gels and then transferred to nitrocellulose. After transfer, blots were blocked in TBST with 5% NFDM, then incubated with primary antibody, either anti-FLAG (F7425, Sigma) or anti-Gapdh (MAB374, MilliporeSigma; Burlington,

Massachusetts), washed in TBST, incubated with appropriate horseradish peroxidase

(HRP)-conjugated goat anti-rabbit IgG (111-035-144, Jackson ImmunoResearch; Seattle,

WA) or donkey anti-mouse IgG (715-035-150, Jackson ImmunoResearch) secondary antibody, washed again, and developed using an ECL developing kit (Amersham,

Piscataway, NJ), much as previously described140. Quantification of protein bands was performed by densitometry through use of NIH ImageJ software.

3.3.13 Luciferase assay

To assess knockdown of GNE in vitro, a luciferase assay was utilized as previously described179. In brief, the GNE ORF was cloned downstream of the Renilla

102 luciferase stop codon in the Psicheck2 vector (Promega; Madison, WI) using NotI sites; thus, targeting of the GNE ORF with a specific miR shuttle could reduce Renilla luciferase expression. This vector also contains Firefly luciferase driven by the thymidine kinase (TK), which would not be affected by miR shuttle targeting. So, this GNE- containing Psicheck2 vector was transfected with or without U6.GNEmiR4 or

U6.GNEmiR3 into HEK293 cells in a 1:5 molar ratio, and GNEmiR efficacy was assessed by measuring relative expression of Renilla and Firefly luciferase using the Dual

Luciferase Reporter Assay System (Promega). Data is reported as relative luciferase expression to Psicheck2 transfection alone, and errors are reported as SEM for triplicate measurements.

3.3.14 Statistics

Comparisonoftwogroupswasassessedbystudent’stwo-tailed, unpaired t test, and comparison of more than two groups was assessed using one-way analysis of variance (ANOVA) with post-hoc Tukey's pairwise comparison. Survival curves were assessed using the log-rank test, and R square values were determined by linear regression. Measures with p<0.05 were considered significant in all cases. Statistics were analyzed using GraphPad Prism Version 6.03 (GraphPad Software Inc., La Jolla, CA).

3.4 Results

3.4.1 Characterization of the Gne(−/−)hGNED176V-Tg and Gne(−/−)hGNED176V-

TgCmah(−/−) models of GNE myopathy

103 We first bred the Gne(−/−)hGNED176V-Tg mouse model of GNE myopathy with the Cmah(−/−) line, which has a humanized sialylglycome, to create the novel

Gne(−/−)hGNED176V-TgCmah(−/−) GNE myopathy model that would predominantly express a single form of sialic acid (SA), Neu5Ac. We wanted to assess GNE myopathy pathology in this novel model as compared to the original Gne(−/−)hGNED176V-Tg model, as the original model was reported to show marked muscle atrophy, gross motor dysfunction, and GNE myopathy-like histopathology180. In all cases, we compared these mice to littermate controls, which in our case included Gne(+/−), Gne(+/−)hGNED176V-Tg,

Gne(+/−)Cmah(−/−), and Gne(+/−)hGNED176V-Tg Cmah(−/−) mice, as hGNED176V- transgene-expressing and Gne heterozygote mice have not differed from wild type (WT) mice in previous studies108.

Surprisingly, neither Gne(−/−)hGNED176V-Tg nor Gne(−/−)hGNED176V-

TgCmah(−/−) mice exhibited a significant decrease in lifespan over 15 months (Fig. 17A).

We next assessed body weight in these lines; body weight for male and female mice were considered separately, as males generally weigh more than females of the same genotype180. We found no difference in gender-matched weight apart from a change between littermate controls and Gne(−/−)hGNED176V-Tg female mice at 9 and 12 months of age (p<0.01 and p<0.05, respectively) (Fig. 17B for male, Figure 17C for female). We also considered muscle and organ weight; in this case, muscle and organ measurements were normalized to mouse body weight, allowing male and female measurements to be pooled. No changes between any groups were seen in weights of skeletal muscles, including tibialis anterior (TA), gastrocnemius, quadriceps, or triceps (Fig. 17D). Hearts

104 and spleens were significantly enlarged in both Gne(−/−)hGNED176V-Tg and

Gne(−/−)hGNED176V-TgCmah(−/−) mice as compared to littermate controls (p<0.05 for each comparison), but did not differ between Gne(−/−)hGNED176V-Tg and

Gne(−/−)hGNED176V-TgCmah(−/−) mice for either heart or spleen (p=0.843 and p=0.993, respectively) (Fig. 17E). So, apart from minor differences in female body weight at some time points, neither Gne(−/−)hGNED176V-Tg nor Gne(−/−)hGNED176V-TgCmah(−/−) mice exhibited the originally reported alterations in lifespan, mouse weight, and muscle size as compared to littermate controls.

We next wanted to evaluate gross motor function in these lines, which was also reported to be impaired in the original Gne(−/−)hGNED176V-Tg model108. To do this, we assessed treadmill performance as well as forelimb and hindlimb grip strength every three months from 6 to 15 months of age (Fig. 18). We found few significant changes in grip strength and treadmill performance; Gne(−/−)hGNED176V-TgCmah(−/−) hindlimb grip strength was significantly decreased from that of Gne(−/−)hGNED176V-Tg mice at 6 months of age (p<0.05) and Gne(−/−)hGNED176V-Tg forelimb performance was significantly increased as compared to littermate controls at 15 months of age (p<0.05), but all other comparisons were insignificant (Fig. 19). We additionally wanted to assess histopathology in these mice, as Gne(−/−)hGNED176V-Tg muscles were originally reported to show atrophy, rimmed vacuoles, and amyloid-positive inclusions via H&E and Congo Red staining beginning at approximately 10 months of age and worsening over time108. In our hands, however, no histopathology was seen in skeletal muscles of

105 Gne(−/−)hGNED176V-Tg or Gne(−/−)hGNED176V-TgCmah(−/−) mice as assessed by H&E and Congo Red staining even at 15 months of age (Fig. 20).

Given the unexpected lack of histopathology and gross motor deficits in

Gne(−/−)hGNED176V-Tg mice, we next wanted to assess muscle electrophysiology in this line. To do so, we measured absolute force (Fig. 21A), specific force (Fig. 21B), and force drop induced by eccentric contraction (Fig. 21C) in the TA of

Gne(−/−)hGNED176V-Tg mice as compared to littermate controls. Again, measures did not significantly differ between these two groups in all instances (n=4 per group).

Finally, given that the mdx Cmah(−/−) model of muscular dystrophy, but not the mdx line nor the Cmah(−/−) line, develops antibodies to dietary Neu5Gc, we wished to assess antibody titer to Neu5Gc in our novel Gne(−/−)hGNED176V-TgCmah(−/−) model.

We found that 2 of 5 Gne(−/−)hGNED176V-TgCmah(−/−) mice, but no

Gne(−/−)hGNED176V-Tg or Cmah(−/−) mice, exhibited Neu5Gc titers above 0.1µg/ml (Fig.

22). Thus, with a few minor exceptions, Cmah deletion did not alter muscle pathology or function in the Gne(−/−)hGNED176V-Tg mouse model of GNE myopathy.

3.4.2 High dose gavage of free SA, but not long-term SA-enriched food, can be incorporated into muscles and organs in the Gne(−/−)hGNED176V-TgCmah(−/−) model

Given that the novel Gne(−/−)hGNED176V-TgCmah(−/−) model expresses a single form of SA, Neu5Ac, we wishedtoassessthemodel’s utility in visualizing exogenous

Neu5Gc through immunofluorescence with a Neu5Gc-specific antibody, as visualization of local SA expression in muscle after oral therapy had never before been possible. To do this, we first wished to assess both Neu5Gc levels and overall sialylation in these lines. 106 We did this through immunostaining with a Neu5Gc-specific antibody and the peanut agglutinin (PNA) lectin, which recognizes subterminal β-1,3-N-acetylgalactosamine structures that are unmasked when terminal SA residues are absent. As such, Neu5Gc staining would represent levels of the Neu5Gc form of SA, while PNA staining would represent an overall measure of SA loss regardless of Neu5Ac or Neu5Gc form. We found that as compared to WT muscle, endogenous Neu5Gc staining was decreased in

Gne(−/−)hGNED176V-Tg muscle (likely due to overall hyposialylation in this line) and absent in Gne(−/−)hGNED176V-TgCmah(−/−) muscle, confirming lack of endogenous

Neu5Gc SA in our novel model. However, both Gne(−/−)hGNED176V-Tg and

Gne(−/−)hGNED176V-TgCmah(−/−) muscles exhibited a comparable increase in PNA staining, including both elevated overall cytoplasmic staining and development of PNA- positive foci. These foci may reflect staining nuclei of myofibers or staining of mononuclear cells in muscle sections, such as satellite cells, intramuscular fibroblasts, or immune cells (Fig. 23). Together, these results indicate that while Gne(−/−)hGNED176V-

Tg and Gne(−/−)hGNED176V-TgCmah(−/−) muscles have roughly equivalent loss of overall

SA, the Neu5Gc form of SA is specifically absent in Gne(−/−)hGNED176V-TgCmah(−/−) mouse muscles.

As endogenous Neu5Gc was, as expected, absent in Gne(−/−)hGNED176V-

TgCmah(−/−) muscles, we next wanted to evaluate the extent to which exogenous Neu5Gc

SA could be detected in muscles or organs after oral delivery. To test this, we first delivered 2,000 mg/kg of free SA to this model via oral gavage in the form of Neu5Gc

(or Neu5Ac as a control), and assessed Neu5Gc immunostaining in muscle and organs

107 one week after treatment (Fig. 24). This rather high amount of SA is roughly equivalent to therapeutic oral SA doses being given to GNE myopathy patients in current clinical trials128. We found that exogenous Neu5Gc incorporated into the sarcolemmal membrane of skeletal muscles, including TA, gastrocnemius, quadriceps, triceps, and diaphragm.

Interestingly, while the diaphragm exhibited sarcolemmal staining of all fibers, other skeletal muscles, and in particular the gastrocnemius, seemed to show preferential incorporation into small fibers (likely slow-twitch, type I fibers). The heart, by contrast, displayed intracellular Neu5Gc staining rather than sarcolemmal (Fig. 25). In addition,

Neu5Gc showed significant intracellular incorporation into stomach and lung, as well as more sparse intracellular incorporation into kidney. The pancreas exhibited staining in the spaces surrounding acini, possibly indicating Neu5Gc incorporation into capillaries of the ascinar vascular system, while the liver did not show significant Neu5Gc incorporation

(Fig. 26).

We next wished to assess the extent to which Neu5Gc could be incorporated into muscle from dietary sources. To do so, we fed Gne(−/−)hGNED176V-TgCmah(−/−) mice with either standard mouse chow or chow enriched with porcine submaxillary mucus

(PSM), which has an approximately 40 times higher Neu5Gc content than standard mouse chow. After 10 months, we sacrificed these mice and stained muscles and organs with a Neu5Gc-specific antibody, as before (Fig. 27). In this feeding paradigm, skeletal muscles including TA, gastrocnemius, quadriceps, triceps, diaphragm, and heart did not significantly incorporate Neu5Gc (Fig. 28). Similarly, only large intestine, but not kidney, lung, pancreas, or liver, was able to incorporate Neu5Gc from PSM-enriched

108 food (Fig. 29). Thus, muscles appear not to take up SA as Neu5Gc from dietary sources even after many months of exposure, while acute oral treatment with Neu5Gc monosaccharide at very high doses did lead to SA incorporation in muscle sarcolemmal membranes.

3.4.3 Liver-specific expression of GNE gene therapy is more effective than muscle- specific GNE gene therapy at increasing muscle SA levels, but may lead to liver pathology

We next wanted to assess efficacy of GNE gene therapy in rescuing the muscle hyposialylation in GNE myopathy. Specifically, we wanted to assess GNE gene therapy directed to the muscle, where GNE myopathy pathology occurs, and compare it to GNE gene therapy directed to the liver, an easily transducible target with the potential to produce sialylated serum proteins that can be subsequently be scavenged by muscle tissue. To do this, we designed two rAAV constructs to express GNE exclusively in muscle or liver using the muscle creatine kinase (MCK) promoter or a liver-specific promoter (LSP), respectively (Fig. 30). The LSP utilized here is the human thyroxine- binding globulin promoter, which was originally characterized by Hayashi et al. and subsequently developed as a promoter for liver-based therapies 181-183. As our group and others have previously described the muscle-specific nature of the MCK promoter67, 184,

185, we first wanted to confirm the ability of the LSP promoter to express in liver cells in vitro. As expected, we found that the LSP promoter could drive gene expression in liver carcinoma HepG2 cells, albeit with weaker expression than the strong, ubiquitous cytomegalovirus (CMV) promoter used here as a control (Fig. 31). 109 To compare the LSP and MCK promoters in vivo, we injected

Gne(−/−)hGNED176V-Tg and Gne(−/−)hGNED176V-TgCmah(−/−) mice via unilateral, intramuscular (IM) injection with 1E+11 vg/TA and 5E+11 vg/gastrocnemius of rAAVrh74.MCK.GNE in Gne(−/−)hGNED176V-TgCmah(−/−) mice. We injected 1E+11 vg/TA of rAAVrh74.LSP.GNE, which will deliver the same amount of rAAV but not express transgene in skeletal muscle, as a control. Additionally, we injected

Gne(−/−)hGNED176V-Tg and Gne(−/−)hGNED176V-TgCmah(−/−) mice via intraperitoneal

(IP) injection with 2.5E+13/kg of rAAVrh74.LSP.GNE. We also delivered rAAVrh74.MCK.GNE via IP injection in Gne(−/−)hGNED176V-TgCmah(−/−) mice as a control (Table 5, Fig. 32). Ten months after injection, we sacrificed these mice and confirmed biodistribution and GNE expression in each group through qPCR and qRT-

PCR (Fig. 33A and 33B, respectively). As expected, IM injections generally induced higher biodistribution in muscle than liver, with IM AAV.MCK.GNE biodistribution significantly differing from IP injection conditions in both TA and gastrocnemius

(p<0.001 in all cases) (Fig. 33A). Biodistribution in liver, on the other hand, did not significantly differ between injection conditions, likely because some virus travels to the liver via the bloodstream after IM injection. Notably, control IM injection of

AAV.LSP.GNE was only performed in TA and not gastrocnemius of

Gne(−/−)hGNED176V-TgCmah(−/−)mice. As such, AAV.LSP.GNE biodistribution was not assessed in the gastrocnemius muscle; in addition, mice in this IM AAV.LSP.GNE group had markedly lower biodistribution in the liver, as they received less total virus than mice

110 receiving an IM AAV.MCK.GNE injection. This condition, therefore, was not used for statistical analysis in liver.

As normalization of gene expression proved difficult in this paradigm (due to very high levels of mutant D176V GNE transgene expression in skeletal muscle in this mouse model), we first assessed GNE transgene expression arising from rAAV as the change in cycle threshold (dCT) as compared to an 18S control using probes that discriminated the tagged, rAAV-derived GNE transgene from the endogenous human

GNE transgene in the mouse model (Fig. 33B). These probes showed no signal in uninjected control muscles, but could be used to assess relative transgene amounts in rAAV-injected muscles. Of particular importance to note here is that the dCT value is inversely proportional to transcript abundance (so lower numbers reflect higher gene expression). In this comparison, we found that IP injection of AAV.LSP.GNE caused a significantly lower dCT (reflecting higher transcript abundance) in liver than that of all other injection conditions (p<0.05 for all such comparisons). Likewise, in both the TA and the gastrocnemius, IM injection with AAV.MCK.GNE lowered dCT values as compared to all other injection conditions (p<0.001 for all such comparisons).

Unexpectedly, IM injection of AAV.LSP.GNE also induced lower dCT values as compared to IP injection of either AAV.MCK.GNE or AAV.LSP.GNE (p<0.01 for all such comparisons), but to a lesser extent than IM AAV.MCK.GNE. This finding may reflect some leakiness of the LSP promoter, allowing gene expression in muscle. For liver samples, we also normalized GNE dCT values in our injection paradigms to those in uninjected Gne(−/−)hGNED176V-TgCmah(−/−) mice, in which the only GNE expression 111 present is from transgenic expression of GNE with the D176V mutation (Fig. 33C). Here we used a primer set that recognizes the GNE gene itself as opposed to the epitope tag, so that GNE measures reflected both rAAV-injected GNE transgene and endogenous GNE transgene. For this comparison, we found that IP injection of AAV.LSP.GNE increased

GNE expression by 7-fold as compared to uninjected Gne(−/−)hGNED176V-TgCmah(−/−) liver, statistically differing from IM injection of AAV.MCK.GNE (p<0.05). GNE expression levels in the other three injection conditions were similar to those in uninjected mice. As this transgene is highly overexpressed in muscle as compared to liver

(by ~3000-fold) due to the promoter used in creation of this transgenic line, this kind of comparison to uninjected Gne(−/−)hGNED176V-TgCmah(−/−) tissue was not feasible in skeletal muscle. As such, here we also chose to evaluate GNE expression in muscle by normalizing to endogenous mouse Gne levels in uninjected WT mice (Fig. 33D). For this analysis, we used primer sets specific to the epitope tag of rAAV-delivered GNE gene and to the mouse Gne gene. Using this comparison, we found that GNE expression in

Gne(−/−)hGNED176V-TgCmah(−/−) skeletal muscle after IM injection of AAV.MCK.GNE reached levels 20-30-fold higher than endogenous Gne expression in uninjected WT mice; these significantly differed from expression levels in all other injection conditions for the TA (p<0.01 for all such comparisons), and differed from IP injection of

AAV.LSP.GNE for the gastrocnemius (p<0.05). We chose not to do this type of analysis in liver, as endogenous mouse Gne gene expression is much higher (6 times) in liver as it is in skeletal muscle, again making the denominator very high in such comparisons.

112 Western blotting of liver and muscle lysates confirmed protein expression of IM- injected rAAVrh74.MCK.GNE in muscle and IP-injected rAAVrh74.LSP.GNE in liver.

Moreover, protein solubility was not changed by extraction in buffer containing NP40,

SDS, or SDS with 4% urea, suggesting an absence of insoluble protein aggregates (Fig.

34). Liver expression from the LSP promoter was also confirmed in vivo through immunostaining of the FLAG epitope tag in mice after intraperitoneal injection with rAAVrh74.LSP.GNE (Fig. 35). No injection condition altered weight of skeletal muscles, including TA, gastrocnemius, quadriceps, or triceps, or organs, including heart and spleen

(Fig 36). While spleen showed a non-significant trend towards increased size with viral injection, this is likely due to the variability in spleen size in this genotype as a whole.

Thus, delivery of GNE to the liver or the muscle was confirmed at the DNA, RNA, and protein levels, and this long-term expression had no significant effect on muscle or organ weight.

We next wanted to assess the effect muscle- or liver-specific GNE expression on

SA expression in muscle, as muscle hyposialylation is thought to be causative in GNE myopathy pathology. We first did so by comparing Neu5Gc expression in

Gne(−/−)hGNED176V-Tg mice injected with IM rAAVrh74.MCK.GNE or IP rAAVrh74.LSP.GNE. Gne(−/−)hGNED176V-TgCmah(−/−) mice were not assessed in this way as they do not endogenously express the Neu5Gc form of SA, and as such, Neu5Gc expression would be absent regardless of overall sialylation changes. As expected, IM rAAVrh74.MCK.GNE injection increased muscle, but not liver, expression of SA. As seen in the oral gavage feeding paradigm, this Neu5Gc staining was localized to the

113 sarcolemmal membrane. IP injection of rAAVrh74.LSP.GNE also increased cytoplasmic expression of SA in hepatocytes; unexpectedly, however, this liver-directed GNE expression increased muscle SA to an even greater extent than IM rAAVrh74.MCK.GNE injection (Fig. 37). To confirm this finding in Gne(−/−)hGNED176V-TgCmah(−/−) mice, we assessed skeletal muscle and liver staining using the Maackia amurensis agglutinin

(MAA) and Sambucus nigra agglutinin(SNA)lectins,whichrecognizeα-2,3 and α-2,6- linked SA, respectively, on glycoproteins and glycolipids. These lectins can be used to recognize SA in Gne(−/−)hGNED176V-TgCmah(−/−) mice, as they do not discriminate between Neu5Ac and Neu5Gc forms of SA. We also looked at staining with the PNA lectin, which recognizes subterminal β-1,3-N-acetylgalactosamine structures that are unmasked when terminal SA residues are absent186. In these Gne(−/−)hGNED176V-

TgCmah(−/−) mice, IP rAAVrh74.LSP.GNE but not IM rAAVrh74.MCK.GNE increased liver sialylation as evidenced by increased SNA staining and decreased PNA staining

(Fig. 38). Similarly, in Gne(−/−)hGNED176V-Tg mice, we saw that IP injection of rAAVrh74.LSP.GNE, but not IM injection of rAAVrh74.MCK.GNE, increased muscle sialylation, again evidenced by increased MAA staining and decreased PNA staining

(Fig. 39).

While neither the Gne(−/−)hGNED176V-Tg nor the Gne(−/−)hGNED176V-

TgCmah(−/−) model endogenously exhibited muscle histopathology in our hands up to 15 months of age, we wanted to confirm that AAV-mediated overexpression of GNE in muscle and liver did not induce such histopathology in IP rAAVrh74.LSP.GNE- and IM rAAVrh74.MCK.GNE-injected mice. As expected, no histopathology was evident via

114 Congo Red and H&E in skeletal muscle after IM injection of rAAVrh74.MCK.GNE or IP injection of rAAVrh74.LSP.GNE. However, H&E staining of liver revealed that 5/5 IP rAAVrh74.LSP.GNE-injected Gne(−/−)hGNED176V-TgCmah(−/−) mice, but 0/4 IM rAAVrh74.MCK.GNE-injected Gne(−/−)hGNED176V-TgCmah(−/−) mice, showed unexpected, pigment-rich brown lesions within the liver tissue (Fig. 40 20X, Fig. 41

40X). Such brown staining is not normal in liver, or in any tissue, when hematoxylin and eosin staining are used, as hematoxylin stains tissues pink or red and eosin stains nuclei blue. The most likely result of the brown staining in these treated livers was that it reflected a build-up of bile, which would produce such brown staining via H&E. This identified liver histopathology likely arose from GNE transgene expression in the liver resulting from rAAVrh74.LSP.GNE injection, as liver biodistribution of rAAV in rAAVrh74.MCK.GNE-injected livers was similar in terms of the amounts of rAAV vector genomes present, yet these livers did not show such liver pathology. As such,

AAV capsid burden, which was common between these two vectors, likely did not cause this liver-specific pathology, while liver GNE overexpression, which was elevated by the

LSP promoter but absent with MCK, more likely did.

3.4.4 GNE overexpression does not recapitulate GNE myopathy histopathology

It is currently unclear whether the pathology of GNE myopathy is truly caused by a loss of SA from hypomorphic enzyme mutations, or whether it is caused by some aberrant function of mutant GNE, as has been suggested187-189. To address this question, we wanted to manipulate levels of mutant and wild-type GNE expression in the mouse

115 skeletal muscle to assess the extent to which these conditions could recapitulate GNE myopathy histopathology.

First, we wanted to test the possibility that overexpression of mutant GNE itself was causative in GNE myopathy pathology by overexpressing wild-type or mutant GNE in a mouse model expressing only wild-type GNE (Gne(+/-)) or expressing only GNE with the D176V Japanese founder mutation (Gne(−/−)hGNED176V-TgCmah(−/−) ). So, in these models, we overexpressed wild-type GNE (GNEWT), GNE containing the Japanese

D176V founder mutation (GNEDV), or GNE containing the M712T Iranian Jewish founder mutation (GNEMT). We cloned these under the control of the MCK promoter and injected 5E+11/muscle of each virus intramuscularly into the gastrocnemius and quadriceps of Gne(+/−) and Gne(−/−)hGNED176V-TgCmah(−/−) mice, sacrificing one, three, or six months post-injection. To confirm protein expression in muscle after injection, we performed immunoprecipitation and Western blotting of Gne(+/−) gastrocnemius muscle lysates using the FLAG tag, finding that GNE levels significantly increased between 1 and 3 months post-injection and peaked 6 month-post injection. Interestingly, in addition to the monomeric form of GNE predicted at 79kDa, we also observed a higher band that likely represented the dimeric form of GNE (predicted to be 158 kDa) in the presence of

SDS, an ionic detergent expected to solubilize all proteins, and beta-mercaptoethanol, which is expected to reduce all disulfide bonds. This dimeric band was significantly more prominent in the GNEWT-injected muscles as compared to the GNEDV- or GNEMT-injected muscles; at 6 months of age, where overall protein expression was highest in all cases, this dimeric band represented 64% of total wild-type GNE, but a mere 24% and 0% of

116 total GNE with D176V and M712T mutations, respectively (Fig. 42). We next assessed protein localization via immunostaining of GNEWT-, GNEDV- or GNEMT-injected muscles after 6 months, at which time point highest protein expression was achieved. We found that upon overexpression, wild-type and mutant GNE all showed the expected cytoplasmic localization in muscle in both Gne(+/−) and Gne(−/−)hGNED176V-TgCmah(−/−) gastrocnemius muscle (Fig. 43). Likewise, overexpression of GNEWT, GNEDV, or GNEMT did not induce histopathology in these muscles as evidenced by H&E staining (Fig. 44) and Congo Red staining (Fig. 45). So, while wild-type GNE may show different multimerization than mutant GNE when overexpressed in vivo, this did not induce histopathology or alter protein localization in skeletal muscle.

3.4.5 Gne knockdown may recapitulate some GNE myopathy histopathology

As overexpression of mutant GNE was unable to induce GNE myopathy-like pathology in vivo, we next wanted to assess the extent to which SA reduction itself could recapitulate this histopathology. To achieve this, we wanted to reduce endogenous levels of wild-type Gne in mice, thus reducing SA levels without introducing the presence of mutant GNE. Given the fact that complete Gne knockout shows embryonic lethality in mice107, and that Gne(+/−) mice show a roughly 50% reduction in SA without inducing a

GNE myopathy-like muscle phenotype190, we chose to further reduce Gne expression in

Gne(+/−) mice through use of an artificial miRNA shuttle delivered to skeletal muscle via

AAV. To do this, we first designed four artificial miRNA shuttle constructs (Fig. 46) using previously described methods175, and assessed the ability of two of these constructs,

GNEmiR3 and GNEmiR4, to knock down GNE expression in vitro. We found that 117 GNEmiR4 was more effective than GNEmiR3 at reducing GNE expression in cultured cells, effecting reduction by approximately 60% via luciferase assay (Fig. 47A) and 50% via qRT-PCR (Fig. 47B). We subsequently cloned GNEmiR4 under the control of a U6 promoter into an rAAV cassette, including a GFP reporter under the control of a CMV promoter (Fig. 48A).

To test this GNE miR in vivo, one-month-old Gne(+/−) mice were given IM injections of 2E+11 vg/TA of rAAV.U6.GNEmiR.CMV.eGFP and sacrificed one, three, or six months post-injection (Fig. 48B). Fluorescent microscopy revealed strong fluorescence of the GFP reporter, indicating expression of the construct in injected skeletal muscle (Fig. 49). Biodistribution was assessed via qPCR (Fig. 50A), and qRT-

PCR showed that knockdown increased with longer expression of GNE miR shuttle, with the six-month time point showing a 34% reduction in Gne expression in the TA as compared to expression in Gne(+/−) mice (Fig. 50B), with an overall knockdown of 58% relative to WT, compared to a knockdown of only 36% in Gne(+/−) relative to WT.

We next wanted to assess the ability of this Gne knockdown to produce GNE myopathy-like pathology in vivo, including muscle atrophy and histopathology. We found that reduction in Gne expression did not alter the size of injected TA muscles at any time point (Fig. 51). No histopathology was seen in uninjected Gne(+/−) TA muscles at any time point via H&E or Congo Red; likewise, no histopathology was seen in muscles that expressed GNEmiR4 for one or three months. Interestingly, however,

Gne(+/−) TA muscle that expressed GNEmiR4 for six months showed Congo Red-positive staining, indicating the presence of amyloid-positive inclusions (Fig. 52). This Congo

118 Red-positive staining was confirmed by brightfield, polarized light birefringence, and rhodamine fluorescence imaging at 40X magnification (Fig. 53). Thus, reduction of Gne beyond heterozygote levels seemed to recapitulate this very important characteristic of

GNE myopathy histopathology, suggesting that loss of SA alone is sufficient to cause this phenotype.

3.5 Discussion

We have developed a mouse model of GNE myopathy, Gne(−/−)hGNED176V-

TgCmah(−/−), in which oral sialic acid (SA) incorporation into muscles can be visualized through immunostaining. Using this model, we have assessed two dosage paradigms of oral SA, and found that a single, high dose of free SA can effectively be incorporated into the sarcolemmal membrane, while long-term feeding of SA-enriched food cannot. In addition, we have shown that liver-specific GNE gene replacement is more effective than muscle-specific GNE replacement in restoring muscle SA, but that GNE overexpression in liver may lead to pathology. Finally, we have shown that knockdown of endogenous

Gne, but not overexpression of wild-type or mutant GNE protein, recapitulates some histopathology characteristics of GNE myopathy.

In our hands, neither our novel Gne(−/−)hGNED176V-TgCmah(−/−) model nor the original Gne(−/−)hGNED176V-Tg model of GNE myopathy exhibited muscle pathology as was originally reported for this line108. This included minimal change as compared to littermate controls in lifespan, muscle size, histopathology as assessed by H&E and

Congo Red staining, forelimb and hindlimb grip strength, and treadmill performance up to 15 months of age. In addition, Gne(−/−)hGNED176V-Tg did not differ from littermate 119 controls at one year of age in electrophysiology measures of the tibialis anterior (TA).

This was surprising, as the original model was reported to differ from littermate controls in all of these measures beginning at approximately 7 months of age108. Instead, this model showed only enlargement of the heart and spleen as well as hyposialylation in skeletal muscle. This difference from the published literature could be due to a variety of factors, including genetic drift, variances in SA content of chow, or husbandry differences between facilities. Previous knock-in mouse models of GNE myopathy have also lacked a muscle phenotype82, 109, underscoring the need for a robust mouse model with a reproducible muscle-specific pathology.

Interestingly, despite the lack of muscle-specific pathology, we have shown that in some cases our novel Gne(−/−)hGNED176V-TgCmah(−/−) mice develop circulating serum antibody titers to Neu5Gc, while neither Gne(−/−)hGNED176V-Tg nor Cmah(−/−) mice develop appreciable Neu5Gc antibody titers. Thus, it appears that

Gne(−/−)hGNED176V- TgCmah(−/−) mice may have alterations in either incorporation of exogenous SA or immune response to incorporated SA, which would be interesting to explore further, as all humans, being CMAH deficient, also develop anti-Neu5Gc serum antibodies that have the potential to impact muscle disease severity141.

We have also utilized this novel Gne(−/−)hGNED176V-TgCmah(−/−) model to allow us, for the first time, to accurately assess cellular and subcellular localization of oral SA as Neu5Gc, which is not endogenously present in these mice. This new model also is more genetically appropriate as it better mimics the human sialylglycome169. Our data showing differences in SA muscle incorporation between two oral SA dosing

120 regimens and the fact that current preparations of oral SA do not provide significant functional benefits129 highlight the need for careful assessment of the relationship between dosing strategy and SA incorporation into muscle. This novel model will provide an excellent system to efficiently optimize SA delivery to the affected tissue through direct visualization.

As GNE myopathy is a recessive monogenic disease, gene replacement via viral vector is a promising therapeutic avenue. To address this, we have treated our GNE myopathy mouse model with gene therapy vectors that provide the wild-type GNE gene under the control of tissue-specific promoters. We compared GNE directed to muscle, the site of disease pathology, and directed to liver, which is more easily targeted by gene therapy viral vectors. We found that, while liver-directed GNE gene replacement is more effective than muscle-directed GNE gene replacement in restoring muscle SA, GNE overexpression induced liver pathology in all mice tested. Considering that most GNE gene therapy strategies being developed are utilizing a promoter that express in all tissues121, 123, these findings indicate that it may be important to minimize GNE expression in liver in future development of GNE gene therapies.

While loss of SA is generally believed to be the cause of GNE myopathy, there remains some controversy regarding this aspect of the disease, as SA reductions in GNE myopathy muscle are often mild, and do not correlate well with disease severity96, 97, 101,

108. One alternate hypothesis is that the disease-causing mutations in GNE myopathy induce pathogenic protein aggregates, as is the case with other such as superoxide dismutase (SOD), which can form aggregates when mutated in Amyotrophic

121 Lateral Sclerosis (ALS)191. Our studies on overexpression of wild type and mutant GNE proteins, however, show that these proteins do not lead to the formation of inclusion bodies in muscle even when overexpressed; wild type and mutant GNE had similar cytoplasmic localization when overexpressed in the skeletal muscle of Gne(+/−) and

Gne(−/−)hGNED176V-TgCmah(−/−) mice. These proteins differed, however, in multimeric state when analyzed via Western blot; GNEWT existed largely in a dimeric state, while most of GNEDV and virtually all of GNEMT existed in the monomeric state. This is an intriguing finding because the monomeric form is likely not enzymatically active in vivo, as biochemical studies have shown that a multimeric state of GNE is favored when its chemical substrate is present90. In fact, in vitro studies have previously shown that mutations causative of GNE myopathy can affect GNE protein multimerization97, but this is the first such finding in vivo.

Finally, we have shown work indicating that miR shuttle knockdown in the TA

(causing approximately a 34% reduction in Gne expression from Gne(+/−) levels) may recapitulate the Congo Red-positive inclusions seen in GNE myopathy histopathology.

This would represent the first instance of GNE myopathy histopathology seen in the absence of mutant GNE protein, indicating GNE myopathy pathology may be a result of hyposialylation rather than alternative effects of mutant GNE protein that have been suggested by others, for example GNE binding to alpha actinin-1, alpha actinin-2, the collapsin response mediator protein 1 (CRMP-1) or the promyelocytic leukemia zinc finger protein (PLZF)187-189. These data further point to the fact that loss of Gne, perhaps in an inducible deletion model, should be sufficient to produce GNE myopathy-like

122 pathology and may in fact do so in a manner that occurs much more quickly and specifically than is seen with the current mutant transgenic or knock-in GNE myopathy mouse models.

3.6 Acknowledgements

We would like to thank Kristin Heller for her help with electrophysiology, Carlee

Giesige for her help with oral gavage, and Nettie Pyne for her invaluable assistance with the miR shuttle project. We would like to thank Genevieve Faber, Paul Thomas, Ben

Hood, and Haley Guggenheim for assistance with experiments. We would like to thank

Dr. Rui Xu and Deborah Zygmunt for technical assistance. We would like to thank Dr.

Douglas McCarty for kindly providing the LSP promoter, Dr. Scott Harper for kindly providing the U6T6 and rAAV.CMV.eGFP plasmids, Dr. Li Chen for kindly providing

HepG2 cells, Dr. Ichizo Nishino and Dr. Satoru Noguchi for kindly providing the

Gne(−/−)hGNED176V-Tg mouse model, and Dr. Ajit Varki for kindly providing the

Cmah(−/−)mouse model and performing the anti-Neu5Gc serum antibody ELISA assays.

123 Figure 17. Gne(−/−)hGNED176V-Tg mice largely do not differ from

Gne(−/−)hGNED176V-TgCmah(−/−) mice in lifespan, body weight, muscle weight, or organ weight up to 15 months of age. No significant difference in lifespan was found between Gne(−/−)hGNED176V-Tg, Gne(−/−)hGNED176V-TgCmah(−/−), and littermate control mice (A). No significant difference was found between body weights of male (B) and female (C) Gne(−/−)hGNED176V-Tg, Gne(−/−)hGNED176V-TgCmah(−/−), and littermate control mice, apart from the difference between female Gne(−/−)hGNED176V-

Tg and littermate controls at 9 months of age (p<0.01, n=4 for Gne(−/−)hGNED176V-Tg

124 and n=13 for littermate controls) and 12 months of age (p<0.05, n=4 for

Gne(−/−)hGNED176V-Tg and n=12 for littermate controls). No significant difference was found between Gne(−/−)hGNED176V-Tg and Gne(−/−)hGNED176V-TgCmah(−/−)mice at any time point. Weight of tibialis anterior (TA), gastrocnemius, quadriceps, and triceps muscles (D) after normalization to body weight was not significantly different between

Gne(−/−)hGNED176V-Tg, Gne(−/−)hGNED176V-TgCmah(−/−), and littermate control mice.

Normalized heart and spleen weight (E) did not differ between Gne(−/−)hGNED176V-Tg and Gne(−/−)hGNED176V-TgCmah(−/−)mice, but were significantly higher than littermate controls in all cases. * p<0.05, ** p <0.01

125 Figure 18. GNE myopathy mouse model gross motor function study timeline.

Gne(−/−)hGNED176V-Tg, Gne(−/−)hGNED176V-TgCmah(−/−), and littermate control mice were assayed for forelimb and hindlimb grip strength and treadmill endurance from 6 to

15 months of age every 3 months as indicated (At 15 MO, n=4 for Gne(−/−)hGNED176V-

Tg, n=4 for Gne(−/−)hGNED176V-TgCmah(−/−) and n=11 for littermate controls). Upon sacrifice, muscles and organs were harvested, weighted, and processed for analysis.

126 Figure 19. Gne(−/−)hGNED176V-Tg mice do not differ from Gne(−/−)hGNED176V-

TgCmah(−/−) mice in gross motor ability up to 15 months of age. Normalized forelimb and hindlimb grip strength (left) and treadmill time before exhaustion (right) were assessed at 6 month (top), 9 month (top middle), 12 month (bottom middle) and 15 month

(bottom) time points.

127 Gne(−/−)hGNED176V-TgCmah(−/−) mice showed a significant decrease in normalized hindlimb grip strength as compared to Gne(−/−)hGNED176V-Tg mice at 6 months old, and Gne(−/−)hGNED176V-Tg showed a significant increase in normalized forelimb grip strength at 15 months old, but no other changes were seen between any genotypes at any time point. * p<0.05

128 Figure 20. H&E and Congo Red staining of gastrocnemius muscle in

Gne(−/−)hGNED176V-Tg mice, Gne(−/−)hGNED176V-TgCmah(−/−) mice, and littermate controls.

129 3xTg-AD mouse brain (top left), or gastrocnemius muscle from Gne(+/−) (top middle),

Gne(+/−)Cmah(−/−) (top right), Gne(−/−)hGNED176V-Tg (bottom left) or

Gne(−/−)hGNED176V-TgCmah(−/−) (bottom right) mice were stained with H&E (top) or

CongoRed(bottom).Scalebar=50μm.

130 Figure 21. Electrophysiology measures in Gne(−/−)hGNED176V-Tg mice and littermate controls.

131 Absolute force (mN) (A), specific force (mN/mm2) (B), and force during repeated eccentric contractions (expressed as percentage of first contraction) (C) were assessed for the TA of Gne(−/−)hGNED176V-Tg mice or littermate controls. Comparisons were not statistically significant for any measure. Errors are SEM for n=4 animals per condition.

132 Figure 22. Several Gne(−/−)hGNED176V-TgCmah(−/−) but not Gne(−/−)hGNED176V-

Tg mice have increased Neu5Gc antibody serum titers at 15 months of age. Zero of six Cmah(−/−)mice and zero of four Gne(−/−)hGNED176V-Tg mice, but two of five

Gne(−/−)hGNED176V-TgCmah(−/−)mice had Neu5Gc titers above 0.1 µg/ml.

133 Figure 23. Endogenous Neu5Gc is absent in Gne(−/−)hGNED176V-TgCmah(−/−) muscles, but overall SA loss is equivalent in Gne(−/−)hGNED176V-Tg and

Gne(−/−)hGNED176V-TgCmah(−/−)muscles. TA muscles from WT (left)

Gne(−/−)hGNED176V-Tg (middle), or Gne(−/−)hGNED176V-TgCmah(−/−) (right) mice were stained with a highly specific antibody for the Neu5Gc form of sialic acid (top) or with the Peanut Agglutinin (PNA) lectin (bottom), which recognizes subterminal β-1,3-

N-acetylgalactosamine structures that are unmasked in the absence of terminal SA residues. All images are time-matched exposures and images shown are merged with

DAPI, a nuclear marker (blue). Scalebar=50μm.

134 Figure 24. Method to study short-term, high-dose exogenous sialic acid incorporation in a GNE myopathy model. Exogenous Neu5Gc (or Neu5Ac as a control) were administered via oral gavage to Gne(−/−)hGNED176V-TgCmah(−/−) mice, which lack endogenous Neu5Gc. After one week, tissues were isolated and immunostained with an anti-Neu5Gc-specific antibody to determine localization of

Neu5Gc incorporation in tissues.

135 Figure 25. Neu5Gc is incorporated into Gne(−/−)hGNED176V-TgCmah(−/−) muscle tissues after oral gavage. Tibialis anterior (top left), gastrocnemius (top middle), quadriceps (top right), triceps (bottom left), diaphragm (bottom middle), and heart

(bottom right) tissues from Gne(−/−)hGNED176V-TgCmah(−/−) mice gavaged with

Neu5Ac (top) or Neu5Gc (bottom) were stained with an anti-Neu5Gc antibody.

136 Exposures are time-matched within muscle types, and images shown are merged with

DAPI, a nuclear marker (blue). Scalebar=50μm.

137 Figure 26. Neu5Gc is incorporated into several Gne(−/−)hGNED176V-TgCmah(−/−) organs after oral gavage. Stomach (top left), kidney (top middle), lung (top right), pancreas (bottom left), and liver (bottom right) tissues from Gne(−/−)hGNED176V-

TgCmah(−/−) mice gavaged with Neu5Ac (top) or Neu5Gc (bottom) were stained with an anti-Neu5Gc antibody. Exposures are time-matched within muscle types, and images shown are merged with DAPI, a nuclear marker (blue). Scalebar=50μm. 138 Figure 27. Method to study long-term, low-dose exogenous sialic acid incorporation in a GNE myopathy model. Gne(−/−)hGNED176V-TgCmah(−/−) mice, which lack endogenous Neu5Gc, were continuously fed ad libitum with normal food (top) or with food enriched by porcine submaxillary mucus (PSM), which has a high Neu5Gc content.

After ten months, tissues were isolated and immunostained with an anti-Neu5Gc-specific antibody to determine localization of Neu5Gc incorporation in tissues.

139 Figure 28. Neu5Gc is not incorporated into Gne(−/−)hGNED176V-TgCmah(−/−) muscle tissues after feeding of PSM food. Tibialis anterior (top left), gastrocnemius

(top middle), quadriceps (top right), triceps (bottom left), diaphragm (bottom middle), and heart (bottom right) tissues from Gne(−/−)hGNED176V-TgCmah(−/−) mice gavaged with Neu5Ac (top) or Neu5Gc (bottom) were stained with an anti-Neu5Gc antibody.

140 Exposures are time-matched within muscle types, and images shown are merged with

DAPI, a nuclear marker (blue). Scalebar=50μm.

141 Figure 29. Neu5Gc is not incorporated into Gne(−/−)hGNED176V-TgCmah(−/−) organs after feeding of PSM food. Large intestine (top left), kidney (top middle), lung (top right), pancreas (bottom left), and liver (bottom right) tissues from Gne(−/−)hGNED176V-

TgCmah(−/−) mice gavaged with Neu5Ac (top) or Neu5Gc (bottom) were stained with an

142 anti-Neu5Gc antibody. Exposures are time-matched within muscle types, and images shown are merged with DAPI, a nuclear marker (blue). Scalebar=50μm.

143 Figure 30. Constructs for tissue-specific expression of GNE in vivo. The standard rAAV cassette was used to express myc/flag tagged GNE under the control of a muscle- specific MCK promoter (top) or a liver-specific LSP promoter (bottom).

AAV, adeno-associated virus; MCK, muscle creatine kinase; ITR, inverted terminal repeat; SV40, simian virus 40; ORF, open reading frame; LSP, liver-specific promoter;

AMBP, alpha1-microglobulin/bikunin precursor.

144 Figure 31. The LSP promoter is expressed in liver cells in vitro. Liver carcinoma

HepG2 cells were mock transfected (left) or transfected with CMV.GFP (middle) or

LSP.GFP (right) constructs. GFP fluorescence was assessed 48 hours after transfection.

Scalebar=50μm.

145 Inj., injection; IM, intramuscular; IP, intraperitoneal; rAAV, recombinant Adeno- associated virus; vg, vector genomes; LTA, left tibialis anterior; Lgas, left gastrocnemius; kg, kilogram

Table 5. Injection information for expression of tissue-specific GNE in vivo.

Genotype, n number, injection route, rAAV construct, and viral dose are shown for all mice injected with wild-type GNE under the control of a tissue-specific promoter.

146 Figure 32. Method to study muscle- or liver-specific expression of GNE in

Gne(−/−)hGNED176V-Tg and Gne(−/−)hGNED176V-TgCmah(−/−) mice. At 1-2 months of age, Gne(−/−)hGNED176V-Tg and Gne(−/−)hGNED176V-TgCmah(−/−) mice, which exhibit muscle hyposialylation, received unilateral, intramuscular injections of 1E+11 vg/TA and 5E+11 vg/gastrocnemius of rAAVrh74.MCK.GNE or rAAVrh74.LSP.GNE

(top). Alternatively, theses mice received intraperitoneal injections of approximately

2.5E+13 vg/kg of rAAVrh74.MCK.GNE or rAAVrh74.LSP.GNE (bottom). After ten months, muscle tissues and organs were isolated for analysis.

147 Figure 33. qPCR and qRT-PCR analysis for rAAV injection of GNE with tissue- specific promoters. rAAV vector genomes assessed via quantitative polymerase chain reaction (qPCR) and normalized to total genomic DNA (A) and changes in GNE gene expression assessed via qRT-PCR and expressed as dCT (as normalized to 18S) for tibialis anterior, gastrocnemius, and liver (B), expressed as fold change in GNE from uninjected Gne(−/−)hGNED176V-TgCmah(−/−) for liver (C), or expressed as fold change from Gne in uninjected wild type for skeletal muscle (D) are shown for

Gne(−/−)hGNED176V-TgCmah(−/−)mice that were injected intramuscularly with 1E+11 vg/TA and 5E+11 vg/gastrocnemius or intraperitoneally with approximately 2.5E+13 vg/kg of rAAVrh74.MCK.GNE or rAAVrh74.LSP.GNE. For qPCR, comparisons with

Gne(−/−)hGNED176V-TgCmah(−/−) +IP AAV.MCK.GNE were not made for

148 gastrocnemius, as this muscle was uninjected in this condition, or for liver, as this condition received less total virus than equivalent conditions. A dashed line in (C) is shown at a fold change of one, representing no change from GNE expression in

Gne(−/−)hGNED176V-TgCmah(−/−) liver. Dataaremean±SEMfor3-5 mice per group.

* p <0.05, ** p <0.01, *** p <0.001, **** p <0.0001

149 Figure 34. Western blot analysis of GNE protein expression in muscle and liver extracted with varying detergents. Liver (left blot) or quadriceps (right blot) lysates from Gne(−/−)hGNED176V-TgCmah(−/−)mice that were uninjected, injected intraperitoneally with approximately 2.5E+13 vg/kg of rAAVrh74.LSP.GNE (left blot), or intramuscularly with 5E+11 rAAV.MCK.GNE (right blot) were extracted in detergents containing NP-40 (NP), sodium dodecyl sulfate (SDS), or SDS and 4% urea

(Ur) as marked, separated by SDS-PAGE, and immunoblotted with an antibody to

FLAG. Anti-Gapdh blots were done as a control for serum protein loading and transfer.

150 Figure 35. The LSP promoter expresses GNE in liver in vivo. Liver sections from

Gne(−/−)hGNED176V-TgCmah(−/−) mice that were uninjected (top), injected intramuscularly with 5E+11 rAAVrh74.MCK.GNE (middle), or injected intraperitoneally with approximately 2.5E+13 vg/kg of rAAVrh74.LSP.GNE (bottom) were stained with an anti-FLAG antibody. Exposures are time-matched, and images shown are merged with

DAPI, a nuclear marker (blue). Scale bar = 100 μm. 151 Figure 36. Normalized muscle and organ weights of Gne(−/−)hGNED176V-

TgCmah(−/−)mice injected with GNE targeted to muscle or liver. Weight of tibialis anterior, gastrocnemius, quadriceps, and triceps muscles (A) or heart and spleen (B) after normalization to body weight for Gne(−/−)hGNED176V-TgCmah(−/−)mice that were

152 injected intramuscularly with 1E+11 vg/TA and 5E+11 vg/gastrocnemius or intraperitoneally with approximately 2.5E+13 vg/kg of rAAVrh74.MCK.GNE or rAAVrh74.LSP.GNE. Comparisons were not statistically significant at any time point.

153 Figure 37. Intraperitoneal injection of GNE under a liver-specific promoter increases muscle and liver sialic acid in Gne(−/−)hGNED176V-Tg mice. Gastrocnemius muscle (left) or liver (right) from Gne(−/−)hGNED176V-Tg mice that were uninjected

(top), injected intramuscularly with 5E+11 rAAVrh74.MCK.GNE (middle), or injected intraperitoneally with approximately 2.5E+13 vg/kg of rAAVrh74.LSP.GNE (bottom)

154 were stained with an anti-Neu5Gc antibody. Exposures are time-matched, and images shown are merged with DAPI, a nuclear marker (blue). Scalebar=100μm.

155 Figure 38. Intraperitoneal injection of GNE under a liver-specific promoter increases liver sialic acid. Liver tissues from Gne(−/−)hGNED176V-TgCmah(−/−) mice that were uninjected (top), injected intramuscularly with 5E+11 rAAVrh74.MCK.GNE

(middle), or injected intraperitoneally with approximately 2.5E+13 vg/kg of rAAVrh74.LSP.GNE (bottom) were stained with the sialylation-specific lectin SNA (left) or the asialylation-specific lectin PNA (right).

156 Exposures are time-matched, and images shown are merged with DAPI, a nuclear marker

(blue). Scalebar=50μm.

157 Figure 39. Intraperitoneal injection of GNE under a liver-specific promoter increases muscle sialic acid. Gastrocnemius muscle from Gne(−/−)hGNED176V-

TgCmah(−/−) mice that were uninjected (top), injected intramuscularly with 5E+11 rAAVrh74.MCK.GNE (middle), or injected intraperitoneally with approximately

2.5E+13 vg/kg of rAAVrh74.LSP.GNE (bottom) were stained with the sialylation- specific lectin MAA (left) or the asialylation-specific lectin PNA (right).

158 Exposures are time-matched, and images shown are merged with DAPI, a nuclear marker

(blue). Scalebar=50μm.

159 Figure 40. H&E staining of liver and skeletal muscle and Congo Red staining of skeletal muscle in virus-injected Gne(−/−)hGNED176V-TgCmah(−/−)mice. Liver (top) or skeletal muscle (middle/bottom) tissues from Gne(−/−)hGNED176V-TgCmah(−/−) mice that were uninjected (left), injected intramuscularly with 5E+11 rAAVrh74.MCK.GNE

(middle), or injected intraperitoneally with approximately 2.5E+13 vg/kg of rAAVrh74.LSP.GNE (right) were stained with H&E (top/middle) or Congo Red

(bottom). Arrows indicate liver pathology. Scale bar = 50μm.

160 Figure 41. H&E staining of liver in virus-injected Gne(−/−)hGNED176V-TgCmah(−/−) mice. Liver tissues from Gne(−/−)hGNED176V-TgCmah(−/−) mice that were uninjected

(left), injected intramuscularly with 5E+11 rAAVrh74.MCK.GNE (middle), or injected intraperitoneally with approximately 2.5E+13 vg/kg of rAAVrh74.LSP.GNE (right) were stained with H&E. Arrows indicate liver pathology. Scale bar = 25 μm.

161 Figure 42. Immunoprecipitation and western blot analysis of wild-type or mutant

GNE delivered under tissue-specific promoters. Gastrocnemius tissues from Gne(+/−) or

Gne(−/−)hGNED176V-TgCmah(−/−) (left blot, uninjected control only) mice that were uninjected or injected intramuscularly with 5E+11 rAAV.MCK.GNEWT, DV, or MT were immunoprecipitated with FLAG beads, separated by SDS-PAGE, and immunoblotted with an antibody to FLAG.

162 Figure 43. Wild type and mutant GNE under a muscle-specific promoter are expressed in vivo. Gastrocnemius from Gne(+/−) mice (left) or Gne(−/−)hGNED176V- 163 TgCmah(−/−) (right) that were uninjected (top), or injected intramuscularly with 5E+11 rAAVrh74.MCK.GNEWT (top middle), rAAVrh74.MCK.GNEDV (bottom middle), or rAAVrh74.MCK.GNEMT (bottom) were stained with an anti-FLAG antibody. Exposures are time-matched, and images shown are merged with DAPI, a nuclear marker (blue).

Scalebar=50μm.

164 Figure 44. H&E staining of gastrocnemius in mice injected with wild-type or mutant

GNE. Gastrocnemius from Gne(+/−) mice (left) or Gne(−/−)hGNED176V-TgCmah(−/−) 165 (right) that were uninjected (top), or injected intramuscularly with 5E+11 rAAVrh74.MCK.GNEWT (top middle), rAAVrh74.MCK.GNEDV (bottom middle), or rAAVrh74.MCK.GNEMT (bottom) were stained with H&E. Scalebar=50μm.

166 Figure 45. Congo Red staining of gastrocnemius in mice injected with wild-type or mutant GNE. Gastrocnemius from Gne(+/−) mice (left) or Gne(−/−)hGNED176V-

TgCmah(−/−) (right) that were uninjected (top), or injected intramuscularly with 5E+11 167 rAAVrh74.MCK.GNEWT (top middle), rAAVrh74.MCK.GNEDV (bottom middle), or rAAVrh74.MCK.GNEMT (bottom) were stained with Congo Red. Scalebar=50μm.

168 Figure 46. Artificial miR shuttle constructs for knockdown of Gne expression. Four artificial miR shuttle constructs were designed for the purpose of knocking down Gne gene expression. Underlined sequence indicates the mature guide strand after processing.

169 Figure 47. Luciferase expression and qRT-PCR analysis for GNE miR shuttle constructs in vitro. A. The GNE ORF was cloned downstream of luciferase in a plasmid co-expressing Renilla luciferase and co-transfected with GNE miR shuttle constructs 3 or

4 into HEK293 cells. Firefly luciferase expression was assessed and normalized to

Renilla luciferase expression. B. GNE gene expression was assessed by qRT-PCR after mock transfection or transfection of GNE miR shuttle constructs 3 or 4 and normalized to mock transfected control. * p<0.05, ** p <0.01, *** p <0.001

170 Figure 48. In vivo miRNA shuttle study construct map and timeline. A. The standard rAAV cassette was used to dually express GNEmiR4 under the control of the U6 promoter and GFP under the control of the ubiquitous CMV promoter. B. Gne(+/−)mice were injected intramuscularly with 2E+11vg/TA of rAAV.U6.GNEmiR4.CMV.GFP and sacrificed 1, 3, and 6 months post-injection as indicated (n=2 for each timepoint). Upon sacrifice, tibialis anterior muscles were harvested, weighted, and processed for analysis. rAAV, recombinant adeno-associated virus; miR, micro RNA; CMV, cytomegalovirus; eGFP, enhanced green fluorescent protein; ITR, inverted terminal repeat; ORF, open reading frame; SV40, simian virus 40; MO, months old.

171 Figure 49. rAAV.U6.GNEmiR4.CMV.GFP is expressed in vivo. Tibialis anterior from

Gne(+/−) mice that were uninjected (left) or injected intramuscularly with 1E+12vg/TA of rAAVrh74.U6.GNEmiR4.CMV.GFP for six months (right) were assessed for green fluorescence. Exposures are time-matched, and images shown are merged with DAPI, a nuclear marker (blue). Scalebar=50μm.

172 Figure 50. qPCR and qRT-PCR analysis for rAAV delivery of GNE miR shuttle. rAAV vector genomes assessed via quantitative polymerase chain reaction (qPCR) and normalized to total genomic DNA (A) and changes in Gne gene expression assessed via qRT-PCR and expressed as relative to wild-type Gne (B) are shown for tibialis anterior of uninjected wild type or Gne(+/−) mice or Gne(+/−) mice that were injected intramuscularly

173 with 2E+11 vg/TA of rAAVrh74.U6.GNEmiR4.CMV.GFP for 1,3, or 6 months. Data are mean±SEMfor2 mice per group. * p<0.05

174 Figure 51. Normalized muscle weights of Gne(+/−)mice injected with rAAV.U6.GNEmiR4.CMV.GFP. Weight of TA muscles after normalization to body weight for Gne(+/−) mice that were uninjected or injected intramuscularly with 2E+11/TA of rAAV.U6.GNEmiR4.CMV.GFP for one month, three months, or six months.

Comparisons were not statistically significant at any time point.

175 Figure 52. H&E and Congo Red staining of tibialis anterior in mice injected with rAAV.U6.GNEmiR4.CMV.GFP. TA muscles from Gne(+/−) mice that were uninjected

(top) or injected intramuscularly with 2E+11vg/TA of rAAV.U6.GNEmiR4.CMV.GFP 176 for one month (top middle), three months (bottom middle) or six months (bottom) were stained with H&E (left) or Congo Red (right). Arrows indicate Congo Red-positivity.

Scalebar=50μm.

177 Figure 53. Congo Red positivity of tibialis anterior in mice injected with rAAV.U6.GNEmiR4.CMV.GFP. TA muscles from Gne(+/−) mice that were injected intramuscularly with 2E+11vg/TA of rAAV.U6.GNEmiR4.CMV.GFP for six months were stained with Congo Red and visualized with brightfield (left), polarized light

(middle) or rhodamine fluorescence (right) imaging. Arrows indicate the same Congo

Red-positive inclusion in each panel. Scale bar = 25μm.

178 Chapter 4. Conclusions and future directions

We first assessed the utility of N-terminal alpha dystroglycan (αDG-N), a secreted component of the dystrophin-associated glycoprotein (DAG) complex, as a serum biomarker for patients with neuromuscular disease. We developed an ELISA-based assay to assess αDG-N, finding that a serum-immobilized indirect ELISA was more favorable than sandwich or competition ELISAs for this purpose. Utilizing this assay and Western blotting, we found that patients with Duchenne muscular dystrophy (DMD) show a significant decrease in levels of serum αDG-N as compared to otherwise normal patients, likely reflecting a decrease and/or instability of the DAG complex. We further show that serum αDG-N levels are not dependent on age or gender in patient or otherwise normal populations. This reduction as mirrored in a decrease serum αDG-N in mdx Utrn-/-, but not mdx mice, as compared to their wild-type counterparts. Golden retriever muscular dystrophy (GRMD) dogs show a non-significant trend towards decreased serum αDG-N as compared to their unaffected golden retriever (GR) littermates. Finally, we show via immunostaining and Western blot that αDG-N has little to no expression in mouse skeletal muscle, but is strongly expressed in intramuscular peripheral nerves. While this work is a promising avenue for a future clinical biomarker for DMD, several key issues would first need to be addressed: namely, the αDG-N ELISA assay requires additional refinement, and evidence is needed that DAG complex instability is causative, secondary to dystrophin deficiency, of the decrease in αDG-N.

For future development of αDG-N as a biomarker for DMD clinical trials, it would first be important to improve the ELISA assay for αDG-N quantification. The 179 serum-immobilized ELISA with the 2A3 antibody used herein for the quantification of serum αDG-N, while showing favorable intra-assay variability and spike-in recovery, was limited by poor inter-assay variability of approximately 73%. This issue of inter- assay variability likely stems from the high serum dilutions necessary to fit within the assay’snarrowrangeofdetectability,sometimesnecessitatingdifferentserumdilution factors between plates. For future work with αDG-N, development of a sandwich ELISA could circumvent these issues, as this method generally has superior specificity and robustness192. Unfortunately, the paucity of antibodies specific to αDG-N has complicated development of such an assay. Although a sandwich ELISA using the 2A3 antibody with a second αDG-N antibody, 3B4, was originally published as effective in measuring αDG-N139, our work has shown that these antibodies bind to αDG-N in a mutually exclusive manner, precluding their use in a sandwich ELISA. So, creation of a true matched pair of αDG-N antibodies and development of a sandwich ELISA would be critical for future translation of αDG-N as a serum biomarker for DMD.

Additionally, while serumαDG-N is differentially abundant in DMD serum as compared to otherwise normal controls, it is not yet ready for use as a biomarker, partially because it is unclear ifthedecreaseofαDG-N in serum is actually caused by

DAG complex instability in DMD muscle. To address this, future work should address the extenttowhichserumαDG-N levels are returned to baseline when dystrophin, and thus the DAG complex, are therapeutically restored to the muscle membrane. This could be done by utilizing mdx Utrn-/- mice, which have shown a robust reduction in serum

αDG-N as compared to wild type mice. So, untreated mdx Utrn-/- mice could be

180 compared to those treated with a therapy such as exon skipping to restore dystrophin, and thus the DAG complex, to the muscle membrane. This increaseinserumαDG-N could be correlated with restoration of dystrophin and other members of the DAG complex in skeletal muscle as well as rescue of gross motor function in these mice. Taken together, theseresultswouldindicatethat,infact,serumαDG-N levels reflect presence or stability of the DAG complex in the muscle membrane, which is disrupted in DMD pathophysiology. Importantly, this would provide evidence of a link between serum

αDG-N levels, dystrophin restoration, and functional motor correction in this disease.

In addition, further work should be done to assess serum αDG-N levels in GRMD dogs as compared to their GR counterparts. In our hands, GRMD dogs showed a non- significant trend towards a decrease in αDG-N as compared to GR dogs (n=5 per group).

The GRMD model of muscular dystrophy, however, is notably phenotypically heterogeneous, as it is a breed rather than an inbred line, which introduces a host of genetic modifiers19, 20. Thus, an increased sample size, stratification of GRMD dogs by disease severity, or an improved ELISA assay may reveal a significant difference in serum αDG-N levels between these groups. Such a finding would be useful in further validation of αDG-N as a biomarker in αDG-N, bolstering the case for its use as an outcome measure in DMD clinical trials.

Next, we studied pre-existing antibody titers to recombinant adeno-associated virus (rAAV). This is of interest because rAAV is a commonly used gene therapy vector for the delivery of therapeutic transgenes in a variety of human diseases, but pre-existing serum antibodies to viral capsid proteins can greatly inhibit rAAV transduction of tissues.

181 We have assayed serum from patients with DMD, Becker muscular dystrophy (DMD),

Inclusion Body Myositis (IBM), and GNE myopathy (GNE). These were compared to serum from otherwise normal human subjects to determine the extent of pre-existing serum antibodies to rAAVrh74, rAAV1, rAAV2, rAAV6, rAAV8 and rAAV9. In almost all cases, patients with measurable titers to one rAAV serotype showed titers to all other serotypes tested, with average titers to rAAV2 being highest in all instances. 26% of all young normal subjects (<18 years old) had measurable rAAV titers to all serotypes tested, and this percentage increased to almost 50% in adult normal subjects (>18 years old). 50% of all IBM and GNE patients also had antibody titers to all rAAV serotypes, while only 18% of DMD and 0% of BMD patients did.

In the future, it would be useful to extend this work to additional patient populations, as rAAV-based gene therapies are being developed for a host of disorders134,

193-197. Based on our findings, it would be particularly useful to compare these titers in diseases of early onset, such as spinal muscular atrophy (SMA), and diseases of late onset,suchasParkinson’sdisease;rAAV-based gene therapies are currently being developed for both conditions, and our studies would suggest that the treatable patient populationmaybesignificantlysmallerforpatientswithParkinson’sdiseasethanSMA due to pre-existing rAAV antibody titers that increase in amount and preponderance with the age of the patient.

Serum antibody titers specific to Neu5Gc-containing glycans were also assessed in DMD patients; we found that while there was a trend towards an increase in Neu5Gc- containing glycans in DMD patients as compared with otherwise normal patients, none

182 reached statistical significance. Future work may explore the extent to which severity of disease may impact titers to Neu5Gc-containing glycans in DMD patients and whether or not these titers may exacerbate disease pathology.

We have next developed the Gne(−/−)hGNED176V-TgCmah(−/−) model for GNE myopathy that expresses only a single form of sialic acid (SA), Neu5Ac, and thus more accurately models the human SA repertoire. We first compared this model and the original model, Gne(−/−)hGNED176V-Tg, to littermate controls, finding that, with a few minor exceptions, neither shows the originally published phenotype of muscle weakness and GNE myopathy-like histopathology. They do, however, show muscle hyposialylation. We next used the Gne(−/−)hGNED176V-TgCmah(−/−) model to visualize uptake of Neu5Gc SA, which these mice do not endogenously express. We found that short-term, high doses of free SA are able to incorporate into the sarcolemma, while glycoprotein-conjugated SA doses delivered by long-term feeding are not. We next found that liver-directed GNE gene therapy is more effective than muscle-directed GNE gene therapy in restoring SA to skeletal muscle, but may induce liver pathology that is transgene-dependent. Finally, we have shown that knockdown of endogenous, wild-type

Gne, but not overexpression of wild-type or mutant GNE, in a mouse model may recapitulate some of the findings of GNE myopathy.

As oral SA therapies have unfortunately shown little efficacy in the treatment of

GNE myopathy patients129, the Gne(−/−)hGNED176V-TgCmah(−/−) model may prove useful in refining these therapies by allowing direct visualization of oral SA incorporation into target tissues. This will allow comparison of various dosage paradigms, which may

183 vary in dose, frequency, packaging, and/or conjugation. Importantly, this model will allow for assessment of duration of incorporation as well; our studies showed incorporation of high doses of free SA into skeletal muscle one week post-injection, but the duration of this incorporation has yet to be assessed; this could be done by adding longer time points post-gavage to identify how long orally-derived SA persists in skeletal muscle. This question of therapeutic duration would be important in determination of dosage strategy for these patients.

One problem with clinical SA supplementation is that there is a poor understanding in the field of what factors affect oral SA uptake into skeletal muscle cells, including hyposialylation itself. In fact, SA uptake into skeletal muscle is known to involve endocytosis, which has shown to be perturbed in conditions of low SA115, 198, 199.

Our novel GNE myopathy model could be used to address this question by assessing the effects of endogenous sialylation on SA scavenging efficiency in skeletal muscle in vivo.

Specifically, oral SA incorporation into skeletal muscle could be compared in

Gne(−/−)hGNED176V-TgCmah(−/−) and Cmah(−/−) mice; if hyposialylation decreases SA scavenging, then Gne(−/−)hGNED176V-TgCmah(−/−) mice would show less incorporated

SA in skeletal muscle than Cmah(−/−) mice after receiving equivalent oral SA doses. This better understanding of SA scavenging in the disease could inform dosing strategies and potentially reveal new therapeutic targets. Additionally, new connections made between cell surface glycosylation and SA scavenging could, in turn, better inform therapeutic strategiesfordiseasessuchasosteoporosisandAlzheimer’sdisease,where hyposialylation has been implicated200-202.

184 Our GNE gene therapy studies showed that overexpression of GNE in liver after intraperitoneal delivery of approximately 2.5E+13 vector genomes (vg)/kilogram (kg) of rAAV.LSP.GNE induced liver pathology in Gne(−/−)hGNED176V-TgCmah(−/−) mice. This raises important questions for future gene therapies for GNE myopathy, particularly considering that: 1) most GNE gene therapies currently being developed are utilizing a systemic delivery and ubiquitous promoter121, 123, which is likely to result in significant

GNE expression in liver, and 2) rAAV-based, systemic doses for neuromuscular disorders are frequently even higher than those that we tested, reaching 1-2E+14 vg/kg in current clinical trials197, 203. As such, studies to confirm the safety and efficacy of GNE overexpression via gene therapy are particularly important. It would be useful to assess

SA restoration and potential liver toxicity after intravenous administration of GNE under the control of liver-specific, ubiquitous, and muscle-specific promoters over a range of doses in a mouse model of GNE myopathy. These kinds of studies could help to inform the safe translation of GNE gene therapy into the clinic.

Finally, we found that knockdown of endogenous Gne in the tibialis anterior of

Gne(+/−) mice through use of a miRNA shuttle induced Congo Red-positive inclusions reminiscent of GNE myopathy histopathology. Given that the GneM712T/M712T and GneV572L

/ V572L knock-in models of GNE myopathy show no muscle phenotype82, 109, and that the

Gne(−/−)hGNED176V-Tg model did not show a muscle-specific phenotype in our hands, knockdown of endogenous Gne may be an interesting avenue to pursue for the development of a new GNE myopathy mouse model. While Congo Red-positive inclusions within muscle fibers are a hallmark of GNE myopathy histopathology, other

185 aspects of the disease, such as rimmed vacuoles and muscle atrophy, were not seen in this experimental paradigm. This could be due in part to the relatively modest knockdown of

Gne; notably, even with a large, intramuscular viral dose (2E+11 vg/tibialis anterior) and six months of expression, the GNE miR shuttle tested here only achieved a 34% reduction in the tibialis anterior as compared to untreated Gne(+/−) mice. As such, a more robust knockdown of Gne expression could lead to more severe GNE myopathy-like pathology. This idea could be tested either through design and validation of a more efficient miR shuttle, or perhaps more effectively through a mouse model genetically limiting Gne production. As complete knockout of Gne is embryonically lethal in mice, conditional and/or muscle-specific knockout lines could be constructed to restrict Gne production temporally or by tissue. In this model, Gne could be titrated in skeletal muscle to determine the expression level that best mimics GNE myopathy histopathology, including Congo Red-positive inclusions and rimmed vacuoles. Muscle size, gross motor function, and ex vivo electrophysiology could also be assessed in these mice to determine the extent to which more robust Gne knockdown could recapitulate other aspects of GNE myopathy pathophysiology.

As a whole, this work develops several avenues for the assessment and treatment of neuromuscular disorders, including DMD and GNE myopathy. With further development, these tools may prove beneficial for patients with these devastating disorders.

186 Bibliography

[1] Cardamone M, Darras BT, Ryan MM: Inherited myopathies and muscular dystrophies. Seminars in neurology 2008, 28:250-9. [2] Hoffman EP, Brown RH, Jr., Kunkel LM: Dystrophin: the protein product of the Duchenne muscular dystrophy locus. Cell 1987, 51:919-28. [3] Koenig M, Hoffman EP, Bertelson CJ, Monaco AP, Feener C, Kunkel LM: Complete cloning of the Duchenne muscular dystrophy (DMD) cDNA and preliminary genomic organization of the DMD gene in normal and affected individuals. Cell 1987, 50:509-17. [4] Manzur AY, Muntoni F: Diagnosis and new treatments in muscular dystrophies. Postgraduate medical journal 2009, 85:622-30. [5] Wein N, Alfano L, Flanigan KM: Genetics and emerging treatments for Duchenne and Becker muscular dystrophy. Pediatric clinics of North America 2015, 62:723-42. [6] Brooks SV, Faulkner JA: Contractile properties of skeletal muscles from young, adult and aged mice. The Journal of physiology 1988, 404:71-82. [7] Fong PY, Turner PR, Denetclaw WF, Steinhardt RA: Increased activity of calcium leak channels in myotubes of Duchenne human and mdx mouse origin. Science 1990, 250:673-6. [8] Turner PR, Westwood T, Regen CM, Steinhardt RA: Increased protein degradation results from elevated free calcium levels found in muscle from mdx mice. Nature 1988, 335:735-8. [9] Stedman HH, Sweeney HL, Shrager JB, Maguire HC, Panettieri RA, Petrof B, Narusawa M, Leferovich JM, Sladky JT, Kelly AM: The mdx mouse diaphragm reproduces the degenerative changes of Duchenne muscular dystrophy. Nature 1991, 352:536-9. [10] Ervasti JM, Ohlendieck K, Kahl SD, Gaver MG, Campbell KP: Deficiency of a glycoprotein component of the dystrophin complex in dystrophic muscle. Nature 1990, 345:315-9. [11] Ervasti JM, Campbell KP: A role for the dystrophin-glycoprotein complex as a transmembrane linker between laminin and actin. The Journal of cell biology 1993, 122:809-23. [12] Martin PT: Dystroglycan glycosylation and its role in matrix binding in skeletal muscle. Glycobiology 2003, 13:55R-66R. [13] Ervasti JM, Campbell KP: Membrane organization of the dystrophin-glycoprotein complex. Cell 1991, 66:1121-31. [14] Ibraghimov-Beskrovnaya O, Ervasti JM, Leveille CJ, Slaughter CA, Sernett SW, Campbell KP: Primary structure of dystrophin-associated glycoproteins linking dystrophin to the extracellular matrix. Nature 1992, 355:696-702. 187 [15] Yucel N, Chang AC, Day JW, Rosenthal N, Blau HM: Humanizing the mdx mouse model of DMD: the long and the short of it. NPJ Regenerative medicine 2018, 3:4. [16] Helliwell TR, Man NT, Morris GE, Davies KE: The dystrophin-related protein, utrophin, is expressed on the sarcolemma of regenerating human skeletal muscle fibres in dystrophies and inflammatory myopathies. Neuromuscular disorders : NMD 1992, 2:177- 84. [17] Deconinck AE, Rafael JA, Skinner JA, Brown SC, Potter AC, Metzinger L, Watt DJ, Dickson JG, Tinsley JM, Davies KE: Utrophin-dystrophin-deficient mice as a model for Duchenne muscular dystrophy. Cell 1997, 90:717-27. [18] Kornegay JN: The golden retriever model of Duchenne muscular dystrophy. Skeletal muscle 2017, 7:9. [19] Bello L, Kesari A, Gordish-Dressman H, Cnaan A, Morgenroth LP, Punetha J, Duong T, Henricson EK, Pegoraro E, McDonald CM, Hoffman EP, Cooperative International Neuromuscular Research Group I: Genetic modifiers of ambulation in the Cooperative International Neuromuscular Research Group Duchenne Natural History Study. Annals of neurology 2015, 77:684-96. [20] Vieira NM, Elvers I, Alexander MS, Moreira YB, Eran A, Gomes JP, Marshall JL, Karlsson EK, Verjovski-Almeida S, Lindblad-Toh K, Kunkel LM, Zatz M: Jagged 1 Rescues the Duchenne Muscular Dystrophy Phenotype. Cell 2015, 163:1204-13. [21] Scott JM, Li S, Harper SQ, Welikson R, Bourque D, DelloRusso C, Hauschka SD, Chamberlain JS: Viral vectors for gene transfer of micro-, mini-, or full-length dystrophin. Neuromuscular disorders : NMD 2002, 12 Suppl 1:S23-9. [22] Duan D: Dystrophin Gene Replacement and Gene Repair Therapy for Duchenne Muscular Dystrophy in 2016: An Interview. Human gene therapy Clinical development 2016, 27:9-18. [23] Nguyen HH, Jayasinha V, Xia B, Hoyte K, Martin PT: Overexpression of the cytotoxic T cell GalNAc transferase in skeletal muscle inhibits muscular dystrophy in mdx mice. Proceedings of the National Academy of Sciences of the United States of America 2002, 99:5616-21. [24] Xu R, Camboni M, Martin PT: Postnatal overexpression of the CT GalNAc transferase inhibits muscular dystrophy in mdx mice without altering muscle growth or neuromuscular development: evidence for a utrophin-independent mechanism. Neuromuscular disorders : NMD 2007, 17:209-20. [25] Kole R, Krieg AM: Exon skipping therapy for Duchenne muscular dystrophy. Advanced drug delivery reviews 2015, 87:104-7. [26] Syed YY: Eteplirsen: First Global Approval. Drugs 2016, 76:1699-704. [27] Popplewell LJ, Malerba A, Dickson G: Optimizing antisense oligonucleotides using phosphorodiamidate morpholino oligomers. Methods in molecular biology 2012, 867:143-67. [28] Yin H, Lu Q, Wood M: Effective exon skipping and restoration of dystrophin expression by peptide nucleic acid antisense oligonucleotides in mdx mice. Molecular therapy : the journal of the American Society of Gene Therapy 2008, 16:38-45. [29] Malik V, Rodino-Klapac LR, Viollet L, Wall C, King W, Al-Dahhak R, Lewis S, Shilling CJ, Kota J, Serrano-Munuera C, Hayes J, Mahan JD, Campbell KJ, Banwell B, 188 Dasouki M, Watts V, Sivakumar K, Bien-Willner R, Flanigan KM, Sahenk Z, Barohn RJ, Walker CM, Mendell JR: Gentamicin-induced readthrough of stop codons in Duchenne muscular dystrophy. Annals of neurology 2010, 67:771-80. [30] Bajek A, Porowinska D, Kloskowski T, Brzoska E, Ciemerych MA, Drewa T: Cell therapy in Duchenne muscular dystrophy treatment: clinical trials overview. Critical reviews in eukaryotic gene expression 2015, 25:1-11. [31] Benedetti S, Hoshiya H, Tedesco FS: Repair or replace? Exploiting novel gene and cell therapy strategies for muscular dystrophies. The FEBS journal 2013, 280:4263-80. [32] Hathout Y, Seol H, Han MH, Zhang A, Brown KJ, Hoffman EP: Clinical utility of serum biomarkers in Duchenne muscular dystrophy. Clinical proteomics 2016, 13:9. [33] Wilton SD, Fletcher S, Flanigan KM: Dystrophin as a therapeutic biomarker: are we ignoring data from the past? Neuromuscular disorders : NMD 2014, 24:463-6. [34] Vila MC, Klimek MB, Novak JS, Rayavarapu S, Uaesoontrachoon K, Boehler JF, Fiorillo AA, Hogarth MW, Zhang A, Shaughnessy C, Gordish-Dressman H, Burki U, Straub V, Lu QL, Partridge TA, Brown KJ, Hathout Y, van den Anker J, Hoffman EP, Nagaraju K: Elusive sources of variability of dystrophin rescue by exon skipping. Skeletal muscle 2015, 5:44. [35] Califf RM: Scientific Dispute Regarding Accelerated Approval of Sarepta Therapeutics' Eteplirsen (NDA 206488) - Commissioner's Decision. Edited by Research CfDEa. 2016. [36] Hyser CL, Griggs RC, Mendell JR, Polakowska R, Quirk S, Brooke MH, Fenichel GM, Doherty RA: Use of serum creatine kinase, pyruvate kinase, and genetic linkage for carrier detection in Duchenne and Becker dystrophy. Neurology 1987, 37:4-10. [37] Mendell JR, Shilling C, Leslie ND, Flanigan KM, al-Dahhak R, Gastier-Foster J, Kneile K, Dunn DM, Duval B, Aoyagi A, Hamil C, Mahmoud M, Roush K, Bird L, Rankin C, Lilly H, Street N, Chandrasekar R, Weiss RB: Evidence-based path to newborn screening for Duchenne muscular dystrophy. Annals of neurology 2012, 71:304-13. [38] Ozawa E, Hagiwara Y, Yoshida M: Creatine kinase, cell membrane and Duchenne muscular dystrophy. Molecular and cellular biochemistry 1999, 190:143-51. [39] Ferlini A, Flanigan KM, Lochmuller H, Muntoni F, t Hoen PA, McNally E: 204th ENMC International Workshop on Biomarkers in Duchenne Muscular Dystrophy 24-26 January 2014, Naarden, The Netherlands. Neuromuscular disorders : NMD 2015, 25:184- 98. [40] Gasper MC, Gilchrist JM: Creatine kinase: a review of its use in the diagnosis of muscle disease. Medicine and health, Rhode Island 2005, 88:398, 400-4. [41] McNally EM: New approaches in the therapy of cardiomyopathy in muscular dystrophy. Annual review of medicine 2007, 58:75-88. [42] Cynthia Martin F, Hiller M, Spitali P, Oonk S, Dalebout H, Palmblad M, Chaouch A, Guglieri M, Straub V, Lochmuller H, Niks EH, Verschuuren JJ, Aartsma-Rus A, Deelder AM, van der Burgt YE, t Hoen PA: Fibronectin is a serum biomarker for Duchenne muscular dystrophy. Proteomics Clinical applications 2014, 8:269-78. [43] Hathout Y, Marathi RL, Rayavarapu S, Zhang A, Brown KJ, Seol H, Gordish- Dressman H, Cirak S, Bello L, Nagaraju K, Partridge T, Hoffman EP, Takeda S, Mah JK, 189 Henricson E, McDonald C: Discovery of serum protein biomarkers in the mdx mouse model and cross-species comparison to Duchenne muscular dystrophy patients. Human molecular genetics 2014, 23:6458-69. [44] Hathout Y, Brody E, Clemens PR, Cripe L, DeLisle RK, Furlong P, Gordish- Dressman H, Hache L, Henricson E, Hoffman EP, Kobayashi YM, Lorts A, Mah JK, McDonald C, Mehler B, Nelson S, Nikrad M, Singer B, Steele F, Sterling D, Sweeney HL, Williams S, Gold L: Large-scale serum protein biomarker discovery in Duchenne muscular dystrophy. Proceedings of the National Academy of Sciences of the United States of America 2015, 112:7153-8. [45] Nadarajah VD, van Putten M, Chaouch A, Garrood P, Straub V, Lochmuller H, Ginjaar HB, Aartsma-Rus AM, van Ommen GJ, den Dunnen JT, t Hoen PA: Serum matrix metalloproteinase-9 (MMP-9) as a biomarker for monitoring disease progression in Duchenne muscular dystrophy (DMD). Neuromuscular disorders : NMD 2011, 21:569-78. [46] Hirata A, Masuda S, Tamura T, Kai K, Ojima K, Fukase A, Motoyoshi K, Kamakura K, Miyagoe-Suzuki Y, Takeda S: Expression profiling of cytokines and related genes in regenerating skeletal muscle after cardiotoxin injection: a role for osteopontin. The American journal of pathology 2003, 163:203-15. [47] Anaya-Segura MA, Garcia-Martinez FA, Montes-Almanza LA, Diaz BG, Avila- Ramirez G, Alvarez-Maya I, Coral-Vazquez RM, Mondragon-Teran P, Escobar-Cedillo RE, Garcia-Calderon N, Vazquez-Cardenas NA, Garcia S, Lopez-Hernandez LB: Non- Invasive Biomarkers for Duchenne Muscular Dystrophy and Carrier Detection. Molecules 2015, 20:11154-72. [48] Rouillon J, Zocevic A, Leger T, Garcia C, Camadro JM, Udd B, Wong B, Servais L, Voit T, Svinartchouk F: Proteomics profiling of urine reveals specific titin fragments as biomarkers of Duchenne muscular dystrophy. Neuromuscular disorders : NMD 2014, 24:563-73. [49] Zaharieva IT, Calissano M, Scoto M, Preston M, Cirak S, Feng L, Collins J, Kole R, Guglieri M, Straub V, Bushby K, Ferlini A, Morgan JE, Muntoni F: Dystromirs as serum biomarkers for monitoring the disease severity in Duchenne muscular Dystrophy. PloS one 2013, 8:e80263. [50] Matsuzaka Y, Kishi S, Aoki Y, Komaki H, Oya Y, Takeda S, Hashido K: Three novel serum biomarkers, miR-1, miR-133a, and miR-206 for Limb-girdle muscular dystrophy, Facioscapulohumeral muscular dystrophy, and Becker muscular dystrophy. Environmental health and preventive medicine 2014, 19:452-8. [51] Saito F, Saito-Arai Y, Nakamura A, Shimizu T, Matsumura K: Processing and secretion of the N-terminal domain of alpha-dystroglycan in cell culture media. FEBS letters 2008, 582:439-44. [52] Singh J, Itahana Y, Knight-Krajewski S, Kanagawa M, Campbell KP, Bissell MJ, Muschler J: Proteolytic enzymes and altered glycosylation modulate dystroglycan function in carcinoma cells. Cancer research 2004, 64:6152-9. [53] Yoon JH, Xu R, Martin P: A Method to Produce and Purify Full-Length Recombinant Alpha Dystroglycan: Analysis of N- and O-Linked Monosaccharide

190 Composition in CHO Cells with or without LARGE Overexpression. PLoS currents 2013, 5. [54] Hesse C, Johansson I, Mattsson N, Bremell D, Andreasson U, Halim A, Anckarsater R, Blennow K, Anckarsater H, Zetterberg H, Larson G, Hagberg L, Grahn A: The N- terminal domain of alpha-dystroglycan, released as a 38 kDa protein, is increased in cerebrospinal fluid in patients with Lyme neuroborreliosis. Biochemical and biophysical research communications 2011, 412:494-9. [55] Saito F, Saito-Arai Y, Nakamura-Okuma A, Ikeda M, Hagiwara H, Masaki T, Shimizu T, Matsumura K: Secretion of N-terminal domain of alpha-dystroglycan in cerebrospinal fluid. Biochemical and biophysical research communications 2011, 411:365-9. [56] Bozzi M, Morlacchi S, Bigotti MG, Sciandra F, Brancaccio A: Functional diversity of dystroglycan. Matrix biology : journal of the International Society for Matrix Biology 2009, 28:179-87. [57] Daya S, Berns KI: Gene therapy using adeno-associated virus vectors. Clinical microbiology reviews 2008, 21:583-93. [58] Deyle DR, Russell DW: Adeno-associated virus vector integration. Current opinion in molecular therapeutics 2009, 11:442-7. [59] Myers MW, Laughlin CA, Jay FT, Carter BJ: Adenovirus helper function for growth of adeno-associated virus: effect of temperature-sensitive mutations in adenovirus early gene region 2. Journal of virology 1980, 35:65-75. [60] Goncalves MA: Adeno-associated virus: from defective virus to effective vector. Virology journal 2005, 2:43. [61] Ayuso E, Mingozzi F, Bosch F: Production, purification and characterization of adeno-associated vectors. Current gene therapy 2010, 10:423-36. [62] Zincarelli C, Soltys S, Rengo G, Rabinowitz JE: Analysis of AAV serotypes 1-9 mediated gene expression and tropism in mice after systemic injection. Molecular therapy : the journal of the American Society of Gene Therapy 2008, 16:1073-80. [63] Bartel M, Schaffer D, Buning H: Enhancing the Clinical Potential of AAV Vectors by Capsid Engineering to Evade Pre-Existing Immunity. Frontiers in microbiology 2011, 2:204. [64] Mikals K, Nam HJ, Van Vliet K, Vandenberghe LH, Mays LE, McKenna R, Wilson JM, Agbandje-McKenna M: The structure of AAVrh32.33, a novel gene delivery vector. Journal of structural biology 2014, 186:308-17. [65] Adriouch S, Franck E, Drouot L, Bonneau C, Jolinon N, Salvetti A, Boyer O: Improved Immunological Tolerance Following Combination Therapy with CTLA-4/Ig and AAV-Mediated PD-L1/2 Muscle Gene Transfer. Frontiers in microbiology 2011, 2:199. [66] Calcedo R, Wilson JM: Humoral Immune Response to AAV. Frontiers in immunology 2013, 4:341. [67] Chicoine LG, Rodino-Klapac LR, Shao G, Xu R, Bremer WG, Camboni M, Golden B, Montgomery CL, Shontz K, Heller KN, Griffin DA, Lewis S, Coley BD, Walker CM, Clark KR, Sahenk Z, Mendell JR, Martin PT: Vascular delivery of rAAVrh74.MCK.GALGT2 to the gastrocnemius muscle of the rhesus macaque 191 stimulates the expression of dystrophin and laminin alpha2 surrogates. Molecular therapy : the journal of the American Society of Gene Therapy 2014, 22:713-24. [68] Goldstein JA, McNally EM: Mechanisms of muscle weakness in muscular dystrophy. The Journal of general physiology 2010, 136:29-34. [69] Rosenberg AS, Puig M, Nagaraju K, Hoffman EP, Villalta SA, Rao VA, Wakefield LM, Woodcock J: Immune-mediated pathology in Duchenne muscular dystrophy. Science translational medicine 2015, 7:299rv4. [70] Ng R, Banks GB, Hall JK, Muir LA, Ramos JN, Wicki J, Odom GL, Konieczny P, Seto J, Chamberlain JR, Chamberlain JS: Animal models of muscular dystrophy. Progress in molecular biology and translational science 2012, 105:83-111. [71] Stanley P, Cummings RD: Structures Common to Different Glycans. Essentials of Glycobiology. Edited by nd, Varki A, Cummings RD, Esko JD, Freeze HH, Stanley P, Bertozzi CR, Hart GW, Etzler ME. Cold Spring Harbor (NY), 2009. [72] Essentials of Glycobiology. Edited by nd, Varki A, Cummings RD, Esko JD, Freeze HH, Stanley P, Bertozzi CR, Hart GW, Etzler ME. Cold Spring Harbor (NY), 2009. [73] Varki NM, Strobert E, Dick EJ, Jr., Benirschke K, Varki A: Biomedical differences between human and nonhuman hominids: potential roles for uniquely human aspects of sialic acid biology. Annual review of pathology 2011, 6:365-93. [74] Okerblom J, Varki A: Biochemical, Cellular, Physiological, and Pathological Consequences of Human Loss of N-Glycolylneuraminic Acid. Chembiochem : a European journal of chemical biology 2017, 18:1155-71. [75] Chandrasekharan K, Yoon JH, Xu Y, deVries S, Camboni M, Janssen PM, Varki A, Martin PT: A human-specific deletion in mouse Cmah increases disease severity in the mdx model of Duchenne muscular dystrophy. Science translational medicine 2010, 2:42ra54. [76] Nishino I, Carrillo-Carrasco N, Argov Z: GNE myopathy: current update and future therapy. Journal of neurology, neurosurgery, and psychiatry 2015, 86:385-92. [77] Amouri R, Driss A, Murayama K, Kefi M, Nishino I, Hentati F: Allelic heterogeneity of GNE gene mutation in two Tunisian families with autosomal recessive inclusion body myopathy. Neuromuscular disorders : NMD 2005, 15:361-3. [78] Huizing M, Krasnewich DM: Hereditary inclusion body myopathy: a decade of progress. Biochimica et biophysica acta 2009, 1792:881-7. [79] Broccolini A, Ricci E, Cassandrini D, Gliubizzi C, Bruno C, Tonoli E, Silvestri G, Pescatori M, Rodolico C, Sinicropi S, Servidei S, Zara F, Minetti C, Tonali PA, Mirabella M: Novel GNE mutations in Italian families with autosomal recessive hereditary inclusion-body myopathy. Human mutation 2004, 23:632. [80] O'Ferrall EK, Sinnreich M: GNE-Related Myopathy. GeneReviews((R)). Edited by Adam MP, Ardinger HH, Pagon RA, Wallace SE, Bean LJH, Stephens K, Amemiya A. Seattle (WA), 1993. [81] Broccolini A, Gidaro T, De Cristofaro R, Morosetti R, Gliubizzi C, Ricci E, Tonali PA, Mirabella M: Hyposialylation of neprilysin possibly affects its expression and enzymatic activity in hereditary inclusion-body myopathy muscle. Journal of neurochemistry 2008, 105:971-81.

192 [82] Ito M, Sugihara K, Asaka T, Toyama T, Yoshihara T, Furuichi K, Wada T, Asano M: Glycoprotein hyposialylation gives rise to a nephrotic-like syndrome that is prevented by sialic acid administration in GNE V572L point-mutant mice. PloS one 2012, 7:e29873. [83] Argov Z, Yarom R: "Rimmed vacuole myopathy" sparing the quadriceps. A unique disorder in Iranian Jews. Journal of the neurological sciences 1984, 64:33-43. [84] Sivakumar K, Dalakas MC: The spectrum of familial inclusion body myopathies in 13 families and a description of a quadriceps-sparing phenotype in non-Iranian Jews. Neurology 1996, 47:977-84. [85] Hinderlich S, Stasche R, Zeitler R, Reutter W: A bifunctional enzyme catalyzes the first two steps in N-acetylneuraminic acid biosynthesis of rat liver. Purification and characterization of UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase. The Journal of biological chemistry 1997, 272:24313-8. [86] Stasche R, Hinderlich S, Weise C, Effertz K, Lucka L, Moormann P, Reutter W: A bifunctional enzyme catalyzes the first two steps in N-acetylneuraminic acid biosynthesis of rat liver. Molecular cloning and functional expression of UDP-N-acetyl-glucosamine 2-epimerase/N-acetylmannosamine kinase. The Journal of biological chemistry 1997, 272:24319-24. [87] Eisenberg I, Avidan N, Potikha T, Hochner H, Chen M, Olender T, Barash M, Shemesh M, Sadeh M, Grabov-Nardini G, Shmilevich I, Friedmann A, Karpati G, Bradley WG, Baumbach L, Lancet D, Asher EB, Beckmann JS, Argov Z, Mitrani- Rosenbaum S: The UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase gene is mutated in recessive hereditary inclusion body myopathy. Nature genetics 2001, 29:83-7. [88] Reinke SO, Hinderlich S: Prediction of three different isoforms of the human UDP- N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase. FEBS letters 2007, 581:3327-31. [89] Yardeni T, Choekyi T, Jacobs K, Ciccone C, Patzel K, Anikster Y, Gahl WA, Kurochkina N, Huizing M: Identification, tissue distribution, and molecular modeling of novel human isoforms of the key enzyme in sialic acid synthesis, UDP-GlcNAc 2- epimerase/ManNAc kinase. Biochemistry 2011, 50:8914-25. [90] Ghaderi D, Strauss HM, Reinke S, Cirak S, Reutter W, Lucka L, Hinderlich S: Evidence for dynamic interplay of different oligomeric states of UDP-N- acetylglucosamine 2-epimerase/N-acetylmannosamine kinase by biophysical methods. Journal of molecular biology 2007, 369:746-58. [91] Martinez J, Nguyen LD, Hinderlich S, Zimmer R, Tauberger E, Reutter W, Saenger W, Fan H, Moniot S: Crystal structures of N-acetylmannosamine kinase provide insights into enzyme activity and inhibition. The Journal of biological chemistry 2012, 287:13656-65. [92] Tong Y, Tempel W, Nedyalkova L, Mackenzie F, Park HW: Crystal structure of the N-acetylmannosamine kinase domain of GNE. PloS one 2009, 4:e7165. [93] Leroy JG, Seppala R, Huizing M, Dacremont G, De Simpel H, Van Coster RN, Orvisky E, Krasnewich DM, Gahl WA: Dominant inheritance of sialuria, an inborn error of feedback inhibition. American journal of human genetics 2001, 68:1419-27. 193 [94] Weiss P, Tietze F, Gahl WA, Seppala R, Ashwell G: Identification of the metabolic defect in sialuria. The Journal of biological chemistry 1989, 264:17635-6. [95] Krause S, Aleo A, Hinderlich S, Merlini L, Tournev I, Walter MC, Argov Z, Mitrani-Rosenbaum S, Lochmuller H: GNE protein expression and subcellular distribution are unaltered in HIBM. Neurology 2007, 69:655-9. [96] Hinderlich S, Salama I, Eisenberg I, Potikha T, Mantey LR, Yarema KJ, Horstkorte R, Argov Z, Sadeh M, Reutter W, Mitrani-Rosenbaum S: The homozygous M712T mutation of UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase results in reduced enzyme activities but not in altered overall cellular sialylation in hereditary inclusion body myopathy. FEBS letters 2004, 566:105-9. [97] Noguchi S, Keira Y, Murayama K, Ogawa M, Fujita M, Kawahara G, Oya Y, Imazawa M, Goto Y, Hayashi YK, Nonaka I, Nishino I: Reduction of UDP-N- acetylglucosamine 2-epimerase/N-acetylmannosamine kinase activity and sialylation in distal myopathy with rimmed vacuoles. The Journal of biological chemistry 2004, 279:11402-7. [98] Leoyklang P, Class B, Noguchi S, Gahl WA, Carrillo N, Nishino I, Huizing M, Malicdan MC: Quantification of lectin fluorescence in GNE myopathy muscle biopsies. Muscle & nerve 2018. [99] Voermans NC, Guillard M, Doedee R, Lammens M, Huizing M, Padberg GW, Wevers RA, van Engelen BG, Lefeber DJ: Clinical features, lectin staining, and a novel GNE frameshift mutation in hereditary inclusion body myopathy. Clinical neuropathology 2010, 29:71-7. [100] Chan YM, Lee P, Jungles S, Morris G, Cadaoas J, Skrinar A, Vellard M, Kakkis E: Substantial deficiency of free sialic acid in muscles of patients with GNE myopathy and in a mouse model. PloS one 2017, 12:e0173261. [101] Salama I, Hinderlich S, Shlomai Z, Eisenberg I, Krause S, Yarema K, Argov Z, Lochmuller H, Reutter W, Dabby R, Sadeh M, Ben-Bassat H, Mitrani-Rosenbaum S: No overall hyposialylation in hereditary inclusion body myopathy myoblasts carrying the homozygous M712T GNE mutation. Biochemical and biophysical research communications 2005, 328:221-6. [102] Saito F, Tomimitsu H, Arai K, Nakai S, Kanda T, Shimizu T, Mizusawa H, Matsumura K: A Japanese patient with distal myopathy with rimmed vacuoles: missense mutations in the epimerase domain of the UDP-N-acetylglucosamine 2-epimerase/N- acetylmannosamine kinase (GNE) gene accompanied by hyposialylation of skeletal muscle glycoproteins. Neuromuscular disorders : NMD 2004, 14:158-61. [103] Huizing M, Rakocevic G, Sparks SE, Mamali I, Shatunov A, Goldfarb L, Krasnewich D, Gahl WA, Dalakas MC: Hypoglycosylation of alpha-dystroglycan in patients with hereditary IBM due to GNE mutations. Molecular genetics and metabolism 2004, 81:196-202. [104] Ricci E, Broccolini A, Gidaro T, Morosetti R, Gliubizzi C, Frusciante R, Di Lella GM, Tonali PA, Mirabella M: NCAM is hyposialylated in hereditary inclusion body myopathy due to GNE mutations. Neurology 2006, 66:755-8. [105] Tajima Y, Uyama E, Go S, Sato C, Tao N, Kotani M, Hino H, Suzuki A, Sanai Y, Kitajima K, Sakuraba H: Distal myopathy with rimmed vacuoles: impaired O-glycan 194 formation in muscular glycoproteins. The American journal of pathology 2005, 166:1121-30. [106] Broccolini A, Gliubizzi C, Pavoni E, Gidaro T, Morosetti R, Sciandra F, Giardina B, Tonali P, Ricci E, Brancaccio A, Mirabella M: alpha-Dystroglycan does not play a major pathogenic role in autosomal recessive hereditary inclusion-body myopathy. Neuromuscular disorders : NMD 2005, 15:177-84. [107] Schwarzkopf M, Knobeloch KP, Rohde E, Hinderlich S, Wiechens N, Lucka L, Horak I, Reutter W, Horstkorte R: Sialylation is essential for early development in mice. Proceedings of the National Academy of Sciences of the United States of America 2002, 99:5267-70. [108] Malicdan MC, Noguchi S, Nonaka I, Hayashi YK, Nishino I: A Gne knockout mouse expressing human GNE D176V mutation develops features similar to distal myopathy with rimmed vacuoles or hereditary inclusion body myopathy. Human molecular genetics 2007, 16:2669-82. [109] Galeano B, Klootwijk R, Manoli I, Sun M, Ciccone C, Darvish D, Starost MF, Zerfas PM, Hoffmann VJ, Hoogstraten-Miller S, Krasnewich DM, Gahl WA, Huizing M: Mutation in the key enzyme of sialic acid biosynthesis causes severe glomerular proteinuria and is rescued by N-acetylmannosamine. The Journal of clinical investigation 2007, 117:1585-94. [110] Malicdan MC, Noguchi S, Hayashi YK, Nishino I: Muscle weakness correlates with muscle atrophy and precedes the development of inclusion body or rimmed vacuoles in the mouse model of DMRV/hIBM. Physiological genomics 2008, 35:106-15. [111] Yan C, Ikezoe K, Nonaka I: Apoptotic muscle fiber degeneration in distal myopathy with rimmed vacuoles. Acta neuropathologica 2001, 101:9-16. [112] Amsili S, Shlomai Z, Levitzki R, Krause S, Lochmuller H, Ben-Bassat H, Mitrani- Rosenbaum S: Characterization of hereditary inclusion body myopathy myoblasts: possible primary impairment of apoptotic events. Cell death and differentiation 2007, 14:1916-24. [113] Broccolini A, Mirabella M: Hereditary inclusion-body myopathies. Biochimica et biophysica acta 2015, 1852:644-50. [114] Zhang YW, Thompson R, Zhang H, Xu H: APP processing in Alzheimer's disease. Molecular brain 2011, 4:3. [115] Bosch-Morato M, Iriondo C, Guivernau B, Valls-Comamala V, Vidal N, Olive M, Querfurth H, Munoz FJ: Increased amyloid beta-peptide uptake in skeletal muscle is induced by hyposialylation and may account for apoptosis in GNE myopathy. Oncotarget 2016, 7:13354-71. [116] Li H, Chen Q, Liu F, Zhang X, Li W, Liu S, Zhao Y, Gong Y, Yan C: Unfolded protein response and activated degradative pathways regulation in GNE myopathy. PloS one 2013, 8:e58116. [117] Fischer C, Kleinschnitz K, Wrede A, Muth I, Kruse N, Nishino I, Schmidt J: Cell stress molecules in the skeletal muscle of GNE myopathy. BMC neurology 2013, 13:24. [118] Cho A, Christine M, Malicdan V, Miyakawa M, Nonaka I, Nishino I, Noguchi S: Sialic acid deficiency is associated with oxidative stress leading to muscle atrophy and weakness in GNE myopathy. Human molecular genetics 2017, 26:3081-93. 195 [119] Singh R, Arya R: GNE Myopathy and Cell Apoptosis: A Comparative Mutation Analysis. Molecular neurobiology 2016, 53:3088-101. [120] Sparks S, Rakocevic G, Joe G, Manoli I, Shrader J, Harris-Love M, Sonies B, Ciccone C, Dorward H, Krasnewich D, Huizing M, Dalakas MC, Gahl WA: Intravenous immune globulin in hereditary inclusion body myopathy: a pilot study. BMC neurology 2007, 7:3. [121] Nemunaitis G, Jay CM, Maples PB, Gahl WA, Huizing M, Yardeni T, Tong AW, Phadke AP, Pappen BO, Bedell C, Allen H, Hernandez C, Templeton NS, Kuhn J, Senzer N, Nemunaitis J: Hereditary inclusion body myopathy: single patient response to intravenous dosing of GNE gene lipoplex. Human gene therapy 2011, 22:1331-41. [122] Tal-Goldberg T, Lorain S, Mitrani-Rosenbaum S: Correction of the Middle Eastern M712T mutation causing GNE myopathy by trans-splicing. Neuromolecular medicine 2014, 16:322-31. [123] Mitrani-Rosenbaum S, Yakovlev L, Becker Cohen M, Telem M, Elbaz M, Yanay N, Yotvat H, Ben Shlomo U, Harazi A, Fellig Y, Argov Z, Sela I: Sustained expression and safety of human GNE in normal mice after gene transfer based on AAV8 systemic delivery. Neuromuscular disorders : NMD 2012, 22:1015-24. [124] Keppler OT, Hinderlich S, Langner J, Schwartz-Albiez R, Reutter W, Pawlita M: UDP-GlcNAc 2-epimerase: a regulator of cell surface sialylation. Science 1999, 284:1372-6. [125] Niethamer TK, Yardeni T, Leoyklang P, Ciccone C, Astiz-Martinez A, Jacobs K, Dorward HM, Zerfas PM, Gahl WA, Huizing M: Oral monosaccharide therapies to reverse renal and muscle hyposialylation in a mouse model of GNE myopathy. Molecular genetics and metabolism 2012, 107:748-55. [126] Malicdan MC, Noguchi S, Hayashi YK, Nonaka I, Nishino I: Prophylactic treatment with sialic acid metabolites precludes the development of the myopathic phenotype in the DMRV-hIBM mouse model. Nature medicine 2009, 15:690-5. [127] Xu X, Wang AQ, Latham LL, Celeste F, Ciccone C, Malicdan MC, Goldspiel B, Terse P, Cradock J, Yang N, Yorke S, McKew JC, Gahl WA, Huizing M, Carrillo N: Safety, pharmacokinetics and sialic acid production after oral administration of N- acetylmannosamine (ManNAc) to subjects with GNE myopathy. Molecular genetics and metabolism 2017, 122:126-34. [128] Argov Z, Caraco Y, Lau H, Pestronk A, Shieh PB, Skrinar A, Koutsoukos T, Ahmed R, Martinisi J, Kakkis E: Aceneuramic Acid Extended Release Administration Maintains Upper Limb Muscle Strength in a 48-week Study of Subjects with GNE Myopathy: Results from a Phase 2, Randomized, Controlled Study. Journal of neuromuscular diseases 2016, 3:49-66. [129] Martins R: Ultragenyx Announces Top-Line Results from Phase 3 Study of Ace- ER in GNE Myopathy. 2017. [130] Nowak KJ, Davies KE: Duchenne muscular dystrophy and dystrophin: pathogenesis and opportunities for treatment. EMBO reports 2004, 5:872-6. [131] Ervasti JM: Dystrophin, its interactions with other proteins, and implications for muscular dystrophy. Biochimica et biophysica acta 2007, 1772:108-17.

196 [132] Ohlendieck K, Matsumura K, Ionasescu VV, Towbin JA, Bosch EP, Weinstein SL, Sernett SW, Campbell KP: Duchenne muscular dystrophy: deficiency of dystrophin- associated proteins in the sarcolemma. Neurology 1993, 43:795-800. [133] Kanagawa M, Saito F, Kunz S, Yoshida-Moriguchi T, Barresi R, Kobayashi YM, Muschler J, Dumanski JP, Michele DE, Oldstone MB, Campbell KP: Molecular recognition by LARGE is essential for expression of functional dystroglycan. Cell 2004, 117:953-64. [134] Mendell JR, Sahenk Z, Malik V, Gomez AM, Flanigan KM, Lowes LP, Alfano LN, Berry K, Meadows E, Lewis S, Braun L, Shontz K, Rouhana M, Clark KR, Rosales XQ, Al-Zaidy S, Govoni A, Rodino-Klapac LR, Hogan MJ, Kaspar BK: A phase 1/2a follistatin gene therapy trial for becker muscular dystrophy. Molecular therapy : the journal of the American Society of Gene Therapy 2015, 23:192-201. [135] Bowles DE, McPhee SW, Li C, Gray SJ, Samulski JJ, Camp AS, Li J, Wang B, Monahan PE, Rabinowitz JE, Grieger JC, Govindasamy L, Agbandje-McKenna M, Xiao X, Samulski RJ: Phase 1 gene therapy for Duchenne muscular dystrophy using a translational optimized AAV vector. Molecular therapy : the journal of the American Society of Gene Therapy 2012, 20:443-55. [136] Varki A: Multiple changes in sialic acid biology during human evolution. Glycoconjugate journal 2009, 26:231-45. [137] Yoon JH, Chandrasekharan K, Xu R, Glass M, Singhal N, Martin PT: The synaptic CT carbohydrate modulates binding and expression of extracellular matrix proteins in skeletal muscle: Partial dependence on utrophin. Molecular and cellular neurosciences 2009, 41:448-63. [138] Flanigan KM, Dunn DM, von Niederhausern A, Soltanzadeh P, Gappmaier E, Howard MT, Sampson JB, Mendell JR, Wall C, King WM, Pestronk A, Florence JM, Connolly AM, Mathews KD, Stephan CM, Laubenthal KS, Wong BL, Morehart PJ, Meyer A, Finkel RS, Bonnemann CG, Medne L, Day JW, Dalton JC, Margolis MK, Hinton VJ, United Dystrophinopathy Project C, Weiss RB: Mutational spectrum of DMD mutations in dystrophinopathy patients: application of modern diagnostic techniques to a large cohort. Human mutation 2009, 30:1657-66. [139] Heng S, Vollenhoven B, Rombauts LJ, Nie G: A High-Throughput Assay for the Detection of alpha-Dystroglycan N-Terminus in Human Uterine Fluid to Determine Uterine Receptivity. Journal of biomolecular screening 2016, 21:408-13. [140] Chandraskeharan K, Martin PT: Embryonic overexpression of Galgt2 inhibits skeletal muscle growth via activation of myostatin signaling. Muscle & nerve 2009, 39:25-41. [141] Padler-Karavani V, Yu H, Cao H, Chokhawala H, Karp F, Varki N, Chen X, Varki A: Diversity in specificity, abundance, and composition of anti-Neu5Gc antibodies in normal humans: potential implications for disease. Glycobiology 2008, 18:818-30. [142] Needham M, Mastaglia FL: Sporadic inclusion body myositis: a continuing puzzle. Neuromuscular disorders : NMD 2008, 18:6-16. [143] Sicinski P, Geng Y, Ryder-Cook AS, Barnard EA, Darlison MG, Barnard PJ: The molecular basis of muscular dystrophy in the mdx mouse: a point mutation. Science 1989, 244:1578-80. 197 [144] Tinsley JM, Potter AC, Phelps SR, Fisher R, Trickett JI, Davies KE: Amelioration of the dystrophic phenotype of mdx mice using a truncated utrophin transgene. Nature 1996, 384:349-53. [145] Rafael JA, Tinsley JM, Potter AC, Deconinck AE, Davies KE: Skeletal muscle- specific expression of a utrophin transgene rescues utrophin-dystrophin deficient mice. Nature genetics 1998, 19:79-82. [146] Grady RM, Teng H, Nichol MC, Cunningham JC, Wilkinson RS, Sanes JR: Skeletal and cardiac myopathies in mice lacking utrophin and dystrophin: a model for Duchenne muscular dystrophy. Cell 1997, 90:729-38. [147] Jarad G, Miner JH: The Pax3-Cre transgene exhibits a rostrocaudal gradient of expression in the skeletal muscle lineage. Genesis 2009, 47:1-6. [148] Reverberi R: The statistical analysis of immunohaematological data. Blood transfusion = Trasfusione del sangue 2008, 6:37-45. [149] Chicoine LG, Montgomery CL, Bremer WG, Shontz KM, Griffin DA, Heller KN, Lewis S, Malik V, Grose WE, Shilling CJ, Campbell KJ, Preston TJ, Coley BD, Martin PT, Walker CM, Clark KR, Sahenk Z, Mendell JR, Rodino-Klapac LR: Plasmapheresis eliminates the negative impact of AAV antibodies on microdystrophin gene expression following vascular delivery. Molecular therapy : the journal of the American Society of Gene Therapy 2014, 22:338-47. [150] Ohlendieck K, Ervasti JM, Matsumura K, Kahl SD, Leveille CJ, Campbell KP: Dystrophin-related protein is localized to neuromuscular junctions of adult skeletal muscle. Neuron 1991, 7:499-508. [151] Matsumura K, Ervasti JM, Ohlendieck K, Kahl SD, Campbell KP: Association of dystrophin-related protein with dystrophin-associated proteins in mdx mouse muscle. Nature 1992, 360:588-91. [152] van Deutekom JC, Janson AA, Ginjaar IB, Frankhuizen WS, Aartsma-Rus A, Bremmer-Bout M, den Dunnen JT, Koop K, van der Kooi AJ, Goemans NM, de Kimpe SJ, Ekhart PF, Venneker EH, Platenburg GJ, Verschuuren JJ, van Ommen GJ: Local dystrophin restoration with antisense PRO051. The New England journal of medicine 2007, 357:2677-86. [153] Goemans NM, Tulinius M, van den Akker JT, Burm BE, Ekhart PF, Heuvelmans N, Holling T, Janson AA, Platenburg GJ, Sipkens JA, Sitsen JM, Aartsma-Rus A, van Ommen GJ, Buyse G, Darin N, Verschuuren JJ, Campion GV, de Kimpe SJ, van Deutekom JC: Systemic administration of PRO051 in Duchenne's muscular dystrophy. The New England journal of medicine 2011, 364:1513-22. [154] Mendell JR, Rodino-Klapac LR, Sahenk Z, Roush K, Bird L, Lowes LP, Alfano L, Gomez AM, Lewis S, Kota J, Malik V, Shontz K, Walker CM, Flanigan KM, Corridore M, Kean JR, Allen HD, Shilling C, Melia KR, Sazani P, Saoud JB, Kaye EM, Eteplirsen Study G: Eteplirsen for the treatment of Duchenne muscular dystrophy. Annals of neurology 2013, 74:637-47. [155] Cirak S, Feng L, Anthony K, Arechavala-Gomeza V, Torelli S, Sewry C, Morgan JE, Muntoni F: Restoration of the dystrophin-associated glycoprotein complex after exon skipping therapy in Duchenne muscular dystrophy. Molecular therapy : the journal of the American Society of Gene Therapy 2012, 20:462-7. 198 [156] Finkel RS, Flanigan KM, Wong B, Bonnemann C, Sampson J, Sweeney HL, Reha A, Northcutt VJ, Elfring G, Barth J, Peltz SW: Phase 2a study of -mediated dystrophin production in patients with Duchenne muscular dystrophy. PloS one 2013, 8:e81302. [157] Rezniczek GA, Konieczny P, Nikolic B, Reipert S, Schneller D, Abrahamsberg C, Davies KE, Winder SJ, Wiche G: Plectin 1f scaffolding at the sarcolemma of dystrophic (mdx) muscle fibers through multiple interactions with beta-dystroglycan. The Journal of cell biology 2007, 176:965-77. [158] McMillan JR, Akiyama M, Rouan F, Mellerio JE, Lane EB, Leigh IM, Owaribe K, Wiche G, Fujii N, Uitto J, Eady RA, Shimizu H: Plectin defects in epidermolysis bullosa simplex with muscular dystrophy. Muscle & nerve 2007, 35:24-35. [159] Mendell JR, Rodino-Klapac LR, Rosales XQ, Coley BD, Galloway G, Lewis S, Malik V, Shilling C, Byrne BJ, Conlon T, Campbell KJ, Bremer WG, Taylor LE, Flanigan KM, Gastier-Foster JM, Astbury C, Kota J, Sahenk Z, Walker CM, Clark KR: Sustained alpha-sarcoglycan gene expression after gene transfer in limb-girdle muscular dystrophy, type 2D. Annals of neurology 2010, 68:629-38. [160] Calcedo R, Vandenberghe LH, Gao G, Lin J, Wilson JM: Worldwide epidemiology of neutralizing antibodies to adeno-associated viruses. The Journal of infectious diseases 2009, 199:381-90. [161] Calcedo R, Wilson JM: AAV Natural Infection Induces Broad Cross-Neutralizing Antibody Responses to Multiple AAV Serotypes in Chimpanzees. Human gene therapy Clinical development 2016, 27:79-82. [162] Flotte TR, Trapnell BC, Humphries M, Carey B, Calcedo R, Rouhani F, Campbell- Thompson M, Yachnis AT, Sandhaus RA, McElvaney NG, Mueller C, Messina LM, Wilson JM, Brantly M, Knop DR, Ye GJ, Chulay JD: Phase 2 clinical trial of a recombinant adeno-associated viral vector expressing alpha1-antitrypsin: interim results. Human gene therapy 2011, 22:1239-47. [163] Thwaite R, Pages G, Chillon M, Bosch A: AAVrh.10 immunogenicity in mice and humans. Relevance of antibody cross-reactivity in human gene therapy. Gene therapy 2015, 22:196-201. [164] Riviere C, Danos O, Douar AM: Long-term expression and repeated administration of AAV type 1, 2 and 5 vectors in skeletal muscle of immunocompetent adult mice. Gene therapy 2006, 13:1300-8. [165] Crowe KE, Shao G, Flanigan KM, Martin PT: N-terminal alpha Dystroglycan (alphaDG-N): A Potential Serum Biomarker for Duchenne Muscular Dystrophy. Journal of neuromuscular diseases 2016, 3:247-60. [166] Zygmunt DA, Crowe KE, Flanigan KM, Martin PT: Comparison of Serum rAAV Serotype-Specific Antibodies in Patients with Duchenne Muscular Dystrophy, Becker Muscular Dystrophy, Inclusion Body Myositis, or GNE Myopathy. Human gene therapy 2017, 28:737-46. [167] Varki A, Schauer R: Sialic Acids. Essentials of Glycobiology. Edited by nd, Varki A, Cummings RD, Esko JD, Freeze HH, Stanley P, Bertozzi CR, Hart GW, Etzler ME. Cold Spring Harbor (NY), 2009.

199 [168] Yonekawa T, Malicdan MC, Cho A, Hayashi YK, Nonaka I, Mine T, Yamamoto T, Nishino I, Noguchi S: Sialyllactose ameliorates myopathic phenotypes in symptomatic GNE myopathy model mice. Brain : a journal of neurology 2014, 137:2670-9. [169] Varki A: Loss of N-glycolylneuraminic acid in humans: Mechanisms, consequences, and implications for hominid evolution. American journal of physical anthropology 2001, Suppl 33:54-69. [170] Naito Y, Takematsu H, Koyama S, Miyake S, Yamamoto H, Fujinawa R, Sugai M, Okuno Y, Tsujimoto G, Yamaji T, Hashimoto Y, Itohara S, Kawasaki T, Suzuki A, Kozutsumi Y: Germinal center marker GL7 probes activation-dependent repression of N- glycolylneuraminic acid, a sialic acid species involved in the negative modulation of B- cell activation. Molecular and cellular biology 2007, 27:3008-22. [171] Diaz SL, Padler-Karavani V, Ghaderi D, Hurtado-Ziola N, Yu H, Chen X, Brinkman-Van der Linden EC, Varki A, Varki NM: Sensitive and specific detection of the non-human sialic Acid N-glycolylneuraminic acid in human tissues and biotherapeutic products. PloS one 2009, 4:e4241. [172] Hedlund M, Tangvoranuntakul P, Takematsu H, Long JM, Housley GD, Kozutsumi Y, Suzuki A, Wynshaw-Boris A, Ryan AF, Gallo RL, Varki N, Varki A: N- glycolylneuraminic acid deficiency in mice: implications for human biology and evolution. Molecular and cellular biology 2007, 27:4340-6. [173] Hakim CH, Li D, Duan D: Monitoring murine skeletal muscle function for muscle gene therapy. Methods in molecular biology 2011, 709:75-89. [174] Singhal N, Martin PT: A role for Galgt1 in skeletal muscle regeneration. Skeletal muscle 2015, 5:3. [175] Harper SQ: RNA interference techniques. New York: Humana Press, 2011. [176] Xiao X, Li J, Samulski RJ: Production of high-titer recombinant adeno-associated virus vectors in the absence of helper adenovirus. Journal of virology 1998, 72:2224-32. [177] Zolotukhin S, Byrne BJ, Mason E, Zolotukhin I, Potter M, Chesnut K, Summerford C, Samulski RJ, Muzyczka N: Recombinant adeno-associated virus purification using novel methods improves infectious titer and yield. Gene therapy 1999, 6:973-85. [178] Livak KJ, Schmittgen TD: Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 2001, 25:402-8. [179] Wallace LM, Liu J, Domire JS, Garwick-Coppens SE, Guckes SM, Mendell JR, Flanigan KM, Harper SQ: RNA interference inhibits DUX4-induced muscle toxicity in vivo: implications for a targeted FSHD therapy. Molecular therapy : the journal of the American Society of Gene Therapy 2012, 20:1417-23. [180] Swallow JG, Koteja P, Carter PA, Garland T, Jr.: Food consumption and body composition in mice selected for high wheel-running activity. Journal of comparative physiology B, Biochemical, systemic, and environmental physiology 2001, 171:651-9. [181] Hayashi Y, Mori Y, Janssen OE, Sunthornthepvarakul T, Weiss RE, Takeda K, Weinberg M, Seo H, Bell GI, Refetoff S: Human thyroxine-binding globulin gene: complete sequence and transcriptional regulation. Molecular endocrinology 1993, 7:1049-60. [182] Bish LT, Sleeper MM, Forbes SC, Morine KJ, Reynolds C, Singletary GE, Trafny D, Pham J, Bogan J, Kornegay JN, Vandenborne K, Walter GA, Sweeney HL: Long-term 200 systemic myostatin inhibition via liver-targeted gene transfer in golden retriever muscular dystrophy. Human gene therapy 2011, 22:1499-509. [183] Gao GP, Alvira MR, Wang L, Calcedo R, Johnston J, Wilson JM: Novel adeno- associated viruses from rhesus monkeys as vectors for human gene therapy. Proceedings of the National Academy of Sciences of the United States of America 2002, 99:11854-9. [184] Sternberg EA, Spizz G, Perry WM, Vizard D, Weil T, Olson EN: Identification of upstream and intragenic regulatory elements that confer cell-type-restricted and differentiation-specific expression on the muscle creatine kinase gene. Molecular and cellular biology 1988, 8:2896-909. [185] Donoviel DB, Shield MA, Buskin JN, Haugen HS, Clegg CH, Hauschka SD: Analysis of muscle creatine kinase gene regulatory elements in skeletal and cardiac muscles of transgenic mice. Molecular and cellular biology 1996, 16:1649-58. [186] Rogerieux F, Belaise M, Terzidis-Trabelsi H, Greffard A, Pilatte Y, Lambre CR: Determination of the sialic acid linkage specificity of sialidases using lectins in a solid phase assay. Analytical biochemistry 1993, 211:200-4. [187] Weidemann W, Stelzl U, Lisewski U, Bork K, Wanker EE, Hinderlich S, Horstkorte R: The collapsin response mediator protein 1 (CRMP-1) and the promyelocytic leukemia zinc finger protein (PLZF) bind to UDP-N-acetylglucosamine 2- epimerase/N-acetylmannosamine kinase (GNE), the key enzyme of sialic acid biosynthesis. FEBS letters 2006, 580:6649-54. [188] Amsili S, Zer H, Hinderlich S, Krause S, Becker-Cohen M, MacArthur DG, North KN, Mitrani-Rosenbaum S: UDP-N-acetylglucosamine 2-epimerase/N- acetylmannosamine kinase (GNE) binds to alpha-actinin 1: novel pathways in skeletal muscle? PloS one 2008, 3:e2477. [189] Harazi A, Becker-Cohen M, Zer H, Moshel O, Hinderlich S, Mitrani-Rosenbaum S: The Interaction of UDP-N-Acetylglucosamine 2-Epimerase/N-Acetylmannosamine Kinase (GNE) and Alpha-Actinin 2 Is Altered in GNE Myopathy M743T Mutant. Molecular neurobiology 2017, 54:2928-38. [190] Hanisch F, Weidemann W, Grossmann M, Joshi PR, Holzhausen HJ, Stoltenburg G, Weis J, Zierz S, Horstkorte R: Sialylation and muscle performance: sialic acid is a marker of muscle ageing. PloS one 2013, 8:e80520. [191] Banci L, Bertini I, Boca M, Girotto S, Martinelli M, Valentine JS, Vieru M: SOD1 and amyotrophic lateral sclerosis: mutations and oligomerization. PloS one 2008, 3:e1677. [192] Cox KL, Devanarayan V, Kriauciunas A, Manetta J, Montrose C, Sittampalam S: Immunoassay Methods. Assay Guidance Manual. Edited by Sittampalam GS, Coussens NP, Brimacombe K, Grossman A, Arkin M, Auld D, Austin C, Baell J, Bejcek B, Chung TDY, Dahlin JL, Devanaryan V, Foley TL, Glicksman M, Hall MD, Hass JV, Inglese J, Iversen PW, Kahl SD, Kales SC, Lal-Nag M, Li Z, McGee J, McManus O, Riss T, Trask OJ, Jr., Weidner JR, Xia M, Xu X. Bethesda (MD), 2004. [193] Spencer HT, Riley BE, Doering CB: State of the art: gene therapy of haemophilia. Haemophilia : the official journal of the World Federation of Hemophilia 2016, 22 Suppl 5:66-71.

201 [194] Hoyng SA, de Winter F, Tannemaat MR, Blits B, Malessy MJ, Verhaagen J: Gene therapy and peripheral nerve repair: a perspective. Frontiers in molecular neuroscience 2015, 8:32. [195] Kalia LV, Kalia SK, Lang AE: Disease-modifying strategies for Parkinson's disease. Movement disorders : official journal of the Movement Disorder Society 2015, 30:1442-50. [196] Rincon MY, VandenDriessche T, Chuah MK: Gene therapy for cardiovascular disease: advances in vector development, targeting, and delivery for clinical translation. Cardiovascular research 2015, 108:4-20. [197] Mendell JR, Al-Zaidy S, Shell R, Arnold WD, Rodino-Klapac LR, Prior TW, Lowes L, Alfano L, Berry K, Church K, Kissel JT, Nagendran S, L'Italien J, Sproule DM, Wells C, Cardenas JA, Heitzer MD, Kaspar A, Corcoran S, Braun L, Likhite S, Miranda C, Meyer K, Foust KD, Burghes AHM, Kaspar BK: Single-Dose Gene-Replacement Therapy for Spinal Muscular Atrophy. The New England journal of medicine 2017, 377:1713-22. [198] Videira PA, Amado IF, Crespo HJ, Alguero MC, Dall'Olio F, Cabral MG, Trindade H: Surface alpha 2-3- and alpha 2-6-sialylation of human monocytes and derived dendritic cells and its influence on endocytosis. Glycoconjugate journal 2008, 25:259-68. [199] Morales CR: Role of sialic acid in the endocytosis of prosaposin by the nonciliated cells of the rat efferent ducts. Molecular reproduction and development 1998, 51:156-66. [200] Susheela AK, Das TK, Khurana JS, Jayaswal A, Dave PK: Circulating levels of sialic acid and glycosaminoglycans: a diagnostic test for ankylosing spondylitis. Annals of the rheumatic diseases 1988, 47:833-7. [201] Deveci H, Nur G, Çiçek H, Karapehlivan M: Evaluation of oxidative stress factors in patients with osteoporosis, 2017. [202] Schedin-Weiss S, Winblad B, Tjernberg LO: The role of protein glycosylation in Alzheimer disease. The FEBS journal 2014, 281:46-62. [203] Duan D: Micro-Dystrophin Gene Therapy Goes Systemic in Duchenne Muscular Dystrophy Patients. Human gene therapy 2018.

202