Neuromuscular Synaptic Proteins in Development and Regeneration

University of Pennsylvania ScholarlyCommons

Publicly Accessible Penn Dissertations

2017

Neuromuscular Synaptic Proteins In Development And Regeneration

Katherine Diane Gribble University of Pennsylvania, [email protected]

Follow this and additional works at: https://repository.upenn.edu/edissertations

Part of the Biology Commons, and the Neuroscience and Neurobiology Commons

Recommended Citation Gribble, Katherine Diane, "Neuromuscular Synaptic Proteins In Development And Regeneration" (2017). Publicly Accessible Penn Dissertations. 2318. https://repository.upenn.edu/edissertations/2318

This paper is posted at ScholarlyCommons. https://repository.upenn.edu/edissertations/2318 For more information, please contact [email protected]. Neuromuscular Synaptic Proteins In Development And Regeneration

Abstract Following injury, peripheral nerves reestablish neuromuscular connections with their developmental targets. While the molecular pathways that govern peripheral nerve development and neuromuscular synapse formation are mostly understood, it is unclear whether the same pathways are reemployed to promote regeneration. For example, neuromuscular synapses form through the Agrin-Lrp4-MuSK signaling pathway that clusters acetylcholine receptors beneath motor axon terminals, yet in vivo role of this pathway in peripheral nerve regeneration has not been examined. To determine whether this pathway is employed during regeneration, I used a previously established assay to transect spinal motor nerves and continuously monitor regeneration in live, intact zebrafish. Using live imaging I find that in ebrz afish pioneering axons establish a regenerative path for follower axons. I find that this process requires the synaptic receptor lrp4 and that in lrp4 mutants pioneer axon regrowth is unaffected while follower axons frequently stall at the site of injury, providing evidence for molecular diversity between pioneering and follower axons during regeneration. I demonstrate that Lrp4 promotes regeneration independent of membrane anchoring and of MuSK co-receptor signaling essential for synaptic development. Finally, I show that Lrp4 coordinates the realignment of denervated Schwann cells with regenerating axons, consistent with a model by which Lrp4 is repurposed to promote sustained peripheral nerve regeneration via axon-glia interactions. In a broader context, my findings demonstrate that certain molecular players involved in development are reused in regeneration in novel contexts, lending credence to the argument that studying mechanisms of synapse development enhances our understanding of mechanisms of peripheral nerve regeneration.

Degree Type Dissertation

Degree Name Doctor of Philosophy (PhD)

Graduate Group Neuroscience

First Advisor Michael Granato

Subject Categories Biology | Neuroscience and Neurobiology

This dissertation is available at ScholarlyCommons: https://repository.upenn.edu/edissertations/2318 NEUROMUSCULAR SYNAPTIC PROTEINS IN DEVELOPMENT AND REGENERATION

Katherine D. Gribble

A DISSERTATION

Neuroscience

Presented to the Faculties of the University of Pennsylvania

Partial Fulfillment of the Requirements for the

Degree of Doctor of Philosophy

2017

Supervisor of Dissertation

______

Michael Granato, Ph.D. Professor of Cell and Developmental Biology

Graduate Group Chairperson

______

Joshua Gold, Ph.D. Professor of Neuroscience

Dissertation Committee:

Jonathan Raper, Ph.D., Professor of Neuroscience Greg Bashaw, Ph.D., Professor of Neuroscience Judith Grinspan, Ph.D., Research Professor of Neurology Wenqin Luo, M.D., Ph.D., Assistant Professor of Neuroscience

NEUROMUSCULAR SYNAPTIC PROTEINS IN DEVELOPMENT AND REGENERATION

2017

Katherine Diane Gribble

This work is licensed under the

Creative Commons Attribution-

NonCommercial-ShareAlike 3.0

License

To view a copy of this license, visit https://creativecommons.org/licenses/by-nc-sa/3.0/us/

To William and Arthur

And to my family

iii

ACKNOWLEDGMENTS

I am deeply grateful to my advisor, Michael Granato, for his unwavering support and guidance during graduate school. Michael is an advisor who values more than the perfect experiment, and who has taught me how to present well, how to write in a compelling way (which is hopefully evident throughout this document), networking skills, and fortitude and patience as a scientist and a person.

Many thanks to my dissertation committee—Jonathan Raper, Wenqin Luo, Amin Ghabrial, Judy Grinspan, and Greg Bashaw—for their helpful experimental suggestions and career advice, and for being a great group of people to interact and work with.

I would like to thank the past and current members of the Granato lab for the camping and kayaking and biking trips; “lab meeting” brunches that spiraled into no-work Fridays; cookie breaks; happy hours; all the conferences we attended together that felt like summer camp; and most of all for the supportive environment built around mentorship and the unspoken convention that senior members actively mentor more junior members in all things scientific and not. Graduate students and post-docs in other labs are envious of our lab’s bond, and I charge the current and future members of the lab with striving to maintain such a positive, supportive workplace.

To my dear friends: thank you for the laughs and the love and for providing distractions when they were much needed.

Thank you a thousand times to my family. I am the person I am because of their love, constant cheerleading, willingness to listen, and the numerous family vacations throughout the years that refreshed and rejuvenated me.

I am infinitely fortunate to have found a true life-partner in my husband, William. I joke that he has served as an unofficial member of my dissertation committee, but it’s true. William gives me experimental and career advice; he gives me the unwavering and unconditional support of a spouse. He is my sounding board and my rock. Finally, I would like to thank our son Arthur. It may seem silly to thank a 10-month-old for his support during my seven-year dissertation… but the joy that he brings me overcomes all.

ABSTRACT

NEUROMUSCULAR SYNAPTIC PROTEINS IN DEVELOPMENT AND REGENERATION

Katherine D. Gribble

Michael Granato

Following injury, peripheral nerves reestablish neuromuscular connections with their developmental targets. While the molecular pathways that govern peripheral nerve development and neuromuscular synapse formation are mostly understood, it is unclear whether the same pathways are reemployed to promote regeneration. For example, neuromuscular synapses form through the Agrin-Lrp4-MuSK signaling pathway that clusters acetylcholine receptors beneath motor axon terminals, yet in vivo role of this pathway in peripheral nerve regeneration has not been examined. To determine whether this pathway is employed during regeneration, I used a previously established assay to transect spinal motor nerves and continuously monitor regeneration in live, intact zebrafish. Using live imaging I find that in zebrafish pioneering axons establish a regenerative path for follower axons. I find that this process requires the synaptic receptor lrp4 and that in lrp4 mutants pioneer axon regrowth is unaffected while follower axons frequently stall at the site of injury, providing evidence for molecular diversity between pioneering and follower axons during regeneration. I demonstrate that Lrp4 promotes regeneration independent of membrane anchoring and of MuSK co-receptor signaling essential for synaptic development. Finally, I show that Lrp4 coordinates the realignment of denervated Schwann cells with regenerating axons, consistent with a model by which Lrp4 is repurposed to promote sustained peripheral nerve regeneration via axon- glia interactions. In a broader context, my findings demonstrate that certain molecular players involved in development are reused in regeneration in novel contexts, lending credence to the argument that studying mechanisms of synapse development enhances our understanding of mechanisms of peripheral nerve regeneration.

TABLE OF CONTENTS

ACKNOWLEDGMENTS ...... IV

ABSTRACT ...... V

LIST OF ILLUSTRATIONS ...... X

CHAPTER 1: INTRODUCTION ...... 1

1.1 Overview: Coordinated body movements require neuromuscular synapses ...... 1

1.2 Constructing neuromuscular synapses ...... 3

1.3 Positioning neuromuscular synapses ...... 6

1.4 Are neuromuscular synaptic proteins required for motor axon regeneration? ...... 9

1.5 Figures ...... 12

CHAPTER 2: THE ROLES OF LRP4 AND DOK-7 IN NEUROMUSCULAR SYNAPSE

POSITIONING AND FORMATION ...... 14

2.1 Abstract ...... 14

2.2 Introduction ...... 14

2.3 Experimental Procedures ...... 16

2.3.1 Zebrafish strains and animal care ...... 16

2.3.2 Generation of TALEN mutant alleles ...... 17

2.3.3 Whole mount immunohistochemistry and imaging ...... 17

2.4 Results ...... 19

2.4.1 Zebrafish lrp4 is required for neuromuscular synapse formation, but not synapse

positioning ...... 19

2.4.2 Zebrafish lrp4 is expressed in developing hindbrain, but is not required for performance

of acoustic startle response ...... 21

2.4.3 Zebrafish Dok-7 is not required for AChR pre-patterning or synapse formation ...... 22

2.5 Discussion ...... 24

2.6 Figures ...... 28

CHAPTER 3: THE SYNAPTIC RECEPTOR LRP4 PROMOTES PERIPHERAL NERVE

REGENERATION ...... 36

3.1 Abstract ...... 36

3.2 Introduction ...... 36

3.3 Experimental Procedures ...... 38

3.3.1 Zebrafish strains and animal care ...... 38

3.3.2 Molecular identification of ennui gene ...... 38

3.3.3 Generation of agrinp190 mutants ...... 39

3.3.4 Spinal motor nerve transection assays ...... 39

3.3.5 Sparse neuronal labeling ...... 40

3.3.6 Whole mount immunohistochemistry and imaging ...... 40

3.4 Results ...... 41

3.4.1 Pioneering axons establish a regenerative path for follower axons ...... 41

3.4.2 lrp4 promotes robust nerve regeneration ...... 42

vii

3.4.3 lrp4 coordinates the realignment of denervated Schwann cells with regenerating axons

3.4.4 Lrp4 promotes peripheral nerve regeneration independently of MuSK signaling ...... 45

3.4.5 A truncated form of Lrp4 promotes nerve regeneration ...... 46

3.4.6 Lrp4 does not act in motor neurons or muscle cells to promote axon regeneration .. 47

3.5 Discussion ...... 48

3.5.1 Lrp4 promotes peripheral nerve regeneration independently of its role in development

3.5.2 Lrp4 reveals molecular diversity between pioneering and follower axons in

regeneration ...... 49

3.5.3 Does Lrp4 coordinate axonal regeneration with Schwann cell plasticity? ...... 51

3.6 Figures ...... 53

3.7 Supplemental Movies ...... 71

CHAPTER 4: DISCUSSION ...... 74

4.1 The future of synapse positioning ...... 74

4.2 Lrp4 is a multi-functional cell-surface receptor and ligand ...... 76

4.3 Neuron-extrinsic cues to promote axon regeneration ...... 79

CHAPTER 5: APPENDIX ...... 82

5.1 Generation of MuSKΔKringle mutant zebrafish ...... 82

5.1.1 sgRNA design and current results ...... 82

5.1.2 MuSK Kringle domain next steps ...... 83

viii

5.2 Juvenile Nerve Transection Protocol ...... 84

5.2.1 Part 1: Raising the fish (the first 5 days) ...... 84

5.2.2 Part 2: Nerve transection (days 28-34) ...... 84

BIBLIOGRAPHY ...... 89

LIST OF ILLUSTRATIONS

Figure 1-1 : An Agrin-Lrp4-MuSK signaling pathway constructs neuromuscular synapses ...... 12

Figure 1-2 : A Wnt/MuSK signaling pathway positions neuromuscular synapses before motor axon

arrival ...... 13

Figure 2-1 : Zebrafish Lrp4 is required for synapse formation but not AChR pre-patterning ...... 28

Figure 2-2 : AChR pre-pattern width is unaffected in lrp4 mutants ...... 30

Figure 2-3 : En passant neuromuscular synapses are absent in lrp4 mutants at 36 hpf ...... 31

Figure 2-4 : Neuromuscular synapses are reduced and disordered in lrp4 mutants at 5 dpf ...... 32

Figure 2-5 : Lrp4 is not required for habituation to a loud acoustic stimulus ...... 33

Figure 2-6 : Zebrafish Dok-7 is not required for synapse formation or AChR pre-patterning ...... 34

Figure 3-1 : Pioneering axons establish a regenerative path for follower axons ...... 53

Figure 3-2 : lrp4 is required for axonal growth during regeneration ...... 56

Figure 3-3 : Schwann cell morphology during early regeneration is disrupted in lrp4 mutants ..... 58

Figure 3-4 : lrp4 acts in an Agrin/MuSK-independent pathway to promote axonal regeneration .. 61

Figure 3-5 : A truncated version of Lrp4 promotes nerve regeneration ...... 63

Figure 3-6 : Specifically expressing lrp4 in motor neurons or muscle cells does not rescue axon

regeneration ...... 65

Figure 3-7 : Skeletal muscle-specific lrp4 expression rescues the developmental neuromuscular

synaptic phenotype ...... 67

Figure 3-8 : Zebrafish Agrin is required for neuromuscular synapse formation ...... 69 x

Figure 5-1 : Sequences of MuSK Kringle sgRNAs and whether sgRNAs worked ...... 82

Figure 5-2 : Zebrafish MuSK domain structure ...... 83

Figure 5-3 : Peripheral motor nerves in 4-week-old fish regenerate after transection injury ...... 87

CHAPTER 1: INTRODUCTION

1.1 Overview: Coordinated body movements require neuromuscular synapses

Think for a moment about your morning. You woke up, showered and got dressed. You brewed and sipped on a cup of coffee while chewing on toast and turning the pages of the newspaper. You typed an email. Perhaps you walked the dog, or helped your children get ready for school. Eventually you packed your work bag, locked up the house, and began your own commute into work, whether by foot, bike, bus, or car. What is a critical feature that these quotidian events share? They all require coordinated, precise body movements. It is a powerful feature of our bodies most of us take for granted after childhood, that if we choose to perform an action, our bodies will obey and the action will be executed flawlessly.

Performing any movement requires the complex interactions of many neurons in the brain and spinal cord through the functional units of neuronal communication called synapses. It is estimated that there are 100 trillion synapses in a person’s body that allow him or her to perceive, experience, and interact with the world. Indeed, there are so many types of synapses in the human body, and the majority are packed into the brain, that it would be impossible to try to understand how each one forms and changes throughout an individual’s life. However, one synapse that is easier to study because of its size and accessibility is the synapse between spinal motor neurons and their muscle fiber targets, called the neuromuscular synapse. Because neuromuscular synapses are much larger than brain synapses and easy to access experimentally, studying them as a “model” synapse has contributed greatly to our understanding of synaptogenesis in general and to our understanding of neuromuscular disorders. Indeed, understanding the molecular mechanisms regulating how neuromuscular synapses develop, and how they can change with disease or injury, are central questions in modern neuroscience.

How, generally, do synapses form? The process of synapse formation includes the selection of the future synaptic site and the coordinated differentiation of pre-synaptic neurons and

post-synaptic target structures. The location on a post-synaptic cell where a synapse forms is often highly stereotyped. For example, in zebrafish, cranial nerve VIII synapses onto the lateral dendrite of the Mauthner neuron, while spiral fiber neuron axons synapse at the axon hillock of the Mauthner neuron (Korn and Faber, 2005). Similarly, neuromuscular synapses usually form at equatorial locations along the muscle fiber, giving rise to a focal synapse at the center of each individual muscle fiber (Hall and Sanes, 1993). Thus, even on a specific cell or cell type, synaptic inputs occur at different stereotyped positions, but the molecular mechanisms regulating synapse positioning remain poorly understood. As I will discuss below, the neuromuscular synapse is an excellent model synapse for studying synapse positioning as well as pre- and post-synaptic differentiation.

Early observations about neuromuscular synapses revolved around the post-synaptic side of the synapse, where dense clusters of nicotinic acetylcholine receptors (nAChRs or AChRs) are localized to bind motor axon-released acetylcholine. Structural and biochemical studies of AChRs were facilitated by using the electric organ of Torpedo electric rays, which resembles muscle but is more densely packed with cholinergic synapses, thereby providing a rich source of AChRs

(Karlsson et al., 1972; Schmidt and Raftery, 1972; 1973; reviewed in Changeux et al., 1984;

Salpeter and Loring, 1985). Indeed, as I will discuss in the next section, many of the synaptic organizing components of neuromuscular synapses were originally identified from Torpedo electric organs. Furthermore, the discovery of snake venom neurotoxins that bind acetylcholine receptors, such as a-bungarotoxin, allowed for direct visualization and immunohistochemical studies of

AChRs (Chang and Lee, 1963; reviewed in Salpeter and Loring, 1985). Using such tools, researchers observed that clusters of AChRs change dramatically throughout development: early, before the pre-synaptic motor neuron has contacted the muscle fiber, loose AChR clusters form in the center area of muscle cells in what is called the AChR “pre-pattern” (Bevan and Steinbach,

1977; Steinbach, 1981; Ziskind-Conhaim and Dennis, 1981; Ziskind-Conhaim and Bennett, 1982;

Lin et al., 2001; Yang et al., 2001); later, when the motor neuron contacts the muscle fiber, AChR clusters become denser and with sharper borders, packing together directly beneath the motor 2

axon terminal to form one of the largest chemical synapses found in the body (for review Sanes and Lichtman, 2001). In mammals, these mature clusters resemble pretzels and, when combined with fluorescent labeling of motor axons, the whole structure resembles a budding rose, with the motor axon “stalk” ending in a dense AChR “rosette”. These observations about AChR cluster behavior throughout development raised important questions about neuromuscular synapse formation specifically, and synapse formation in general: (1) What forces or signals regulate the changing AChR clusters to construct neuromuscular synapses? (2) Which partner directs synapse formation: the pre-synaptic motor axon or the post-synaptic muscle cell? (3) Which entity, motor axon or muscle cell, decides where to form the future synapse? (4) How does the motor axon maintain such a faraway synapse, and can motor axons regenerate to their original targets after sustaining injury? In the next sections of this introduction, I will address these questions within the framework of the relevant molecular players and their actions across developmental time.

1.2 Constructing neuromuscular synapses

When Agrin was identified, it was beheld as one of the first synaptic organizing factors.

Agrin is a motor neuron-derived proteoglycan that localizes in the synaptic basal lamina and is so named because it is a potent “aggregator” of AChRs when applied to cultured muscle cells (Godfrey et al., 1984; Nitkin et al., 1987; Campanelli et al., 1991; Ruegg et al., 1992). Indeed, Agrin is such a potent builder of neuromuscular synapses that it is both necessary and sufficient for mature AChR cluster formation. In mice lacking the neural isoform of Agrin, neuromuscular synapses fail to form and AChRs are evenly distributed along the length of muscle fibers rather than clustered tightly beneath motor axon terminals (Gautam et al., 1996). Agrin is sufficient to drive the formation of

AChR clusters both in denervated muscles and in adult muscles at ectopic locations (Jones et al.,

1997; Bezakova et al., 2001). Clearly, Agrin is a key orchestrator of neuromuscular synapse formation, but the whole story isn’t quite so simple. For example, in Agrin mutant mice, while mature

AChR clusters fail to form, the aneural pre-patterned AChR clusters develop normally, demonstrating that Agrin is not required for pre-patterning (Lin et al., 2001; Yang et al., 2001). 3

Another interesting twist in the Agrin story is that motor neurons, Schwann cells, and muscle cells all produce Agrin, but only the motor neuron-derived isoform which includes certain exons at the

C-terminus can induce AChR clusters (Ferns et al., 1992; 1993; Hoch et al., 1993; Gesemann et al., 1995; Burgess et al., 1999). Collectively, these studies demonstrate that the pre-synaptic motor neuron acts as a key organizer of synapses by signaling its arrival at the muscle cell through Agrin.

How does Agrin communicate with the post-synaptic muscle cell to cluster AChRs? In searching for an Agrin receptor on muscle cells, another key organizer of neuromuscular synapses was identified: the receptor tyrosine kinase called muscle-specific kinase, or MuSK (Jennings et al., 1993; Valenzuela et al., 1995). MuSK localizes to neuromuscular synapses, and genetic deletion of Musk results in the absence of neuromuscular synapses (DeChiara et al., 1996). In

Musk mutant mice, both pre-patterned AChRs and mature AChR clusters are absent, indicating that signaling through MuSK is a critical step in all arrangements of clustered AChRs (Lin et al.,

2001; Yang et al., 2001). When MuSK was discovered, it was assumed that MuSK was the Agrin receptor, but it quickly became clear that neural Agrin and the MuSK ectodomain do not directly bind, although neural Agrin does cluster AChRs through a protein complex that includes MuSK

(Glass et al., 1996). Like Agrin, MuSK is necessary and sufficient to form neuromuscular synapses: expressing ectopic MuSK in muscle cells restores neuromuscular synapses in animals lacking

Agrin (Kim and Burden, 2008). When Musk is absent, it was also observed that motor axons grow in a wider territory along the muscle cell, suggesting that MuSK plays a highly directive role in neuromuscular synapse formation by directing where synapses form (Zhang and Granato, 2000;

Kim and Burden, 2008). In section 1.3, I will present more evidence that strongly suggests MuSK is a synapse positioner.

An unanswered question in the field of neuromuscular synapse formation for a decade concerned the identity of the Agrin receptor. As previously stated, it was known that Agrin acted through MuSK to cluster post-synaptic receptors, but how the Agrin signal was communicated to

MuSK remained unknown, and a hypothetical molecule termed Myotube-Associated Specificity

Component (MASC) was presented as the missing link between Agrin and MuSK (Glass et al.,

1996). The concept of MASC was used until the Agrin receptor was discovered to be the low- density lipoprotein receptor (LDLR)-related protein 4 (Lrp4). Lrp4 was identified in a forward genetic screen for limb development: mice lacking Lrp4 showed severe limb and digit defects, and they died at birth because they could not breathe; further inspection revealed that they had neuromuscular junction defects that resembled those of Musk and Agrin mutant mice (Weatherbee et al., 2006). Just like Musk mutants, Lrp4 mutants lack both pre-patterned AChRs and mature

AChR clusters; and like MuSK, Lrp4 acts downstream of Agrin (Weatherbee et al., 2006). Soon thereafter, it was demonstrated through direct-binding studies that Lrp4 is the Agrin receptor, the missing link between Agrin and MuSK signaling in neuromuscular synapse formation (Kim et al.,

2008; Zhang et al., 2008). Lrp4 is a single-pass transmembrane domain protein that has a large ectodomain to bind Agrin and MuSK, but a very short intracellular domain that is thought to be non- signaling (Gomez and Burden, 2011; Pohlkamp et al., 2015). Lrp4 and MuSK form an interesting signaling pair because they each perform a critical signaling function that the other cannot: Lrp4 acts as the Agrin-binding ectodomain, while MuSK transduces the Agrin slash motor-neuron-arrival signal intracellularly through its kinase domain. But the story of Lrp4 continues to deepen: recent reports demonstrate that the Lrp4 ectodomain, probably through cleavage from the muscle cell surface, binds to the distal end of motor axons to promote pre-synaptic differentiation (Wu et al.,

2012; Yumoto et al., 2012). Additionally, Lrp4 is expressed in several areas of the central nervous system, including hippocampus, olfactory bulb, cerebellum, and is found in synaptic fractions (Tian et al., 2006), and has recently been demonstrated to play a role in CNS synapse function in mice and Drosophila (Gomez et al., 2014; Sun et al., 2016; Mosca et al., 2017). In Chapter 3, I will discuss my own research about a novel role for Lrp4 in peripheral motor axon regeneration.

How is the Agrin-Lrp4/MuSK signal transduced within muscle cells to cluster AChRs? Many intracellular signaling and adaptor proteins interact with the MuSK kinase domain (for in-depth 5

review see Wu et al., 2010), but here I will only mention two key players: the docking protein 7

(Dok-7) and the receptor-associated protein of the synapse (Rapsyn). Rapsyn is an intracellular scaffolding protein that associates with acetylcholine receptors as early as in the Golgi (Neubig et al., 1979; Burden et al., 1983; Park et al., 2012) and is required for AChR clustering and therefore neuromuscular synapse formation (Gautam et al., 1995). Dok-7 is a phosphotyrosine binding (PTB) adaptor protein that binds to a specific target motif on the MuSK kinase domain, stabilizing the active state of MuSK and promoting synapse formation; not surprisingly, Dok-7 mutant mice fail to form neuromuscular synapses, just like Musk, Agrin, Lrp4, and Rapsn mutants (Okada et al., 2006).

Together, the molecular pathway consisting of Agrin-Lrp4/MuSK-Dok-7-Rapsyn is essential for forming and maintaining neuromuscular synapses throughout an organism’s life (Figure 1-1).

Studying the molecular players of this pathway has contributed to our understanding of how other synapses form: Rapsyn informs us that scaffolding proteins may be required to shape post-synaptic receptor clusters; Dok-7 informs us about intracellular adaptor and signaling proteins that might instruct these receptors to cluster; and Agrin-Lrp4/MuSK informs us about the complexities of ligand-receptor signaling partners, and how these signaling pathways are not as simple as one ligand binding to one signaling receptor.

1.3 Positioning neuromuscular synapses

We understand well the molecules and mechanisms necessary to construct neuromuscular synapses, so what questions remain unanswered? One question is, how are synapses positioned during development? Neuromuscular synapses on mammalian skeletal muscles often form at equatorial locations in the center of the muscle fiber, but it is poorly understood how or why these synapses form there. Furthermore, understanding synapse positioning is a puzzling question in the study of the central nervous system, where more complex circuits develop stereotypically across organisms. How are these stereotyped synapses placed correctly?

Appreciating the nuances of neuromuscular synapse positioning began when researchers made observations about early AChR clusters. The earliest AChR clusters appear before the motor axon reaches the muscle fiber, and this so-called AChR pre-patterning happens independently of the nerve or nerve-derived factors like Agrin (Yang et al., 2000; Lin et al., 2001; Yang et al., 2001).

A major question in the field was whether AChR pre-patterning was a prerequisite for mature AChR clusters to form. Some studies showed that when a growing motor axon made contact with pre- patterned AChR clusters, these clusters were incorporated into mature neuromuscular synapses

(Flanagan-Steet et al., 2005; Panzer et al., 2006), suggesting that the AChR pre-patterned clusters were required for mature cluster formation. However, our lab subsequently performed an elegant experiment to address this question directly: using zebrafish embryos lacking musk, we induced expression of a heat-shock musk rescue transgene after the AChR pre-patterning stage. We observed that mature neuromuscular synapses formed normally, but they formed in aberrant locations (Jing et al., 2009). These experiments demonstrated that the AChR pre-pattern is not required for mature synapse formation per se, and instead pre-patterning is required to direct where future neuromuscular synapses will form.

Knowing that MuSK and its signaling partners—Lrp4, Dok-7, Rapsyn—are required for

AChR pre-patterning (Gautam et al., 1995; DeChiara et al., 1996; Okada et al., 2006; Weatherbee et al., 2006), but Agrin the canonical ligand of this signaling pathway is not required (Lin et al.,

2001), our lab wondered which ligand binds and activates MuSK to initiate pre-patterning. A prior study by Luo et al. (2002) reported that the cytoplasmic protein Disheveled, and specifically the

DEP domain of Disheveled that is active in non-canonical Wnt planar cell polarity signaling, is required for AChR clustering. In addition, other studies demonstrated critical roles for Wnt ligands in central nervous system synapse formation in various organisms (Hall et al., 2000; Packard et al.,

2002; Klassen and Shen, 2007). These findings led our lab to hypothesize that Wnt ligands bind

MuSK, initiating AChR pre-patterning and confining both AChR clusters and extending motor axon growth cones to the central region of the muscle fiber. We showed that Wnt11r and MuSK 7

genetically interact; that like MuSK, Wnt11r is required for AChR pre-patterning; and Wnt11r binds the MuSK ectodomain and activates MuSK (Jing et al., 2009). We subsequently showed that another Wnt ligand, Wnt4a, also binds and activates MuSK. Wnt/MuSK binding leads to MuSK endocytosis into the muscle cell and Rab11-dependent trafficking to position MuSK-positive endosomes at the center of the muscle cell, precisely where AChRs will cluster (Gordon et al.,

2012). There, MuSK interacts with Disheveled and other components of the planar cell polarity signaling pathway to establish a ‘central zone’ that restricts where AChRs cluster, where motor axons grow and ultimately where synapses will form (Figure 1-2).

Many outstanding questions remain regarding this pathway, including what roles Lrp4,

Dok-7, and Rapsyn play, since they are all required for AChR pre-patterning (Gautam et al., 1995;

Okada et al., 2006; Weatherbee et al., 2006). The role of Lrp4 in pre-patterning is unclear, because the only known role for Lrp4 in this pathway is to bind Agrin (Kim et al., 2008; Zhang et al., 2008), but Agrin is dispensable for pre-patterning (Lin et al., 2001). Lrp4’s closest relatives are the Wnt- binding proteins Lrp5 and Lrp6 (He et al., 2004), and all three proteins share structural homology of their ectodomains (Choi et al., 2009), though the precise ligand-binding domain is different between Lrp4 and Lrp5/6. I hypothesized that before the motor axon contacts the muscle, Lrp4 stabilizes the Wnt-MuSK interaction to promote pre-patterning. As I will discuss more in Chapter 2,

I found that Lrp4 is unexpectedly dispensable for pre-patterning in zebrafish, and this finding combined with new research on Wnt/MuSK binding in mice from Dr. Steven Burden’s lab led us to conclude that Lrp4’s role during neuromuscular synapse formation diverged during vertebrate evolution (Remédio et al., 2016). How does Dok-7 promote AChR pre-patterning? Dok-7 binds the

MuSK kinase domain to stabilize its activated state, but does it stay bound after MuSK endocytosis and trafficking? In other words, is Dok-7-dependent MuSK activation required to localize

Disheveled and establish the central zone? I hypothesized that Dok-7-dependent MuSK activation is required to pre-pattern AChRs and establish the central zone. I will discuss my somewhat murky results on these questions in Chapter 2. Finally, Rapsyn’s role in pre-patterning seems simple 8

because Rapsyn is required for AChRs to cluster. But an interesting finding from zebrafish suggests that Rapsyn is not simply downstream of the Wnt/MuSK pathway to establish pre-patterning and the central zone: in zebrafish rapsyn mutants, AChR pre-patterning is absent, but the motor axons still grow along the muscles within a restricted central territory, suggesting that the central zone forms independently of Rapsyn (Zhang et al., 2004). This raises the question of how and when

Rapsyn binds AChRs, and whether Rapsyn localization to AChRs is dependent on MuSK signaling.

I did not address this question in my dissertation research, but it is nevertheless an interesting future question.

1.4 Are neuromuscular synaptic proteins required for motor axon regeneration?

An intact neuromuscular synapse is critical to maintaining the integrity of the motor axon: in many diseases that lead to loss or degeneration of peripheral motor neurons, one of the earliest pathologies is degeneration of the synapse (Frey et al., 2000; Fischer et al., 2004; Schaefer et al.,

2005). Interestingly, overexpressing MuSK in a mouse model of ALS is protective against synapse loss and axon degeneration (Pérez-García and Burden, 2012), demonstrating that synapse integrity can delay disease onset or progression. Given that reinforcing the neuromuscular synapse can prevent axon degeneration, the other broad question I pursued in my dissertation research, and the focus of Chapter 3, was about whether neuromuscular synaptic proteins can also promote axon regeneration.

Axons of the peripheral nervous system can regenerate after sustaining an injury, a phenomenon which was described by Santiago Ramon y Cajal nearly 100 years ago (1928). After a distal axon is severed from its cell body, it initiates a stereotyped program of Wallerian degeneration to fragment the distal axon (Waller, 1850), leaving behind the cell body and proximal stump of the axon. Thereafter, Schwann cells and infiltrating macrophages engulf and phagocytose the axonal debris (Scherer and Easter, 1984; Stoll et al., 1989; Fernandez-Valle et al., 1995; Perry et al., 1995; Vargas and Barres, 2007), clearing the path for proximal axons to regrow. Axon regeneration begins when proximal axons sprout new growth cones that explore the environment 9

around the injury site, which eventually stabilize into growing axons that cross the injury gap and contact the denervated Schwann cells in the distal nerve stump. Previous studies reported using static time-point analysis during regeneration that regenerative axonal outgrowth is staggered, with small groups of axons extending at different times from the proximal stump (Al-Majed et al., 2000;

McDonald et al., 2005). However, this phenomenon has never been observed using live-cell imaging, so a major outstanding question is whether axons cross the injury gap in groups, or whether single axons pioneer the regenerative path for later-emerging follower axons to fasciculate along. I sought to answer this question in Chapter 3 by using our lab’s previously-established model of peripheral nerve regeneration in larval zebrafish, which features live-cell imaging to observe regenerating axons (Rosenberg et al., 2012; 2014; Isaacman-Beck et al., 2015).

After axons successfully cross the injury gap, they contact and grow along the denervated

Schwann cells of the distal nerve (Dyck and Hopkins, 1972; Scherer and Easter, 1984; Nguyen et al., 2002; Chen et al., 2005). Schwann cells are required for axon regeneration, and the process of axon degeneration initiates dramatic morphological and behavioral changes in distal Schwann cells that drive them towards a de-differentiated, activated, pro-regeneration state. For example, denervated Schwann cells up-regulate diffusible factors like NGF, BDNF, GDNF, and FGF to promote growth cone sprouting as well as axon growth and guidance (Taniuchi et al., 1986;

Heumann et al., 1987; Meyer et al., 1992; Funakoshi et al., 1993; Liu et al., 1995; Zhang et al.,

2000; Scarlato et al., 2001; Höke et al., 2002). They also degrade existing myelin and down- regulate myelin promoting genes by activating transcription factors and transcriptional regulators such as c-Jun, Notch, and Sox2 (Le et al., 2005; for an in-depth review see Jessen and Mirsky,

2008; Parkinson et al., 2008; Woodhoo et al., 2009; Arthur-Farraj et al., 2012). Axon-derived signals can also promote myelination. For example, axon-derived Neuregulin 1 type III determines myelin sheath thickness (Michailov et al., 2004; Taveggia et al., 2005), likely through cleavage by β- secretase-1 (Hu et al., 2006; Willem et al., 2006), and subsequent binding to ErbB receptors on

Schwann cells. Denervated Schwann cells realign with regenerating axons, and their morphology

changes back as they revert from an activated, regeneration-supporting Schwann cell to a pre- injury myelinating Schwann cell (y Cajal, 1928; Holtzman and Novikoff, 1965; Scherer and Easter,

1984; Cheng and Zochodne, 2002). Despite what is known about the mechanisms regulating

Schwann cell de-differentiation, the mechanisms underlying the realignment of denervated

Schwann cells with regenerating axons, and the mechanisms that trigger this pronounced change in Schwann cell morphology as they revert to a pre-injury myelinating Schwann cell, are not well understood. As I will discuss in Chapter 3, I used live-cell imaging in larval zebrafish to answer my questions about Schwann cell morphology changes during axon de- and regeneration. I sought to answer my first question about whether neuromuscular synaptic components are required for peripheral nerve regeneration, and in my studies I found that the Agrin receptor Lrp4 is required for follower axon outgrowth during regeneration, and Schwann cell morphology changes during regeneration.

In this thesis, I hope to convince you of several things. First, that studying neuromuscular synapses is not a hallmark of a bygone era, and we can learn about synaptogenesis in general and synapse positioning by studying this beautiful synapse. Second, that the core molecular components of the neuromuscular synapse do not exist solely to build and maintain this synapse; researchers regularly describe roles for these genes in novel cell types and contexts. And finally, I hope to show you that a good way to identify genes involved in axon regeneration is to study known developmental axon guidance or synapse formation pathways, as they are sometimes recapitulated to promote axon regeneration.

1.5 Figures

Figure 1-1: An Agrin-Lrp4-MuSK signaling pathway constructs neuromuscular synapses

As motor axons contact their muscle fiber targets, they release the proteoglycan Agrin to signal their arrival. Agrin binds to the Lrp4 receptor in the muscle fiber membrane, which changes the conformation of Lrp4 and allows it to bind MuSK and stimulate MuSK phosphorylation. MuSK signals through Dok-7 and Rapsyn to cluster AChRs beneath the motor axon terminal, thus building a stable synapse. After synapse formation, Agrin continues to signal through Lrp4 and MuSK to maintain the synaptic structure.

Figure 1-2: A Wnt/MuSK signaling pathway positions neuromuscular synapses before

motor axon arrival

Before motor axons contact their muscle fiber targets, muscle cells and dorsolateral somites release Wnt11r and Wnt4a ligands that bind and activate MuSK, leading to MuSK endocytosis into the muscle cell. After endocytosis, MuSK is trafficked into Rab11-positive endosomes that localize at the center of the muscle cell. There, MuSK interacts with Disheveled and other members of the non-canonical Wnt PCP signaling pathway, thereby restricting AChR clusters into a distribution that is called the “pre-pattern”, and restricting motor axons to grow along the center region of muscle where future synapses will form.

CHAPTER 2: THE ROLES OF LRP4 AND DOK-7 IN NEUROMUSCULAR SYNAPSE POSITIONING AND FORMATION

Modified from Gordon LR, Gribble KD, Syrett CM, Granato M. (2012). Initiation of synapse formation by Wnt-induced MuSK endocytosis. Development 139(5): 1023-1033. DOI: 10.1242/dev.071555.

Modified from Remédio L, Gribble KD, Lee JK, Kim N, Hallock PT, Delestrée N, Mentis GZ, Froemke RC, Granato M, Burden SJ. (2016). Diverging roles for Lrp4 and Wnt signaling in neuromuscular synapse development during evolution. Genes and Development 30(9): 1058-1069. DOI:10.1101/gad.279745.116

2.1 Abstract

Coordinated body movements depend on the precise synaptic apposition between motor neurons and muscle cells at a specialized area in the center of muscle cells. Before motor axons contact their muscle cell targets, Wnt ligands bind to the MuSK receptor tyrosine kinase on the surface of muscle cells, initiating a planar cell polarity (PCP)-like signaling pathway that restricts growing motor axons and the first AChR clusters to the center of the muscle cell. In mice loss of

Dok-7 and Lrp4, critical signaling proteins to build and maintain neuromuscular synapses, similarly disrupts AChR pre-patterning and leads to motor axon overgrowth, but where Dok-7 and Lrp4 act in the Wnt/MuSK signaling pathway is unclear. I sought to answer these questions by generating zebrafish mutants lacking these genes. I find that loss of lrp4 disrupts en passant neuromuscular synapses but not AChR pre-patterning, which conflicts with the established role of mammalian Lrp4 in AChR pre-patterning. Loss of dok-7 slightly reduces en passant neuromuscular synapses, but again AChR pre-patterning is unaffected. In this chapter I will discuss my efforts to generate these mutants, my results in analyzing AChR pre-patterning and synapse formation in these mutants, and future questions for the field in synapse positioning and synapse formation.

2.2 Introduction

In vertebrates, survival depends upon performing complex coordinated movements. All movements are controlled by the precise synaptic connections between spinal motor neurons and muscle cells, called neuromuscular synapses. Neuromuscular synapses form when the presynaptic

motor neuron extends its axon out of the spinal cord and contacts its muscle fiber target at a specialized zone in the center of the muscle fiber where the future synapse will form. As discussed in Chapter 1, neuromuscular synapse development occurs in two distinct stages. The first stage is marked by loose pre-patterned clusters of post-synaptic acetylcholine receptors forming on the surface of muscle fibers (Bevan and Steinbach, 1977; Steinbach, 1981; Ziskind-Conhaim and

Dennis, 1981; Ziskind-Conhaim and Bennett, 1982). This AChR pre-pattern forms independently of the nerve or nerve-derived factors: when motor neurons or motor neuron-derived Agrin are absent, AChR pre-patterning occurs normally (Yang et al., 2000; Lin et al., 2001; Yang et al., 2001).

Our lab has previously shown that the Wnt ligands Wnt11r and Wnt4a bind MuSK, initiating MuSK endocytosis and trafficking into Rab11-positive endosomes that localize to the center of the muscle cell (Jing et al., 2009; Gordon et al., 2012). There, Wnt/MuSK interacts with other members of the non-canonical Wnt PCP signaling pathway, including Disheveled, Daam1, and Diversin (Gordon et al., 2012), thereby restricting the AChR pre-pattern to a central area on the muscle cell surface.

Disrupting Wnt/MuSK signaling leads to ectopic synapse formation, thus the Wnt/MuSK signaling cascade determines where along the muscle cell future synapses will form (Jing et al., 2009; 2010).

When I started my doctoral research, I wondered what other genes are required to establish the Wnt/MuSK-dependent central zone? In mice, Lrp4 and Dok-7 are required for both

AChR pre-patterning and neuromuscular synapse formation (Okada et al., 2006; Weatherbee et al., 2006). To build neuromuscular synapses during the late stage, Lrp4 binds Agrin and signals through MuSK to tell the muscle cell that the motor axon has arrived (Kim et al., 2008; Zhang et al.,

2008). But I previously told you that AChR pre-patterning occurs independently of the motor nerve or Agrin, so what is the role for Lrp4 in establishing AChR pre-patterning? The closest relatives of

Lrp4 are Lrp5 and Lrp6, which act as Wnt co-receptors with Frizzled proteins in canonical Wnt signaling pathways (Pinson et al., 2000; Tamai et al., 2000). This led me to hypothesize that Lrp4 stabilizes the Wnt/MuSK interaction during AChR pre-patterning, and that loss of Lrp4 weakens this interaction, leading to loss of AChR pre-patterning. Dok-7 is a cytoplasmic protein that 15

associates with the inside of the cell membrane through its pleckstrin-homology domain and binds to phosphorylated Tyrosine residues on the MuSK kinase domain via its phosphotyrosine binding

(PTB) domain. Dok-7 not only stabilizes MuSK activation, but also it promotes MuSK activation

(Inoue et al., 2009). Dok-7 is also required for pre-patterning, a stage when MuSK is endocytosed from the muscle cell membrane and localizes to Rab11-positive endosomes in the perinuclear region of the muscle cell. Where is Dok-7 localized during these intracellular events—does Dok-7 also leave the muscle cell membrane to follow MuSK into Rab11-positive endosomes?

Furthermore, does MuSK need to be in a Dok-7-dependent phosphorylated/activated state to bind

Disheveled and cluster AChRs?

To answer my questions about the roles of Lrp4 and Dok-7 in synapse positioning, I generated zebrafish mutants in these genes using TALENs. Interestingly, I find that in zebrafish neither Lrp4 nor Dok-7 are required for AChR pre-patterning, and Lrp4 but not Dok-7 is required for late neuromuscular synapse formation. My findings were surprising and unexpected, and suggest that mice and zebrafish evolved different ways of positioning, forming and maintaining neuromuscular synapses.

2.3 Experimental Procedures

2.3.1 Zebrafish strains and animal care

p184 p185 p189 Zebrafish lrp4 , lrp4 , and dok-7 , alleles were generated in the TLF background, and all wild-type fish used were TLF strain. Homozygous lrp4 mutants were obtained by crossing heterozygous carriers and identified at 36 hours post-fertilization (hpf) based on swimming motility defects in response to light touch. Both lrp4p184 and lrp4p185 alleles showed identical pre-patterning and neuromuscular phenotypes; thus only the results from the lrp4p184 allele are shown.

Homozygous dok-7p189 fish were obtained by crossing heterozygous carriers and genotyping offspring, as no visible swimming motility defects were observed at any stage. Other transgenic lines used were: Tg(mnx1:GFP)ml2 (Flanagan-Steet et al., 2005) and Tg(smyhc1:mCherry- 16

CAAX)p149 (Jing et al., 2010). All fish were raised and maintained as described previously (Mullins et al., 1994). We performed all experiments involving fish according to animal protocols that were approved by the University of Pennsylvania IACUC.

2.3.2 Generation of TALEN mutant alleles

TALE nuclease plasmids were designed and engineered by the University of Utah Mutation

Generation and Detection Core Facility and sub-cloned into pCS2TAL3-DDD and pCS2TAL3-RRR

(Lrp4) or pCS2TAL3-DD and pCS2TAL3-RR (Dok-7). Lrp4 exon 9 was targeted: the left TALEN was designed to target the sequence 5’-TCCTCCATGTGCGCCCGAT-3’; the right TALEN was designed to target the sequence 5’-ATGGCCGCTGTATTGGACA-3’. Lrp4 TALEN mRNA was transcribed using the SP6 mMessage mMachine Kit (Ambion) and was diluted to 50 pg in 0.1 M

KCl and microinjected into one-cell stage TLF embryos. Dok-7 exon 3 was targeted: the left TALEN was designed to target the sequence 5’-GCTCTCAGTCTGCTCT-3’; the right TALEN was designed to target the sequence 5’-GTCGAAACCCAGCAT-3’. Dok-7 TALEN mRNA was transcribed as above and diluted to 100 pg in 0.1 M KCl and microinjected into one-cell stage TLF embryos.

Successful dsDNA breaks in lrp4 or dok-7 of G0 injected embryos were confirmed by PCR and high-resolution melt analysis. Heterozygous F1 carrier alleles were identified and characterized by

PCR followed by high-resolution melt analysis and sequencing (Dahlem et al., 2012).

2.3.3 Whole mount immunohistochemistry and imaging

Zebrafish embryos at 19–26 hpf were anesthetized in 0.01% Tricaine, fixed in 4% paraformaldehyde with 1% DMSO diluted in phosphate-buffered saline (1x PBS at pH 7.4) for 3h at room temperature, and washed with PBS. Embryos were permeabilized using 0.1% collagenase for 3–7 min at room temperature and then washed thoroughly with PBS. To label AChR clusters, embryos were incubated for 3h at 4°C in 10 μg/mL Alexa fluor 594-conjugated α-BGT (Molecular

Probes) diluted in incubation buffer (0.2% BSA, 0.5% Triton X-100 in 1x PBS) with 1% normal goat serum added. Motor axons were labeled with the znp-1 antibody (1:200 Trevarrow et al., 1990)

overnight at 4°C followed by incubation with goat anti-mouse Alexa fluor 488-conjugated secondary antibody (1:500; Molecular Probes). Stained embryos were immersed overnight and subsequently mounted in VectaShield mounting medium (Vector Laboratories). Embryos were imaged in 1-μm sections using a 60x immersion objective on a Zeiss LSM 710 confocal microscope. Image stacks were compressed into maximum intensity projections in Fiji and converted to 16-bit images using

MetaMorph (Molecular Devices). AChR clusters were counted using the “count nuclei” function with minimum/maximum length set at 4/30 and minimum average intensity set at 15. The results were recorded in GraphPad Prism for statistical analysis.

In situ hybridization for lrp4 mRNA was performed using the RNAscope kit (Advanced Cell

Diagnostics, Inc.) in whole zebrafish embryos expressing Tg(smyhc1:mCherry-CAAX)p149 (Jing et al., 2010) as previously described (Gross-Thebing et al., 2014). Probes against lrp4 were designed and engineered by Advanced Cell Diagnostics. Embryos were imaged in 1-μm sections on a 60x immersion objective on an Olympus spinning disk confocal microscope. Image stacks were compressed into maximum intensity projections and processed using Adobe Photoshop to adjust brightness and contrast. For chromogenic in situ hybridization for lrp4 mRNA in whole-mount zebrafish embryos, 2 kb from the amino-terminus of lrp4 was subcloned into pGEM-T easy;

Digoxigenin-labeled sense probe was transcribed using T7 polymerase, and Digoxigenin-labeled anti-sense probe was transcribed using SP6 polymerase. In situ hybridization protocol was adapted from protocols by Jennifer Rhodes and Chi-Bin Chien (personal communication with Amy Kugath).

Whole-mount embryo staining was visualized using a brightfield dissecting microscope (data not shown).

2.4 Results

2.4.1 Zebrafish lrp4 is required for neuromuscular synapse formation, but not synapse

positioning

In mice, Lrp4 is required during the late synapse formation stage where it acts as the Agrin receptor, as well as during the early AChR pre-patterning stage (Weatherbee et al., 2006; Kim et al., 2008; Zhang et al., 2008). Pre-patterned AChRs form independently of Agrin, so the role of

Lrp4 during this stage is unclear. The closest relatives of Lrp4 are the Wnt co-receptors Lrp5/6, which together with the Frizzled receptor signal through the canonical Wnt pathway (Pinson et al.,

2000; Tamai et al., 2000; for review see He et al., 2004). Furthermore, the Wnt-binding β-propeller domain on Lrp5/6 is conserved in Lrp4, and Lrp4 has been shown in other systems to interact with and modulate Wnt signaling (Ohazama et al., 2008; Choi et al., 2009). In zebrafish, Wnt signaling through the MuSK Frizzled-like Cysteine Rich Domain (CRD) is required to pre-pattern AChRs. I wondered whether muscle pre-patterning in zebrafish is controlled exclusively by a Wnt/MuSK- dependent mechanism or whether Lrp4 also has a role in pre-patterning in zebrafish. I used

TALENS to generate two zebrafish lrp4 mutant alleles with premature stop codons following the N- terminal LDLa repeats, thus deleting most of the MuSK- and Agrin-binding domains (Figure 2-

1A,B). Importantly, these zebrafish lrp4 alleles closely resemble the mouse Lrp4 mutant allele lacking pre-patterning and synapse formation (Weatherbee et al., 2006).

In zebrafish embryos, pre-patterned AChR clusters form in the central region of adaxial muscle cells prior to the arrival of motor axons, about 16 hours post-fertilization (hpf) (Flanagan-

Steet et al., 2005; Panzer et al., 2006). As motor neuron growth cones traverse the muscle territory, between 17 and 29 hpf, some of the aneural pre-patterned AChR clusters become incorporated into stable en passant neuromuscular synapses (Flanagan-Steet et al., 2005).

I stained lrp4 mutant zebrafish embryos at 19.5 hpf with α-bungarotoxin (α-BTX) and an antibody to Synaptotagmin-2 (znp-1; Trevarrow et al., 1990), and found that AChR clusters were 19

present in lrp4 mutant and wild-type sibling embryos (Figure 2-1C,C’). I measured the width of the band of pre-patterned AChR clusters relative to the width of each hemisegment and found no difference between wild-type and lrp4 mutant embryos (wild-type, n = 17 hemisegments, 20.17% ±

1.24%; lrp4 mutant, n = 30 hemisegments, 21.73% ± 1.16%; mean ± SEM, P = 0.4, unpaired

Student’s t-test; Figure 2-2). These findings indicate that zebrafish use a Wnt/MuSK signaling mechanism, exclusive of Lrp4, for muscle pre-patterning. However, Lrp4 is required for neuromuscular synapse formation in zebrafish, as en passant synapses between motor axons and axial muscles fail to form in the absence of Lrp4 (Figure 2-1D,D’; Figure 2-3), demonstrating that

Lrp4 has a conserved and essential role in synapse formation in lower and higher vertebrates. A few abnormally shaped AChR clusters remain on the surface of muscle pioneer cells in lrp4 mutants at 26 hpf (Figure 2-1D’, white arrowheads), although the significance of these clusters is unknown, and they are absent at 36 hpf, at the stage when more en passant synapses exist and before non- focal innervation occurs (Figure 2-3). In mammals, Lrp4 is expressed as myoblasts begin to fuse to form multinucleated fibers and prior to innervation (Kim et al., 2008). I therefore wondered whether lrp4 is expressed in early muscle development and during the pre-patterning stage in zebrafish embryos. Using RNAscope in situ hybridization technology, which is optimized to detect low-level transcripts (Gross-Thebing et al., 2014), I found that lrp4 mRNA is not detectable at the pre-patterning stage (Figure 2-1E) but is readily detectable when neuromuscular synapses are forming at 36 hpf (Figure 2-1E’). These findings are consistent with my data showing that Lrp4 is not required for AChR pre-patterning in zebrafish and indicate that the timing of lrp4 expression in muscle differs in zebrafish and mice.

By 5 days post-fertilization (dpf) neuromuscular synapses are present in lrp4 mutants, although there appears to be a reduction in the number of synapses, and the synapses present appear disorganized (Figure 2-4). This finding is not altogether surprising: the Granato lab previously reported a similar phenotype in zebrafish musk mutants, which show a reduction but not absence of neuromuscular synapses (Zhang et al., 2004). In zebrafish mutants lacking musk or 20

lrp4, non-focal and myoseptal neuromuscular synapses develop and are functional. Why is this the case? Further studies by the lab demonstrated that knocking down β-Dystroglycan in zebrafish musk mutants abolished nearly all the remaining non-focal and myoseptal neuromuscular synapses (Lefebvre et al., 2007). This finding suggested that focal or equatorial synapses require

MuSK signaling, and that non-focal or distributed synapses form through alternative signaling pathways involving β-Dystroglycan. Curiously, in lower vertebrates like birds, fish, and amphibians, muscle fibers often display non-focal or distributed innervation patterns; in mammals, muscle fibers are more often focally innervated (Silver, 1963). Despite the reduction in synapses in lrp4 mutants, they show no apparent defects in normal swimming or in startle-response escape swims (data not shown), and the homozygous mutants are adult viable.

2.4.2 Zebrafish lrp4 is expressed in developing hindbrain, but is not required for

performance of acoustic startle response

Lrp4 is expressed in several regions of the mouse brain, localizes to post-synaptic structures in the central nervous system in mouse and fly, and loss of lrp4 causes cognitive defects in mouse and synaptic defects in fly (Tian et al., 2006; Gomez et al., 2014; Mosca et al., 2017). I therefore wondered whether lrp4 is expressed in developing zebrafish brain and is involved in synapse or circuit formation in the zebrafish central nervous system. Unpublished findings from Dr.

John Kuwada’s lab demonstrated that lrp4 is expressed in developing zebrafish hindbrain structures. To confirm this, I performed chromogenic in situ hybridization in whole-mount zebrafish embryos at 48 hpf and observed lrp4 expression in the developing hindbrain (data not shown), which is the region where the acoustic startle response circuitry and its command Mauthner neuron are localized (Eaton et al., 1977; for review please see 2001; Korn and Faber, 2005). The Mauthner neuron is required to perform the acoustic startle response, and in zebrafish larvae at 5 dpf, an acoustic startle manifests as a strong C-shaped turn away from the perceived stimulus and subsequent escape swims (Foreman and Eaton, 1993). This behavior is highly stereotyped across larvae, and our lab has developed software to detect slight differences in startle latency, turn angle, 21

angular velocity, startle threshold, and habituation to the acoustic stimulus (Wolman et al., 2011). I first assessed startle response execution in 5 dpf wild-type and lrp4 mutant zebrafish larvae, and I observed no differences in execution or latency (data not shown), which suggests that the reduction in synapses in lrp4 mutants does not impair the normal acoustic startle response. I next assayed habituation to the acoustic stimulus and observed that lrp4 mutants habituate like wild-type siblings with no obvious defects (Figure 2-5). Together these results indicate that zebrafish lrp4 is not required for the performance of the acoustic startle response, or assembly of the neural circuitry underlying habituation.

2.4.3 Zebrafish Dok-7 is not required for AChR pre-patterning or synapse formation

Like Lrp4, Dok-7 is required for AChR pre-patterning and synapse formation in mice

(Okada et al., 2006). A previous report in zebrafish using morpholinos to knock down dok-7 suggested that Dok-7 is required for synapse formation (Muller et al., 2010), however it was unclear whether a dok-7 mutant would recapitulate the morphant phenotype. Dok-7 is a cytoplasmic adaptor protein that has at least two bioactive domains at its amino terminal: the pleckstrin homology domain allows it to associate with the cell membrane, and the phosphotyrosine-binding domain allows it to bind and stabilize the phosphorylated Tyrosine residues in the kinase domain of MuSK. Continuous Dok-7 binding to MuSK is required to maintain MuSK activation (Okada et al., 2006; Inoue et al., 2009; Bergamin et al., 2010). Our lab previously showed that during the

AChR pre-pattern/synapse positioning stage, MuSK is endocytosed from the muscle cell membrane and trafficked into perinuclear Rab11-positive endosomes, but where Dok-7 localizes during this time remains unknown. Does Dok-7 stay bound to MuSK as it traffics, thereby furthering its activation, or does it become unbound from MuSK as soon as MuSK is endocytosed?

Furthermore, is Dok-7-mediated MuSK activation necessary to interact with Disheveled and the other PCP components, or is MuSK kinase activity dispensable for interacting with Disheveled?

While I planned to answer these questions during my dissertation, I did not advance far enough in this project, as you will read below. 22

To address my questions regarding Dok-7’s role in synapse positioning and synapse formation in zebrafish, I generated dok-7 zebrafish mutants using TALEN technology. I targeted exon 3 of dok-7, which falls near the end of the amino-terminal pleckstrin homology domain before the beginning of the phosphotyrosine-binding domain, and generated several alleles that all result in frameshifts and premature stop codons. In this chapter I discuss the results from one particular allele—a 23bp deletion resulting in p.S79Cfs*79—although all three alleles analyzed demonstrate the same synaptic phenotypes. The p189 allele causes a frameshift that is predicted to scramble the amino acid sequence halfway through the pleckstrin homology domain, eliminating the entire phosphotyrosine-binding domain and resulting in 79 random amino acids before a stop codon occurs (Figure 2-6A).

I stained dok-7 mutant zebrafish embryos at 19.5 hpf with α-bungarotoxin (α-BTX) and the znp-1 antibody, and found that AChR pre-patterned clusters were present in dok-7 mutant and wild- type sibling embryos (Figure 2-6B,C). I measured the width of the band of pre-patterned AChR clusters relative to the width of each hemisegment and found no difference between wild-type and dok-7 mutant embryos (Figure 2-6F; wild-type n = 18 hemisegments, 36.60% ± 1.85%; dok-7 mutant n = 18 hemisegments, 36.03% ± 2.04%; mean ± SEM, P = 0.83, unpaired Student’s t-test).

These findings suggest that, like for Lrp4, zebrafish use a Wnt/MuSK signaling mechanism, exclusive of Dok-7, for muscle pre-patterning. Unlike the situation with Lrp4, however, Dok-7 is surprisingly either not required or required to a lesser extent for neuromuscular synapse formation in zebrafish. In dok-7 mutants examined at 26 hpf, en passant synapse formation is variable. Every hemisegment examined shows wild-type-looking AChR clusters on the adaxial muscle pioneer cells at the horizontal myoseptum, but the clusters dorsal and ventral to the horizontal myoseptum show variable staining, with some muscle cells showing no clusters and others showing weak clusters (Figure 2-6E, arrowheads). However, this weak phenotype is not sufficient to cause visible swimming defects at any stage, demonstrating that synapses are functional. This might suggest that the dok-7p189 allele is a hypomorph, or that dok-7 is maternally contributed. Importantly, I 23

examined embryos from homozygous mutant mothers and they show the same variable phenotype, indicating that the synapses present are not due to maternal contribution of dok-7. These findings demonstrate that Dok-7 is not required for AChR pre-patterning, and may play a modest role in synapse formation in zebrafish.

2.5 Discussion

Why is Lrp4 dispensable for AChR pre-patterning in zebrafish but not mice? One obvious reason appears to be a shift in the timing of lrp4 expression in zebrafish and mice: lrp4 is abundantly expressed in mouse muscle during the pre-patterning stage (Weatherbee et al., 2006), whereas lrp4 expression in zebrafish is not detectable during AChR pre-patterning (Figure 2-1) (Remédio et al., 2016). Interestingly, the domain structure of MuSK, especially the extracellular domain, is less well conserved between mice and zebrafish (43% identity). MuSK is a single-pass transmembrane protein that contains a signal sequence, three Ig-like domains, a Frizzled-like CRD, a transmembrane domain, and an intracellular tyrosine kinase domain. Interestingly, non-mammalian

MuSK contains an additional Kringle domain between the CRD and the transmembrane domain

(Jennings et al., 1993; Zhang et al., 2004). Kringle domains are named after the pretzel-shaped

Kringle pastry because they form disulfide bridges and hydrogen bonds that give them a round, folded structure. Kringle domains are found throughout blood clotting and fibrinolytic proteins, and are thought to be involved in protein-protein interactions or in regulating proteolysis (Patthy et al.,

1984; Patthy, 1985; Castellino and Beals, 1987; Ikeo et al., 1991). What is the significance of the non-mammalian MuSK Kringle? I hypothesize that the Kringle domain physically blocks the interactions between Lrp4 and MuSK during AChR pre-patterning, so Lrp4 could not adopt a role in AChR pre-patterning until the Kringle was lost during vertebrate evolution.

Conveniently, the nucleotides that comprise the Kringle domain are a multiple of three, so eliminating this modular domain would not disrupt the MuSK open reading frame and perhaps would not be deleterious. I devised a strategy to test my hypothesis about the role of the Kringle

domain in pre-patterning using CRISPR-mediated excision of the Kringle domain from MuSK. I hypothesize that removing the Kringle domain from MuSK would disrupt AChR pre-patterning, so if I removed the Kringle in an embryo that also lacked lrp4, I predict I would see a reduction in AChR pre-patterning. Regrettably, I was too consumed by efforts to publish Chapter 3 of my thesis, and I ran out of time to test my hypothesis. I designed and validated sgRNAs that specifically excise the

Kringle domain (see Appendix Section 5.1 for more information about the status of the project, including sgRNA design and experimental strategy), so this would be an easy and interesting experiment for a future researcher to pursue.

I also showed that lrp4, while expressed in the developing zebrafish hindbrain, is not required for performance of the acoustic startle response or habituation to this response. What role, then, does lrp4 play in the hindbrain? It is possible that Lrp4 regulates synapse formation of other

Mauthner-independent neuronal circuits in the hindbrain. A future member of the lab could perform antibody stains for synaptic markers in larval zebrafish brains to look for changes or defects in lrp4 mutants. In the mouse CNS, lrp4 is expressed in the hippocampus, olfactory bulb, cerebellum, and neocortex, and is found in post-synaptic membranes and astrocytes (Tian et al., 2006; Sun et al.,

2016). Mouse lrp4 mutants demonstrate learning and memory defects as well as impairments in glutamatergic transmission that lead to locomotor and spatial memory defects (Gomez et al., 2014;

Sun et al., 2016). Recently, lrp4 has been shown to localize to pre-synaptic nerve terminals and regulate excitatory synapse number and olfactory attraction behaviors in Drosophila (Mosca et al.,

2017). It would be interesting to probe which cell types in the zebrafish brain require lrp4, and whether loss of lrp4 from these cell types changes specific behaviors.

How can we explain the more confusing finding that zebrafish Dok-7 is not required for

AChR pre-patterning or synapse formation? One possible explanation is that the dok-7 gene was duplicated during a probable teleost genome duplication event (Amores et al., 1998), so a second dok-7 gene may be compensating for the loss of the first gene. However, I have been looking for a

second dok-7 gene in the zebrafish genome for my entire graduate career, and a second gene has never been annotated, nor has a “predicted” second gene emerged. There is one uncharacterized gene in the zebrafish genome that contains a phosphotyrosine-binding domain highly homologous to the Dok-7 PTB domain (72% identity between PTB domains, 46% identity overall;

“uncharacterized protein” LOC100331748, XP_002663080.3), so it is possible that this uncharacterized gene is either an additional dok-7 gene, or a different gene altogether expressed in skeletal muscle whose PTB domain can bind and activate MuSK. The simplest explanation is that zebrafish Dok-7 does not act during synapse positioning or synapse formation; that like Lrp4,

Dok-7 acquired this function later in vertebrate evolution. Another possibility is that the TALEN allele I generated is not a null allele. The allele is a 23-bp deletion that shifts the open reading frame and encodes 89 random amino acids before a premature truncation. With the help of a post- baccalaureate student and technician, we isolated RNA from dok-7 mutants and siblings, made cDNA and sequenced it, and found the same genetic lesion, so we believe that the mutant mRNA is translated into a truncated, partly scrambled Dok-7 protein that either persists or is degraded, but is not predicted to bind the MuSK kinase domain.

My findings with Lrp4 and Dok-7, while surprising, are not wholly unexpected. These findings may be part of a larger story that there are significant differences in MuSK-dependent synapse formation between mammals and non-mammals. One area of intense investigation regards Wnt ligands and MuSK. We previously showed that in zebrafish Wnt11r and Wnt4a bind the MuSK Cysteine-Rich Domain; these Wnts and the MuSK CRD are required for AChR pre- patterning (Jing et al., 2009; 2010; Gordon et al., 2012). In mice, several Wnt ligands are capable of binding the MuSK CRD (Zhang et al., 2012), but whether Wnts and the MuSK CRD are required for AChR pre-patterning remains unresolved. Some reports suggest that in mouse both Wnt ligands and the MuSK CRD are required for AChR pre-patterning (Strochlic et al., 2012; Messéant et al.,

2015; 2017). Messéant et al. (2015) addressed the requirement for the MuSK CRD by generating a transgenic line overexpressing a MuSKΔCRD construct, and they only excised part of the CRD, 26

so it is possible this construct does not accurately represent an endogenous MuSK allele lacking the entire CRD. Indeed, a report that we collaborated with Dr. Steve Burden on suggests that neither Wnts nor the MuSK CRD are required for pre-patterning in mouse, by generating an endogenous allele of MuSK lacking the entire CRD (Remédio et al., 2016). The Wnt conflict may arise from mouse strain differences, as addressed by Messéant et al (2017). Clearly, the precise mechanism and role for Wnts in MuSK-dependent synapse formation in mice is still an outstanding question, but one that is being pursued by multiple groups. I cannot fully explain the reasons why

Lrp4 and Dok-7 play different roles in synapse formation in zebrafish and mice, but this is an interesting question that future researchers could pursue just like the Wnt/MuSK story.

2.6 Figures

Figure 2-1: Zebrafish Lrp4 is required for synapse formation but not AChR pre-patterning

(A) Sequences of wild-type zebrafish lrp4 exon 9 and sequences of lrp4p184 and lrp4p185, two frameshift alleles that were generated using TALENs targeting exon 9. (B) The protein domain structure of wild-type zebrafish Lrp4 (1899 amino acids) and the predicted structure of lrp4p184 (331 amino acids), which resembles the mouse mitt allele and is predicted to lack the β-propeller domains, the EGF-like domains, the transmembrane domain, and the intracellular region. (C,C′)

Lateral views of wild-type and lrp4 mutant zebrafish embryos at the pre-patterning stage (19.5 hours post-fertilization [hpf]) stained for AChRs (red) and motor axons (green). Lrp4 mutant embryos show no reduction or defect in AChR pre-patterning. n = 50 out of 50 embryos. (D,D′)

Lateral views of wild-type and lrp4 mutant zebrafish embryos at the synapse formation stage (26 hpf) stained for AChRs (red) and motor axons (green). In contrast to the pre-patterning stage, lrp4 mutant embryos show a significant reduction in AChRs clustered beneath the motor axon terminals, with only a few “hot spot” AChR clusters remaining at the horizontal myoseptum (white arrowheads); n = 50 out of 50 embryos. (E,E′) Lateral views of in situ hybridizations performed on wild-type zebrafish embryos at the pre-patterning stage (E) and synapse formation stage (E′). Lrp4 mRNA expression is undetectable at the pre-patterning stage (E and zoomed panel) but robustly expressed at later stages (E′ and zoomed panel), consistent with a requirement for zebrafish Lrp4 during synapse formation but not during pre-patterning.

Figure 2-2: AChR pre-pattern width is unaffected in lrp4 mutants

30 n.s.

10 AChR AChR pre-pattern width (% of hemisegment covered) hemisegment (% of 0 Siblings lrp4-/- (n=17 (n=30 hemisegments) hemisegments)

Quantification of AChR pre-pattern width in wild type siblings and lrp4 mutants at 19.5 hpf. I observed no difference in pre-pattern width, suggesting that lrp4 is not required for AChR pre- patterning in zebrafish (wild-type, n = 17 hemisegments, 20.17% ± 1.24%; lrp4 mutant, n = 30 hemisegments, 21.73% ± 1.16%; mean ± SEM, P = 0.4, unpaired Student’s t-test).

Figure 2-3: En passant neuromuscular synapses are absent in lrp4 mutants at 36 hpf

(A,A’) Lateral views of wild-type zebrafish embryo during the synapse formation stage when almost all en passant synapses have formed (36 hpf) stained for AChRs (red) and motor axons (green).

In wild type embryos, many en passant synapses form beneath the motor axon, enabling the embryo to escape a light touch stimulus. (B,B’) In contrast, lrp4 mutant zebrafish embryos show almost no AChR clusters beneath motor axon terminals, which manifests in their inability to fully articulate the tail and failure to escape a light touch.

Figure 2-4: Neuromuscular synapses are reduced and disordered in lrp4 mutants at 5 dpf

(A,A’) Lateral views of wild-type zebrafish larval hemisegment at 5 dpf when many focal and non- focal synapses have formed, stained for AChRs (red) and motor axons (green). In wild type embryos, many synapses have formed beneath the motor axons. (B,B’) In contrast, lrp4 mutant zebrafish embryos show fewer and less organized AChR clusters beneath motor axon terminals.

Figure 2-5: Lrp4 is not required for habituation to a loud acoustic stimulus

100 90 80 70 60 TLF LRP4 sibs 50 LRP4 muts 40

% Response% ennui muts 30 20 10 0 Tap 1-10 STH 1-10 STH 11-20 STH 21-30

Graph showing the percent of larvae that respond to 40 loud auditory stimuli (grouped into bins of

10). In the first bin, “Tap 1-10”, the stimulus is presented every 10 seconds, and the majority of larvae respond to every stimulus. In the second through fourth bins, the loud stimulus is presented every 1 second, and all genotypes of larvae habituate to the stimulus (seen here as a decline in percentage of larvae responding). Lrp4 and ennui mutants (a second allele of Lrp4; see Chapter 3 for more details) habituate like wild type larvae and siblings to the stimulus, indicating that lrp4 is not required for habituation to a loud acoustic stimulus.

Figure 2-6: Zebrafish Dok-7 is not required for synapse formation or AChR pre-patterning

(A) Sequences of wild-type zebrafish dok-7 exon 3 and sequences of dok-7p189, a frameshift allele that was generated using TALENs. Shown in gray is the protein domain structure of wild-type zebrafish Dok-7 (665 amino acids) and the predicted structure of dok-7p189 (157 amino acids), which is predicted to encode a truncated pleckstrin homology (PH) domain, and predicted to lack the phosphotyrosine binding (PTB) domain and the entire carboxy-terminus. (B,C) Lateral views of wild-type and dok-7 mutant zebrafish embryos at the pre-patterning stage (19.5 hours post- fertilization [hpf]) stained for AChRs (magenta) and motor axons (green). Dok-7 mutant embryos show no reduction or defect in AChR pre-patterning. n = 10 out of 10 embryos. (D,E) Lateral views of wild-type and dok-7 mutant zebrafish embryos at the synapse formation stage (26 hpf) stained for AChRs (magenta) and motor axons (green). In contrast to the pre-patterning stage, dok-7 mutant embryos show variable AChR clusters beneath the motor axon terminals, with wild-type- looking AChR clusters at the horizontal myoseptum, and variable clusters along dorsal and ventral muscle cells (white arrowheads; n = 11 out of 11 embryos). (F) Quantification of AChR pre-pattern width in wild type siblings and dok-7 mutants at 19.5 hpf. I observed no difference in pre-pattern width, suggesting that dok-7 is not required for AChR pre-patterning in zebrafish (wild-type n = 18 hemisegments, 36.60% ± 1.85%; dok-7 mutant n = 18 hemisegments, 36.03% ± 2.04%; mean ±

SEM, P = 0.83, unpaired Student’s t-test).

CHAPTER 3: THE SYNAPTIC RECEPTOR LRP4 PROMOTES PERIPHERAL NERVE REGENERATION

Modified from Gribble KD, Walker LJ, Saint-Amant L, Kuwada JY, Granato M. (anticipated 2017). The synaptic receptor Lrp4 promotes peripheral nerve regeneration. Under Revision at Nature Communications.

3.1 Abstract

Early during PNS regeneration, regenerating axons emerge from the proximal nerve stump yet whether they extend simultaneously or whether pioneering axons establish a path for follower axons remains unknown. Moreover, the molecular mechanisms underlying robust regeneration are not fully understood. Using live imaging we demonstrate that in zebrafish pioneering axons establish a regenerative path for follower axons. We find that this process requires the synaptic receptor lrp4 and that in lrp4 mutants pioneers are unaffected while follower axons frequently stall at the injury gap, providing evidence for molecular diversity between pioneering and follower axons in regeneration. We demonstrate that Lrp4 promotes regeneration independent of membrane anchoring and of MuSK co-receptor signaling essential for synaptic development. Finally, we show that Lrp4 coordinates the realignment of denervated Schwann cells with regenerating axons, consistent with a model by which Lrp4 is repurposed to promote sustained peripheral nerve regeneration via axon-glia interactions.

3.2 Introduction

Axons of the peripheral nervous system can regenerate after injury (y Cajal, 1928), yet the molecular mechanisms that promote robust nerve regeneration are not fully understood. After sustaining an injury, peripheral nerves initiate the program of Wallerian degeneration that causes self-destruction of distal axons (Waller, 1850). Distal axon debris is subsequently removed by macrophages and Schwann cells (Scherer and Easter, 1984; Stoll et al., 1989; Fernandez-Valle et al., 1995; Perry et al., 1995; Vargas and Barres, 2007), clearing the path along which axons can

regrow. Axonal regeneration begins when growth cones sprout from the proximal nerve stump and stabilize into growing axons, and the current view is that denervated Schwann cells in the distal nerve stump become activated and provide diffusible factors including NGF, BDNF, GDNF and

FGF that promote growth cone sprouting as well as axonal growth and guidance (Taniuchi et al.,

1986; Heumann et al., 1987; Meyer et al., 1992; Funakoshi et al., 1993; Liu et al., 1995; Zhang et al., 2000; Scarlato et al., 2001; Höke et al., 2002). Although there is evidence that axonal regrowth is staggered (Al-Majed et al., 2000; McDonald et al., 2005), it is unclear whether axons emerge in waves from the nerve stump and fan out in search of their original trajectory, or whether a limited number of axons pioneer a path that later-emerging follower axons then fasciculate with to traverse the injury site and grow back towards their original targets.

Shortly after sprouting from the proximal nerve stump, regenerating axons grow towards and along denervated/activated Schwann cells (Dyck and Hopkins, 1972; Scherer and Easter,

1984; Nguyen et al., 2002; Chen et al., 2005). Schwann cells realign with regenerating axons, and their morphology changes dramatically as they revert from an activated, regeneration-supporting

Schwann cell to a pre-injury myelinating Schwann cell (y Cajal, 1928; Holtzman and Novikoff, 1965;

Scherer and Easter, 1984; Cheng and Zochodne, 2002). Several molecular pathways critical for

Schwann cells to transition from a myelinating Schwann cell to an activated, denervated and more immature Schwann cell have been documented (Scherer, 1997; Le et al., 2005; Jessen and Mirsky,

2008; Parkinson et al., 2008; Woodhoo et al., 2009; Arthur-Farraj et al., 2012). In contrast, the mechanisms underlying the realignment of denervated Schwann cells with regenerating axons, and the mechanisms that trigger this pronounced change in Schwann cell morphology as they revert to a pre-injury myelinating Schwann cell are not well understood.

Here we use live cell imaging in larval zebrafish and demonstrate that upon complete peripheral nerve transection, individual axons emerging from the proximal stump pioneer a path across the injury gap. Later-emerging axons fasciculate with these pioneer axons to cross the injury

gap and to return towards their original synaptic targets. We find that this process requires the synaptic low-density lipoprotein receptor-related protein 4 (Lrp4), and that in lrp4 null mutants

(Remédio et al., 2016) pioneer axons are unaffected while follower axons frequently fail to cross the injury gap and stall. Moreover, we show that lrp4 promotes regeneration independently from its membrane anchor and without signaling through the canonical Lrp4 signaling receptor MuSK.

Instead, lrp4 coordinates the realignment of regenerating axons with denervated Schwann cells.

Together, our findings demonstrate the existence of and the molecular diversity between axonal pioneers and followers in regeneration, and reveal an unexpected in vivo function for lrp4 in peripheral nerve regeneration.

3.3 Experimental Procedures

3.3.1 Zebrafish strains and animal care

Transgenic lines were generated in the Tübingen or TLF genetic background, and all fish used were maintained as previously described (Mullins et al., 1994). The following transgenic lines were used: Tg(mnx1:GFP)ml2 (Flanagan-Steet et al., 2005), Tg(sox10:mRFP)vu234 (Kucenas et al.,

2008), Tg(mnx1:Lrp4-GFP), Tg(α-actin:Lrp4-GFP). The following mutant strains were used: lrp4p184

(Remédio et al., 2016), lrp4mi36, textually referred to as lrp4ennui (Saint-Amant et al., 2007), musktbb72

(Granato et al., 1996; Zhang et al., 2004; Lefebvre et al., 2007), twoth26e, textually referred to as rapsyn (Granato et al., 1996; Ono et al., 2002), and agrinp190. We conducted all experiments using zebrafish according to animal protocols that were approved by the University of Pennsylvania

IACUC regulatory standards.

3.3.2 Molecular identification of ennui gene

By using recombination mapping and PCR-scorable, polymorphic CA repeats the ennui mutation was mapped to a 1cM region on chromosome 7. Examination of over 2,500 meiotic events identified a single non-recombinant CA repeat marker markers in an intron of the lrp4 gene.

Subsequent RT-PCR analysis using mRNA from both wildtype and ennui embryos, revealed a single point mutation in the lrp4 open reading frame that converted Tyrosine 1398 to a stop codon.

This mutation was confirmed by sequencing PCR-amplified genomic DNA in the region of the mutation from wildtype and mutant embryos.

3.3.3 Generation of agrinp190 mutants

The zebrafish agrin (XM_009296860.2) exon 31 splice donor site was targeted for

CRISPR-mediated mutagenesis in order to generate exclusively Z- agrin transcripts in vivo, which has been shown to eliminate the AChR clustering ability of Agrin (Burgess et al., 1999). To generate templates for sgRNA transcription, 16-bp complementary agrin-specific oligonucleotides were ordered from IDTDNA, with BsaI overhangs (underlined) added to allow the oligos to be annealed and then ligated into pDR274 (Forward agrin oligo: 5’ TAGGCAGAGGAGAGGCGCCGGT 3’;

Reverse agrin oligo: 5’ AAACACCGGCGCCTCTCCTCTG 3’) (Hwang et al., 2013). The sgRNAs in pDR274 were transcribed using the T7 MegaShortScript kit (Ambion). Fifty picograms sgRNA and 80 pg Cas9 protein (PNA Bio) were co-microinjected into one-cell stage TLF embryos. Double- stranded DNA breaks in agrin exon 31 of G0 injected embryos were confirmed by PCR and high- resolution melt analysis, and G0 embryos were raised. Heterozygous F1 carrier alleles were identified and characterized by PCR and sequencing, and transcripts were characterized by sequencing from pools of mutant and sibling cDNA.

3.3.4 Spinal motor nerve transection assays

Complete MicroPoint (Andor Technology) laser-mediated transection and live-cell imaging of peripheral motor nerves was performed as previously described (Rosenberg et al., 2012; 2014;

Isaacman-Beck et al., 2015). Nerve regeneration scoring at 48 hpt was performed as follows: image brightness and contrast were altered in the same way for every image across all experiments, and

48 hpt images were anonymized for blind scoring. Using our previously-established regeneration scoring rubric (Rosenberg et al., 2014), we assigned a regeneration score to each transected nerve

based on the number of regenerated fascicles that regrew to the ventral end of the hemisegment.

“Best” regeneration indicates two or more distinct regenerated fascicles; “moderate” indicates one regenerated fascicle; “worst” indicates zero regenerated fascicles that regrow to the ventral end of the hemisegment. Images were subsequently un-blinded and regeneration scores were recorded in GraphPad Prism for statistical analysis. Fisher’s exact tests or Chi-square tests were used to compare the extent of regeneration between different genotypes.

3.3.5 Sparse neuronal labeling

A DNA vector encoding mnx1:mKate DNA was injected into one-cell-stage

Tg(mnx1:GFP)ml2 embryos as previously described (Thermes et al., 2002). Embryos were sorted for mKate signal in individual motor neurons at 24 hpf, and GFP-expressing motor nerves containing a small subset of mKate-expressing axons were transected at 5 dpf and imaged as described above.

3.3.6 Whole mount immunohistochemistry and imaging

To label acetylcholine receptor clusters and motor axons, zebrafish embryos at 26-28 hpf were stained with fluorescently-conjugated ɑ-bungarotoxin and the znp-1 antibody (Trevarrow et al., 1990), and images were captured as previously described (Remédio et al., 2016). An anti-

Sox10 antibody (1:2000; gift from S. Kucenas) was used to label Schwann cell nuclei in 5 dpf and

6 dpf larvae as previously described (Rosenberg et al., 2014). Larvae were mounted in Vectashield

(Vector Laboratories), and imaged in 1µm sections using a 60x objective on a Zeiss LSM 710 confocal microscope. Image stacks were adjusted for brightness and contrast in Fiji, and Schwann cell nuclei were manually counted in each z-plane of an image. The results were recorded in

GraphPad Prism for statistical analysis: two-tailed t-tests were performed to compare Schwann cell numbers pre- and post-transection.

3.4 Results

3.4.1 Pioneering axons establish a regenerative path for follower axons

During development, subsets of neurons extend pioneering axons that outline the trajectory for later emerging follower axons (Bate, 1976; Raper et al., 1983; Eisen et al., 1986; Pittman et al.,

2008; Raper and Mason, 2010; Kim et al., 2011). Whether regenerating axons employ a similar strategy of pioneer and follower axons or whether they emerge from the proximal nerve stump in waves and fan out searching for their original trajectory is currently unknown. We have previously shown that laser-mediated transection of spinal motor nerves in 5 day post-fertilization (dpf) zebrafish larvae results in robust axonal regeneration within 48 hours post-transection (Rosenberg et al., 2012; 2014; Isaacman-Beck et al., 2015). Importantly, we find that regeneration in larval zebrafish is characterized by key features of vertebrate peripheral nerve regeneration, including the ability of Wlds to delay axonal fragmentation and the dependence on Schwann cells for successful regeneration (Rosenberg et al., 2012; 2014).

Spinal motor nerves consist of ~60 individual axons, and to examine whether regenerating axons emerge simultaneously or whether they emerge in a temporal sequence, we transected individual spinal nerves in the mnx1:GFP transgenic line in which all spinal motor axons express

GFP. Time-lapse imaging revealed that starting around 9 hours post-transection (hpt) between two and six axon sprouts emerge from the proximal stump, rapidly extending and retracting

(Supplemental Movie 2). Around 11 hpt, what appeared to be a single axon stabilized and began to pioneer a path across the transection gap (Figure 3-1B magenta arrowhead and ‘P’). At around

18 hpt, the first ‘follower’ axon (Figure 3-1C green arrowhead and ‘1’) emerged from the proximal stump and grew alongside the pioneer axon. Within the next 3-4 hours up to three additional axons emerged and grew along the earlier axons (average 2.2 followers; Figure 3-1D,E,R and

Supplemental Movie 1) across the injury gap and towards their original targets. We noticed that the

early, pioneering axons extend at an average rate of 0.24 µm per minute, while later emerging follower axons extend almost twice as fast (0.47 µm per minute; Figure 3-1P).

To confirm that indeed a single axon pioneered the regenerative path, we used a sparse labeling strategy (Downes et al., 2002; Flanagan-Steet et al., 2005) to genetically label individual neurons and their axons with mKate (magenta, Figure 3-1F) in the background of the mnx1:GFP transgenic line labeling all motor axons. Prior to nerve transection, we selected spinal segments in which only one motor neuron and its axon was labeled with mKate, thereby ensuring single axon resolution (magenta, Figure 3-1F). In the majority of cases, we observed an individual mKate- positive axon extending along GFP-positive axons that had emerged from the proximal nerve stump earlier. In a small number of instances individual mKate-positive axons were the first to emerge from the proximal nerve stump and pioneered a regenerative path, followed by GFP-only expressing axons (Figure 3-1H-J, Supplemental Movie 2). Thus, regenerating axons segregate into a larger pool of fast growing follower axons that tightly fasciculate with a single axon that successfully pioneered a path across the injury site.

3.4.2 lrp4 promotes robust nerve regeneration

We next sought to gain insights into the molecular mechanisms that promote peripheral nerve regeneration. For this we examined a collection of existing mutant lines harboring mutations in genes critical for axonal growth and guidance, as well as in synapse formation. This revealed an unexpected role for the synaptic low-density lipoprotein receptor-related protein 4 (Lrp4). Lrp4 encodes a single-pass transmembrane domain protein best known for its evolutionarily conserved role in neuromuscular synapse development (Weatherbee et al., 2006; Remédio et al., 2016). Live cell imaging following nerve transection revealed that like in wild type animals (Figure 3-1A-J and

Supplemental Movies 1 and 2), pioneer axons in lrp4 mutants emerged from the proximal nerve stump starting around 10 hpt, and navigated over the transection gap and into the ventral myotome with growth rates indistinguishable from those in wild type siblings (Figure 3-1K-O,Q and

Supplemental Movie 3). In contrast to wild type animals, follower axons in lrp4 mutants remained in a state of exploratory extension and retraction, ultimately stalling at the injury gap (Figure 3-1K-

O,R; open blue arrowheads; mean number of followers observed during imaging window=0.7 axons; n=7 nerves). Consequently, this reduced the number of follower axons that crossed the injury gap within the first 20 hpt by more than threefold, from an average of 2.2 to 0.7 follower axons per nerve (Figure 3-1R). Importantly, motor axonal outgrowth during development is unaffected in lrp4 mutants (Figure 3-5) (Remédio et al., 2016), demonstrating a specific role for Lrp4 during axonal regeneration. Thus, while lrp4 appears dispensable for regenerative growth of pioneer axons, it is critical for follower axons to robustly regrow across the injury gap and towards their original targets.

Finally, we wondered whether the reduced ability of follower axons to cross the injury gap was transient or diminished the overall robustness of peripheral nerve regeneration. Prior to nerve transection at 5 days post-fertilization (dpf), individual motor nerves in wild type and lrp4 mutants are comprised of about 60 axons organized into 3 to 4 major fascicles that extend into the ventral myotome (Figure 3-2A,B) (Rosenberg et al., 2012; 2014). By 48 hours post-transection, wild type axons have robustly regenerated and form several fascicles that extend back towards their original targets in the ventral myotome (Figure 3-2A’’ and quantified in Figure 3-2C). In contrast, lrp4 mutants display a marked reduction in the number of axon fascicles that have reached the ventral myotome, concomitantly with a marked increase of fascicles stalling prematurely (Figure 3-2B’’; open blue arrowheads; quantified in Figure 3-2C). Thus, lrp4 plays a critical function for follower axons early during the regeneration process to cross the injury gap, thereby promoting robust peripheral nerve regeneration in vivo.

3.4.3 lrp4 coordinates the realignment of denervated Schwann cells with regenerating

axons

Schwann cells and perineural glia cells have both been shown to play critical roles in peripheral nerve regeneration (Ide et al., 1983; Hall, 1986; Lewis and Kucenas, 2014). For example, in rodents EphB signaling promotes Schwann cell migration and axonal regrowth across the injury site (Parrinello et al., 2010). Similarly, in zebrafish lacking Schwann cells, regenerating axons sprout from the proximal nerve stump but fail to grow across the injury gap (Rosenberg et al., 2014), reminiscent of the phenotype we observe in lrp4 mutants. Given the prominent role of Schwann cells in peripheral nerve regeneration, we examined Schwann cell numbers and morphology before and after injury. We found that in lrp4 mutants the number of Sox10-positive Schwann cells associated with individual motor nerves before and 24 hours post-transection was equivalent to those in wild type animals (Figure 3-3A-D), strongly arguing that Lrp4 is dispensable for Schwann cell development and survival.

We then examined whether lrp4 is important for the stereotyped morphological changes associated with injury-induced Schwann cell activation. For this we utilized a combination of stable transgenic lines in which both motor axons and Schwann cell membranes are labeled (Kucenas et al., 2008). We have previously shown that before injury Schwann cell membranes ensheath individual motor axons (Rosenberg et al., 2012), and that following nerve transection when axons start to fragment, Schwann cell membranes reorganize, changing from a smooth, tube-like appearance to a more rounded and granular morphology, indicative of their transition to an activated, dedifferentiated state (Figure 3-3F) (Rosenberg et al., 2014). Prior to injury, Schwann cell morphology in lrp4 mutants was indistinguishable from that in wild type siblings (Figure 3-3E,I).

Live cell imaging between 10 and 24 hpt revealed that in lrp4 mutants Schwann cell morphology changed upon injury indistinguishable from the process in wild type siblings (Figure 3-3F,J; yellow arrowheads). In wild type larvae Schwann cell membranes re-extend and become smoother to resemble their pre-transection morphology around 19 hpt, just as regenerating axons are 44

contacting the Schwann cells (Figure 3-3G,H, Supplemental Movie 4; Schwann cell membranes marked by yellow arrowheads, axons marked by green arrowheads; n=4 out of 4 nerves). In contrast, in lrp4 mutants Schwann cells appear unable to transition back to their pre-injury morphology and instead retain a rounded and granular morphology, even despite apparent contacts with a pioneering axon (Figure 3-3K,L, Supplemental Movie 5; Schwann cell membranes marked by yellow arrowheads, axons marked by green arrowheads; n=7 out of 8 nerves). Thus, lrp4 is required during the regeneration process to revert the morphology of an activated, more immature Schwann cell to a more differentiated, myelin-producing Schwann cell.

3.4.4 Lrp4 promotes peripheral nerve regeneration independently of MuSK signaling

We next wanted to decipher the molecular mechanisms by which Lrp4 promotes peripheral nerve regeneration, and one obvious signaling pathway is through the MuSK receptor tyrosine kinase which is required for vertebrate neuromuscular synapse development (DeChiara et al.,

1996). There, Lrp4 expressed on the surface of skeletal muscle cells functions as the receptor for motor axon-derived Agrin, and upon Agrin binding Lrp4 activates the muscle-specific kinase

(MuSK) receptor, initiating downstream signaling through Rapsyn eventually leading to the accumulation of AChR clusters underneath the nerve terminal (Figure 3-4G) (Gautam et al., 1995;

Glass et al., 1996; Apel et al., 1997; Weatherbee et al., 2006; Kim et al., 2008; Zhang et al., 2008;

2011; Zong et al., 2012). Because of its role in synapse development as an obligate signaling receptor for Lrp4, we tested whether MuSK is required for peripheral nerve regeneration. For this we examined peripheral nerve regeneration in musk null mutants (Zhang et al., 2004; Lefebvre et al., 2007). Endpoint analysis of peripheral nerve regeneration at 48 hpt did not reveal any significant differences between musk mutants and wild type siblings (Figure 3-4A-B,E). During synapse development Agrin acts as the ligand for Lrp4, so we tested whether Agrin is required for peripheral nerve regeneration. For this we generated agrin mutants that specifically lack the C-terminal exons critical for neuromuscular synapse formation, which show a strong developmental synapse defect similar to lrp4 mutants (Ferns et al., 1992; 1993; Hoch et al., 1993; Gesemann et al., 1995; Burgess 45

et al., 1999) (Figure 3-8). Endpoint analysis of peripheral nerve regeneration at 48 hpt did not reveal any significant differences between agrin mutants and wild type siblings (Figure 3-4C-D,F). Finally,

Rapsyn is a key downstream effector of MuSK signaling (Gautam et al., 1995; Gillespie et al., 1996;

Granato et al., 1996; Apel et al., 1997; Ono et al., 2002), and analysis of rapsyn mutants failed to reveal deficits in peripheral nerve regeneration (sibling n=9 nerves from 7 larvae; mutant n=11 nerves from 8 larvae; Fisher’s exact test p=0.3742; data not shown). Thus, lrp4 promotes regeneration through an Agrin- and MuSK-independent signaling pathway.

3.4.5 A truncated form of Lrp4 promotes nerve regeneration

In addition to its role as a membrane bound MuSK co-receptor, Lrp4 has also been shown to function in bone development, where it modulates Wnt signaling in part by releasing its extracellular domain, and hence sequestering Wnt antagonists (Choi et al., 2009; Dietrich et al.,

2010). Moreover, several reports have shown the Lrp4 ectodomain can be cleaved and that it retains biological activity (Dietrich et al., 2010; Wu et al., 2012; Yumoto et al., 2012; Choi et al.,

2013). We therefore wanted to explore the possibility that Lrp4 promotes peripheral nerve regeneration through an anchorage-independent mechanism. Positional cloning of the ennui mutant (Saint-Amant et al., 2007) revealed that the mutant phenotype is caused by a premature stop codon in Lrp4, just before the transmembrane domain (amino acid 1398 out of 1899), generating a truncated protein lacking membrane anchoring and the intracellular domain while leaving intact almost the entire extracellular domain (Figure 3-5A). We first determined if and to what degree this truncated Lrp4 protein had retained activity to induce neuromuscular synapses.

For this we compared post-synaptic AChR cluster localization between wild type, lrp4 null mutants and lrp4ennui mutant embryos at 26 hpf when motor axons have formed en passant neuromuscular synapses with muscle cells (Figure 3-5B-D) (Saint-Amant et al., 2007). This revealed a very strong reduction of postsynaptic AChR clusters in lrp4ennui mutants, identical to what we had previously reported for lrp4 null mutant animals (Figure 3-5C-D) (Remédio et al., 2016). In contrast to the almost complete inability of lrp4ennui to function in neuromuscular synapse development, peripheral 46

nerve regeneration at 48 hpt was indistinguishable from that in wild type animals, and in stark contrast to the phenotype observed in lrp4 null mutants (Figure 3-5E-G). Thus, a truncated form of

Lrp4 lacking the transmembrane and cytoplasmic domains that lacks activity as a MuSK co- receptor is sufficient to promote regeneration.

3.4.6 Lrp4 does not act in motor neurons or muscle cells to promote axon regeneration

Finally, we wondered which cell type Lrp4 acts in to promote axon regeneration. During development, lrp4 is expressed and required in muscle cells to bind MuSK and promote synapse formation (Weatherbee et al., 2006; Remédio et al., 2016). Based on its developmental role and our findings about regenerating pioneer and follower axons in lrp4 mutants, we hypothesized that lrp4 may act in muscle cells or motor neurons to promote regeneration. To determine the cell-type requirement for lrp4 during regeneration, we expressed rescue transgenes in lrp4 mutants that drive Lrp4-GFP under control of either skeletal muscle-specific [Tg(α-actin:Lrp4-GFP)] or motor neuron-specific [Tg(mnx1:Lrp4-GFP)] promoters. We restored lrp4 expression to motor neurons or muscle cells and then assessed regeneration, and we observed no significant improvements in motor axon regeneration at 48 hours post-transection with rescue in either cell type (Figure 3-6).

This suggests that lrp4 expression in motor neurons or muscle cells is not sufficient to restore wild- type axon regeneration, and suggests that lrp4 is required in another cell type or in a combination of cell types during regeneration. Quantification of regeneration between lrp4 mutants (n=9 nerves) and lrp4 mutants expressing a motor neuron-specific transgene (n=15 nerves) shows no significant differences (Figure 3-6C; Chi-square test p=0.9266). Quantification of regeneration between lrp4 mutants (n=21 nerves) and lrp4 mutants expressing a muscle-specific transgene (n=27 nerves) shows no significant differences (Figure 3-6F; Chi-square test p=0.8505). As a control to verify that

Lrp4-GFP protein is functional, we asked whether Lrp4-GFP in skeletal muscle cells can rescue neuromuscular synapses during development. We see Lrp4-GFP expressed in puncta throughout all muscle cells (Figure 3-7C). Expressing Lrp4-GFP in skeletal muscles of wild-type embryos does not disrupt neuromuscular synapses at 26 hpf (Figure 3-7A-C). Expressing Lrp4-GFP in lrp4 47

mutants fully rescues the neuromuscular synaptic defects observed in lrp4 mutants (Figure 3-7G-

I), indicating that Lrp4-GFP is a functional protein that can rescue loss of lrp4 in zebrafish embryos.

Together our findings demonstrate that Lrp4 acts either in an unknown cell type (perhaps in

Schwann cells) or multiple cell types at once to promote axon regeneration. Combined with our findings that Lrp4 acts independently of MuSK signaling, this supports a model by which Lrp4 is repurposed independently from its role in synapse development to promote sustained axonal regeneration and regulate Schwann cell plasticity, possibly via interactions between these two cell types.

3.5 Discussion

3.5.1 Lrp4 promotes peripheral nerve regeneration independently of its role in

development

Lrp4 function has been studied most extensively in its role as the Agrin receptor expressed on skeletal muscle cells where Lrp4 signals through its obligate co-receptor MuSK (DeChiara et al., 1996; Kim et al., 2008; Zhang et al., 2008). Here, we provide compelling in vivo evidence that

Lrp4 promotes peripheral nerve regeneration independently of MuSK, Agrin, and its downstream effector Rapsyn (Figure 3-4), all critical for neuromuscular synapse development in mammals and also in zebrafish (Gautam et al., 1995; DeChiara et al., 1996; Gautam et al., 1996; Glass et al.,

1996; Ono et al., 2001; Zhang et al., 2004; Lefebvre et al., 2007). Moreover, by using a mutant lrp4 allele that lacks membrane anchoring and the intracellular domain but leaves the ectodomain mostly intact, we demonstrate that in zebrafish Agrin-dependent synapse formation requires membrane-bound Lrp4 to an even greater degree than in mouse mutants with an allele that encodes only the Lrp4 ectodomain (Choi et al., 2013). We also show that restoring lrp4 specifically to muscle cells is not sufficient to rescue axon regeneration (Figure 3-6), although it rescues lrp4 mutant developmental synaptic defects (Figure 3-7). Collectively these findings strongly argue that

Lrp4 promotes regeneration largely independent of its well-defined signaling pathway critical for neuromuscular synapse development.

How could Lrp4 promote peripheral nerve regeneration? Lrp4 is a member of the large family of low-density lipoprotein receptor-related proteins with distinct roles inside but also outside the nervous system (Tian et al., 2006; Choi et al., 2009; Yumoto et al., 2012; Ahn et al., 2013; Sun et al., 2016). Outside the nervous system, Lrp4 plays critical roles in bone, tooth and limb development, primarily by modulating the Wnt and BMP signaling pathways (Choi et al., 2009; Ahn et al., 2013), and both the BMP antagonist Wise and the Wnt antagonist Dickkopf-1 bind Lrp4

(Ohazama et al., 2008; Choi et al., 2009). While Wnt signaling has been shown to play a functional role in CNS regeneration (Liu et al., 2008; Hollis and Zou, 2012; Strand et al., 2016) its role in peripheral nerve regeneration is not well understood, although a recent report suggests that Wnt/β-

Catenin signaling regulates Schwann cell proliferation and migration after sciatic nerve injury (Han et al., 2017). Similarly, BMP signaling can promote axonal regeneration, although most studies have focused on spinal cord injury rather than peripheral nerve injury (Parikh et al., 2011; Zhong and Zou, 2014). Existing zebrafish mutants for both signaling pathways (Mullins et al., 1996; Kuan et al., 2015) will be invaluable in establishing potential roles for BMP and/or Wnt signaling in Lrp4- dependent peripheral nerve regeneration.

3.5.2 Lrp4 reveals molecular diversity between pioneering and follower axons in

regeneration

During development, pioneer neurons form the initial axonal scaffold used by later outgrowing axons. Pioneering neurons have been identified in many organisms ranging from sensory neurons in grasshopper embryos to subplate neurons in the mouse cerebral cortex (Bate,

1976; McConnell et al., 1989; Raper and Mason, 2010). Sequential axonal outgrowth is thought to simplify the many challenges of axonal pathfinding by enabling the majority of axons to simply extend along a preexisting axonal path. While in some cases pioneer axons appear largely

dispensable for follower axons to find their targets (Pike et al., 1992), in many cases pioneering axons are necessary for follower axons to innervate their appropriate targets (Klose and Bentley,

1989; Hidalgo and Brand, 1997; Pittman et al., 2008).

Yet despite their importance in development, whether pioneer axons play a role in regeneration or even whether regenerating axons are organized into pools of pioneer and follower axons is not well understood. We find that after complete nerve transection and a period of intense neurite extension and retraction (Supplemental Movie 2, open blue arrowheads), a single axon emerges from the proximal nerve stump to pioneer a regenerative path across the injury site and back to the ventral myotome. Multiple later emerging axons join the pioneering axon and extend with about twice the speed of the pioneering axon across the injury gap, consistent with the idea that pioneering axons provide a growth-promoting, adhesive substrate. In lrp4 mutants pioneer axons successfully navigate across the injury gap and towards their original targets while follower axons appear unable to join the pioneering axon and frequently fail to cross the injury gap, eventually leading to reduced nerve regeneration (Figure 3-1). Based on these results it is tempting to speculate that during regeneration pioneer axons are important to promote the growth of later emerging follower axons, and that lrp4 produced by either pioneer or follower axons is critical for this interaction. However, we show that restoring lrp4 to motor neurons is not sufficient to rescue axon regeneration, suggesting that motor axon-derived Lrp4 does not promote regeneration

(Figure 3-6). Independent of the underlying mechanisms, our results demonstrate that like in development the majority of regenerating axons take advantage of and extend along a preexisting path, consistent with the idea that developmental strategies are reutilized during regeneration.

Importantly, in development axonal outgrowth of pioneering primary motor axons and follower secondary motor axons is unaffected in lrp4 mutants (Remédio et al., 2016). Thus, our analysis reveals that regenerating axons are organized into temporally segregated pools of pioneer and follower axons, and uncovers lrp4-dependent functional heterogeneity between pioneering and follower axons dedicated to regeneration. 50

3.5.3 Does Lrp4 coordinate axonal regeneration with Schwann cell plasticity?

A hallmark of peripheral nerve regeneration is the remarkable plasticity of Schwann cells.

Already described by Santiago Ramon y Cajal (1928), upon nerve injury Schwann cells undergo a reversible transformation characterized by extensive morphological changes. The initial transformation into a denervated, activated Schwann cell occurs in response to nerve injury, and reflects the conversion into a more immature cell state. Activated Schwann cells are characterized by the loss of myelin sheaths, reactivation of developmental genes and their ability to divide and migrate (Holtzman and Novikoff, 1965; Scherer and Easter, 1984; Cheng and Zochodne, 2002).

While this first transformation has been studied extensively, the molecular mechanisms that trigger the morphological transformation of denervated, activated Schwann cells back to their pre-injury appearance is not well understood. We show that in lrp4 mutants Schwann cell development and injury-induced transformation into an activated state are unaffected, strongly arguing that lrp4 is dispensable for the general molecular mechanisms and cellular forces underlying Schwann cell plasticity. Unlike in wild type animals, Schwann cells in lrp4 mutants fail to return to their original morphology despite the presence of pioneering axons making apparent contacts with denervated

Schwann cells (Figure 3-3L). This strongly argues against the notion that contacts between regenerating axons and denervated Schwann cells are sufficient to trigger changes in Schwann cell morphology characteristic for their re-differentiation.

The analysis of lrp4 mutants reveals a clear defect in the ability of denervated Schwann cells to resume their more differentiated morphology, consistent with several interpretations. One interpretation is that the Schwann cell phenotype is simply a secondary consequence of the reduced number of regenerating axons. While we cannot exclude this possibility, we favor an alternative possibility that lrp4 functions in glial cells, specifically in denervated Schwann cells to coordinate changes in Schwann cell morphology with axonal regeneration. There is precedent for lrp4 functioning in astroglia where lrp4 regulates synaptic transmission (Sun et al., 2016). Moreover, during development lrp4 functions not only in muscle cells as a receptor for postsynaptic 51

differentiation but the lrp4 ectodomain also acts as a muscle-derived retrograde signal that binds preferentially to the distal portion of motor axons (Tian et al., 2006; Choi et al., 2009; Yumoto et al.,

2012; Ahn et al., 2013; Sun et al., 2016). Such bidirectional signaling might also occur during axonal regeneration: as pioneering axons cross the injury gap and contact denervated Schwann cells this might activate Lrp4 signaling within Schwann cells, triggering the morphological changes associated with re-differentiation to a more mature ensheathing Schwann cell.

Concomitantly, contacts between the pioneering axon and denervated Schwann cells might trigger cleavage of the lrp4 ectodomain from the Schwann cell surface and like in development the lrp4 ectodomain might bind to axons to act as a retrograde signal. Binding of the lrp4 ectodomain to pioneering motor axons could trigger a mechanism to make them a growth- promoting, adhesive substrate for follower axons. Such a scenario would explain why in lrp4 null mutants pioneering axons regrow while follower axons fail to extend along the pioneer axon and instead frequently stall. It would also be consistent with our findings that in lrp4ennui mutants encoding only the ectodomain, regeneration is unaffected (Figure 3-5). Independently of the underlying mechanism, our results uncover a novel role for lrp4 in promoting peripheral nerve regeneration. Moreover, we demonstrate that Lrp4 function has been co-opted in regeneration independently of the well-documented MuSK signaling cascade in neuromuscular synapse formation. Finally, we uncover the existence of pioneer and follower axons in peripheral nerve regeneration, and demonstrate differential sensitivity to lrp4 function, revealing the existence of molecular diversity between regenerating axons.

3.6 Figures

Figure 3-1: Pioneering axons establish a regenerative path for follower axons

(A-E) Still images from a time lapse movie showing early regeneration of a wild type motor nerve in which all axons express GFP. (A) Pre-lesion image, with transection site marked by yellow box.

(B) A pioneering axon (magenta arrowhead and ‘P’) extends and crosses the injury gap ~800 minutes post-injury and extends ventrally. (C,D,E) Three follower axons fasciculate with pioneering axon over the course of imaging (numbered green arrowheads). (F-J) Still images from a time lapse movie showing early regeneration of a 5 dpf wild type motor nerve in which all axons express GFP and a single axon expresses mKate (magenta). (F) Pre-lesion image; transection site is marked by yellow box. Scale bar is 10 µm. (G) Axon sprouts in the proximal nerve stump begin to extend and retract by about 400 minutes post-transection. (H) A single pioneering axon (magenta) crosses the injury gap and extends ventrally along distal nerve by 800 minutes post-transection. Magenta arrowheads mark pioneer axon, which grows at a rate of 0.27 µm/min. (I) A follower axon extends and grows along the pioneering axon; green arrowheads mark follower path, which grows at a rate of 0.36 µm/min. (J) By the end of imaging, the magenta pioneer axon has extended ventrally, as shown via a maximum projection image across time (magenta arrowheads). (K-O) Still images from a time lapse movie showing early regeneration of lrp4 mutant motor nerve labeled with GFP. (K)

Pre-lesion image, with transection site marked by yellow box. (L) Early growth cones sprout from the proximal stump (open blue arrowheads) and a pioneer axon extends ventrally around 600 minutes post-transection (magenta arrowhead and ‘P’). (M-O) Throughout imaging, only the pioneer axon extends to ventral myotome; follower axons explore the transection site but do not traverse the injury gap (open blue arrowheads). (P) Quantification of pioneer axon growth rate compared to follower axon growth rate. Pioneer axon average rate of regrowth is 0.25 µm/min

(n=12 nerves from 8 larvae); follower axon average rate of regrowth is 0.47 µm/min (n=7 nerves from 4 larvae; unpaired t-test p=0.0055; t=3.181, df=17). (Q) Quantification of pioneer axon rate of regrowth in siblings compared to lrp4 mutants. Pioneer axons grow at equivalent rates: sibling average rate of regrowth=0.24 µm/min (n=12 nerves from 8 larvae); lrp4-/- average rate of regrowth=0.21 µm/min (n=10 nerves from 4 larvae; unpaired two-tailed t-test p=0.4702, t=0.7361,

df=20). (R) Quantification of number of follower axons in siblings versus lrp4 mutants. In sibling nerves, multiple follower axons extend (average number of follower axons=2.2, n=6 nerves from 2 larvae) while in lrp4 mutants, follower axons fail to extend to the ventral myotome (lrp4 mutant average number of follower axons=0.7, n=7 nerves from 2 larvae). Number of follower axons in siblings versus mutants is significantly different (two-tailed t-test p=0.0206, t=2.703, df=11).

Figure 3-2: lrp4 is required for axonal growth during regeneration

(A) Lateral view of a wild type motor nerve labeled with GFP before laser transection at 5 dpf, with transection site marked by yellow box. Scale bar is 10 µm. (A’) Six hours post-transection (hpt), motor axons distal to the transection site have fully degenerated. (A’’) Forty-eight hours post- transection, motor axons have regrown along their original path; green arrowheads indicate multiple regenerated fascicles; this representative nerve received a score of ‘best’ regeneration.

(B) Lateral view of lrp4 mutant motor nerve labeled with GFP before laser transection at 5 dpf, with transection site marked by yellow box. (B’) Six hours post-transection, distal motor axons have fully degenerated. (B’’) Forty-eight hours post-transection, motor axons have sprouted from the proximal stump but most fail to cross the injury site (open blue arrowheads); green arrowhead indicates stalled fascicle; this representative nerve received a score of ‘worst’ regeneration. (C)

Quantification of nerve regeneration in wild type and lrp4 mutant larvae. In wild type, 85% of nerves regenerate “best” with two or more distinct fascicles (n=26 nerves from 13 larvae). In lrp4 mutants,

70% of nerves regenerate zero or one fascicle (n=28 nerves from 13 larvae) by 48 hours post- transection. Wild type siblings regenerate significantly better than lrp4 mutants (Chi-square p<0.0001, Chi-square=18.48, df=2).

Figure 3-3: Schwann cell morphology during early regeneration is disrupted in lrp4 mutants

(A) Image of a representative pre-transection sibling motor nerve at 5 dpf labeled with GFP and stained with a Sox10 antibody that recognizes Schwann cell nuclei (magenta). Several Schwann cell nuclei are marked by yellow arrowheads; scale bar is 10 µm. (B) Image of a representative sibling motor nerve at 24 hours post-transection labeled with GFP and stained with a Sox10 antibody. Several Schwann cell nuclei are marked by yellow arrowheads. (C) Quantification of

Schwann cell nuclei associated with individual motor nerves pre-transection in siblings and lrp4 mutants. Average number of Schwann cells in siblings=11.9, per nerve, n=16 nerves; average number of Schwann cells in lrp4 mutants=10.78, per nerve, n=36 nerves. Two-tailed t-test p=0.08, t=1.743, df=50. (D) Quantification of Schwann cell nuclei associated with individual motor nerves 58

24 hours post-transection in siblings and lrp4 mutants. Average number of Schwann cells in siblings=11.4, n=20 nerves; average number of Schwann cells in lrp4 mutants=11.6, n=12 nerves.

Two-tailed t-test p=0.78, t=0.2693, df=30. (E-L) Still images from time lapse movies showing dynamics of early regeneration in wild type (E-H) and lrp4 mutant (I-L) motor nerves. Motor axons express GFP and Schwann cell express mRFP to label their membranes (shown in magenta). (E)

Pre-transection image of wild type motor nerve, with transection site marked by yellow box. Pre- injury Schwann cell membranes (shown in magenta and marked by yellow arrowheads) are thin and closely associate with GFP-labeled motor axons. White dashed box indicates the region of the nerve magnified in panels F and G. Scale bar is 10 µm. (F) Magnified still image from time lapse movie of early regeneration showing granular Schwann cell morphology characteristic of a denervated Schwann cell (marked by yellow arrowheads). (G) Magnified still image from time lapse movie showing that Schwann cell membranes gradually revert to a smooth morphology resembling pre-injury (flattened membranes marked by yellow arrowheads). Yellow dashed box marks the region of the nerve magnified in panel H. (H) Magnified still image at the same time point as panel

G, showing regenerating axons (labeled with GFP and marked with green arrowheads) closely associating with flattened Schwann cell membranes (yellow arrowheads; n=4 out of 4 wild type nerves). (I) Pre-transection image of lrp4 mutant motor nerve, with transection site marked by yellow box. Pre-injury Schwann cell membranes (shown in magenta and marked by yellow arrowheads) in lrp4 mutants are also thin and closely associate with GFP-labeled motor axons.

White dashed box indicates the region of the nerve magnified in panels J and K. (J) Magnified still image from time lapse movie of early regeneration in lrp4 mutant showing granular Schwann cell morphology characteristic of a denervated Schwann cell (marked by yellow arrowheads). (K)

Magnified still image from time lapse movie showing that Schwann cell membranes in lrp4 mutant fail to revert to a smooth pre-injury morphology (granular membranes marked by yellow arrowheads). Yellow dashed box marks the region of the nerve magnified in panel L. (L) Magnified still image at the same time point as panel K, showing a regenerating axon (labeled with GFP and

marked with green arrowhead) closely associating with still-granular Schwann cell membranes

(yellow arrowheads; n=7 out of 8 lrp4 mutant nerves).

Figure 3-4: lrp4 acts in an Agrin/MuSK-independent pathway to promote axonal regeneration

(A) Lateral view of wild type GFP-labeled motor nerve before laser transection at 5 dpf. Yellow dashed box indicates transection site, scale bar is 10 µm. (A’) Forty-eight hours post-transection, 61

multiple axon fascicles have regenerated ventrally. Green arrowheads indicate regenerated fascicles. (B) Lateral view of musk mutant motor nerve before laser transection at 5 dpf. Yellow dashed box indicates transection site. (B’) Forty-eight hours post-transection, motor axons have regenerated ventrally. Green arrowheads indicate regenerated fascicles. (C) Wild-type GFP- labeled motor nerve before laser transection at 5 dpf. Yellow dashed box marks transection site.

(C’) By forty-eight hours post-transection, multiple fascicles have regenerated ventrally (marked by green arrowheads). (D) agrin mutant motor nerve labeled with GFP before laser transection at 5 dpf. (D’) Forty-eight hours post-transection, multiple fascicles have regenerated ventrally (green arrowheads). (E) Quantification of nerve regeneration in wild type larvae (n=23 nerves from 11 larvae) and musk mutants (n=17 nerves from 11 larvae) at 48 hours post-transection. musk mutant motor nerves regenerate as well as wild type sibling motor nerves (Fisher’s exact test p=0.4982).

(F) Quantification of nerve regeneration in wild type larvae (n=10 nerves from 4 larvae) and agrin mutants (n=17 nerves from 6 larvae) at 48 hours post-transection. agrin mutant motor nerves regenerate as well as wild type sibling motor nerves (Fisher’s exact test p=0.6125). (G) Schematic of the canonical Agrin-Lrp4-MuSK pathway critical for neuromuscular synapse formation.

Figure 3-5: A truncated version of Lrp4 promotes nerve regeneration

(A) Schematic of lrp4 alleles and their protein structures: wild type full-length Lrp4 is shown at far left (1899 amino acids). lrp4-/- allele encodes a protein of 331 amino acids. lrp4ennui encodes a protein of 1398 amino acids that is predicted to contain almost the entire ectodomain but lack the transmembrane and intracellular domains. (B-D) Lateral views of wild type, lrp4-/-, and lrp4ennui embryos at the stage when en passant neuromuscular synapses form (26 hpf) stained for acetylcholine receptors (AChRs; magenta) and motor axons (green). Scale bar is 10 µm. (B-B’) In wild type embryos, AChR clusters assemble beneath the entire motor axon. (C-D’) In lrp4 mutant 63

and lrp4ennui embryos there is a strong reduction in AChR clustering beneath the motor axon, with only a few AChR clusters observed (white arrowheads); n=20 out of 20 embryos examined for each genotype. (E) Lateral view of pre-transection wild type motor nerve labeled with GFP at 5 dpf.

Transection site is marked by yellow box. Scale bar is 10 µm. (E’) Forty-eight hours post-transection in wild type larvae, multiple fascicles have regenerated ventrally (marked by green arrowheads).

(F) Lateral view of pre-transection motor nerve in lrp4ennui labeled with GFP at 5 dpf. Transection site marked by yellow box. (F’) Forty-eight hours post-transection in lrp4ennui, multiple fascicles have regenerated ventrally (green arrowheads). (G) Quantification of nerve regeneration in wild type larvae (n=26 nerves from 13 larvae) lrp4 mutants (n=28 nerves from 13 larvae) and lrp4ennui (n=24 nerves from 9 larvae) at 48 hours post-transection. lrp4ennui nerves regenerate to the same extent as wild type sibling nerves (wild type siblings “best” regeneration=85% of nerves; lrp4ennui “best” regeneration=83% of nerves; Chi-square test p=0.1672, Chi-square=3.577, df=2).

Figure 3-6: Specifically expressing lrp4 in motor neurons or muscle cells does not rescue axon regeneration

(A) Lateral view of lrp4 mutant motor nerve labeled with GFP before laser transection at 5 dpf.

Yellow dashed box marks transection site, scale bar is 10 µm. (A’) Forty-eight hours post- transection, a single fascicle has regenerated ventrally (green arrowhead). (B) Lateral view of motor 65

nerve in lrp4 mutant expressing Tg(mnx1:Lrp4-GFP) before laser transection at 5 dpf. Yellow dashed box indicates transection site. (B’) Forty-eight hours post-transection, a single regenerating fascicle has stalled prematurely (green arrowhead). (C) Quantification of nerve regeneration at 48 hours post-transection in lrp4 mutant larvae (n=9 nerves) and lrp4 mutants expressing

Tg(mnx1:Lrp4-GFP) (n=15 nerves). lrp4 mutant motor nerves regenerate to the same extent as lrp4 mutants expressing Tg(mnx1:Lrp4-GFP) (Chi-square test p=0.9266). (D) Lateral view of lrp4 mutant motor nerve labeled with GFP before laser transection at 5 dpf. Yellow dashed box marks transection site, scale bar is 10 µm. (D’) Forty-eight hours post-transection, a single fascicle has regenerated ventrally (green arrowhead). (E) Lateral view of motor nerve in lrp4 mutant expressing

Tg(α-actin:Lrp4-GFP) before laser transection at 5 dpf. Yellow dashed box indicates transection site. (E’) Forty-eight hours post-transection, a single regenerating fascicle has regenerated ventrally (green arrowhead). (F) Quantification of nerve regeneration at 48 hours post-transection in lrp4 mutant larvae (n=21 nerves) and lrp4 mutants expressing Tg(α-actin:Lrp4-GFP) (n=27 nerves). lrp4 mutant motor nerves regenerate to the same extent as lrp4 mutants expressing Tg(α- actin:Lrp4-GFP) (Fisher’s exact test p=0.8505).

Figure 3-7: Skeletal muscle-specific lrp4 expression rescues the developmental neuromuscular synaptic phenotype

(A) Acetylcholine receptors (AChRs) appear normal in zebrafish embryos at 26 hpf expressing

Tg(α-actin:Lrp4-GFP) in skeletal muscle cells. (B) Merged image of AChRs and motor axons labeled with znp-1 antibody, showing normal synapses. (C) Localization and expression of Tg(α- actin:Lrp4-GFP) in wild-type muscle cells. Lrp4-GFP localizes in puncta distributed throughout each muscle cell. (D,E) lrp4 mutants at 26 hpf show a dramatic reduction in AChR clusters and neuromuscular synapses. (F) No Lrp4-GFP expression is observed in this mutant, and the muscle

rescue transgene is not detectable by PCR. (G,H) AChR clusters and neuromuscular synapses appear wild type in lrp4 mutants that express the skeletal muscle Lrp4-GFP rescue transgene. (I)

Localization and expression of Tg(α-actin:Lrp4-GFP) in lrp4 mutant muscle cells looks the same as in wild-type embryos.

Figure 3-8: Zebrafish Agrin is required for neuromuscular synapse formation

(A) Summary of transcripts generated through CRISPR/Cas9 injection. The allele used for all experiments was a 7bpDEL;2bpINS at the end of exon 31, which resulted in loss of the exon 31 splice donor site. Wild type sequence is flanked by red arrows, and the genomic loss is marked in the mutant sequence just below. Sequencing cDNA from pools of mutants and siblings yielded four 69

predominant transcripts, whose alternative splicing is shown using green, orange, blue, and purple dashed lines. (B) Exon structure and resulting protein sequence for wild type agrin and the four predominant transcripts identified from sequencing cDNA. Three of four transcripts result in a frameshift and premature stop codon (purple, green, and orange transcripts), while one transcript is an in-frame deletion (blue transcript). All four transcripts result in severe swimming defects in zebrafish larvae at 36 hpf (not shown). (C-C’) At 26 hpf in wild type embryos, neuromuscular synapses (AChR labeling in magenta) are abundant beneath motor axons (green). (D-D’) In agrin mutants, neuromuscular synapses are significantly reduced; white arrowheads show residual

AChR clusters on the muscle cells at the horizontal myoseptum. N = 10 out of 10 embryos analyzed.

3.7 Supplemental Movies

Supplemental Movie 1: Multiple follower axons fasciculate with the pioneer axon

Lateral view of motor axon regeneration starting about 10 hours post-transection in a wild type sibling larva expressing GFP in all motor neurons. Proximal stump of the nerve is shown at the top of the image, from which axon sprouts extend and retract. At ~11 hpt, a pioneer axon extends from the proximal stump, crosses the injury gap, and grows ventrally along the original nerve trajectory

(magenta arrow). Subsequently three follower axons recognize the pioneer and fasciculate with it to grow ventrally (green arrows).

Supplemental Movie 2: A single axon crosses the injury gap and pioneers the regenerative path for subsequent follower axons

Lateral view of motor axon regeneration starting 7 hours post-transection (hpt) in a wild type sibling larva expressing GFP in all motor neurons and mKate in single motor axon. The proximal nerve stump is visible at the top of the movie, and from the start of imaging GFP- and mKate-positive axon sprouts extend and retract in the area around the injury site (open blue arrowheads). At ~11 hpt, a pioneering axon labeled with mKate extends from the proximal stump, crosses the injury gap, and grows ventrally along the original nerve trajectory (denoted with magenta arrowheads). At ~18 hpt, a GFP-positive follower axon (green arrowheads) crosses the injury gap and fasciculates with the pioneer axon. Images were captured every 10 minutes for 13 hours (from 7 hpt to 20 hpt).

Supplemental Movie 3: lrp4 is dispensable for pioneer axon regrowth but required for follower axon regrowth

Lateral view of motor axon regeneration starting about 10 hours post-transection in lrp4 mutant larva expressing GFP in all motor neurons. A pioneering axon (magenta arrow) has almost fully extended when imaging begins. Proximal stump of the nerve is shown at the top of the image, and axon sprouts (blue arrows) extend and retract in the area around the injury site throughout the course of imaging. By the end of imaging period at 20 hours post-transection, no follower axons have fasciculated with the pioneer axon and regenerated ventrally.

Supplemental Movie 4: Time lapse imaging of Schwann cell morphology changes during early regeneration in a wild type sibling larva

Lateral view of motor nerve regeneration starting 9 hours post-transection (hpt) in a wild type sibling expressing GFP in all motor neurons and mRFP in all Schwann cell membranes. When the movie starts at 9 hpt, Schwann cell membranes are discontinuous and granular as they de-differentiate and engulf axonal debris (yellow arrows from 540-590 minutes). As axons start regrowing (green arrowhead) around 11 hpt, Schwann cell membranes begin to flatten and re-extend and revert to their pre-injury morphology (yellow arrows).

Supplemental Movie 5: Time lapse imaging of Schwann cell morphology during early regeneration in lrp4 mutant larva

Lateral view of motor nerve regeneration 7 hours post-transection (hpt) in lrp4 mutant expressing

GFP in all motor neurons and mRFP in all Schwann cells to label Schwann cell membranes. When the movie starts at 7 hpt, Schwann cell membranes are discontinuous and granular as they de-

differentiate and engulf axonal debris (yellow arrows). By the end of imaging at ~19 hpt, Schwann cell membranes have remained granular and discontinuous, and do not revert to their pre-injury morphology (yellow arrows).

CHAPTER 4: DISCUSSION

In my dissertation research, I pursued two different but related avenues of research. In

Chapter 2, I explored the molecular mechanisms of synapse positioning and formation, and found that two genes that are required for neuromuscular synapse formation in mice—lrp4 and dok-7— have alternative functions in zebrafish synapse formation. My research affirms that genes do not always perform the same roles in different model organisms, and shows there is more than one way to build an effective and evolutionarily fit neuromuscular synapse. In Chapter 3, I investigated whether these neuromuscular synaptic proteins are required for peripheral motor axon regeneration, and I found a novel and unexpected role for lrp4 in this process. My findings demonstrate that Lrp4 promotes regeneration through a MuSK- and Agrin-independent pathway, and likely does not act in muscle cells or motor neurons in the context of regeneration. Rather we hypothesize that Lrp4 acts in Schwann cells to promote axon regeneration through interactions with proteins that remain to be identified.

4.1 The future of synapse positioning

Why is studying synapse positioning important? It is easy to gloss over the developmental steps between axon guidance and synapse formation, but it remains an important question. Many synapses in the body are positioned at precise and stereotyped locations: for example, in the zebrafish, cranial nerve VIII synapses onto the lateral dendrite of the Mauthner cell, while spiral fiber axons form synapses at the axon hillock of the Mauthner cell (reviewed in Korn and Faber,

2005). In the cochlea, spiral ganglion neurons synapse with inner and outer hair cells in very precise and specific patterns, transmitting auditory signals from the cochlea into the brain, but how these synapses are positioned correctly remains largely unknown (for review see Coate and Kelley,

2013). One family of trans-synaptic organizers of synapse positioning that was recently identified in Drosophila is the Teneurins, which are involved in organizing both neuromuscular synapses

(Mosca et al., 2012) and olfactory receptor neuron and projection neuron synaptic partner matching

(Hong et al., 2012). Interestingly, Teneurins are single-pass transmembrane proteins with large ectodomains that contain EGF-like repeats and YD (tyrosine-aspartate) repeats; this is like the ectodomain of Lrp4 that contains EGF-like repeats and YWTD β-propeller domains. Perhaps having a large extracellular region with these similar repetitive domains is evidence of a specific mechanism of synaptic partner matching and synapse positioning, and more roles for Lrp4 in synapse positioning will be discovered in the coming years.

Much of what is known about synapse positioning comes from studying neuromuscular synapses. Neuromuscular synapses form predominantly at equatorial locations, giving rise to a focal synapse at the center of an individual muscle fiber. The positioning of neuromuscular synapses is evident even before motor axons contact muscle fibers, and is marked by the accumulation of AChR clusters at the center of the muscle fiber, precisely where the future synapse will form. In C. elegans, neuromuscular synapses are positioned through a mechanism involving

Wnt signals locally restricting a Frizzled receptor to a sub-domain of the pre-synaptic axon that is supposed to remain asynaptic—this Wnt/Frizzled signaling locally inhibits pre-synaptic assembly

(Klassen and Shen, 2007). The Granato lab has shown that before motor axons contact their muscle targets, Wnt ligands bind MuSK in the muscle cell membrane, and this binding leads to endocytosis and trafficking of Wnt/MuSK into Rab11-positive endosomes that localize at the center of the muscle cell. There, MuSK interacts with Disheveled and other members of the Wnt/PCP signaling pathway to dictate where future neuromuscular synapses will form (Jing et al., 2009;

Gordon et al., 2012). Our lab has also shown that extracellular matrix molecules like chondroitin sulfate proteoglycans (CSPGs) are specifically localized on the muscle cell surface surrounding the growing motor axon, and that in the absence of musk, CSPGs are reduced (Zhang et al., 2004).

This suggests that early Wnt/MuSK signaling positions neuromuscular synapses at least partially through targeted deposits of ECM that restrict the growing motor axon and the early AChR clusters.

The neuromuscular synapse has proven to be an excellent model synapse to study mechanisms of synapse positioning. 75

Understanding synapse positioning is an interesting question and could be addressed by meticulously studying a synapse of interest. For example, a screen could be performed to identify genes that are required for the formation of synapses between inner ear hair cells and spiral fiber ganglion neurons, and then yeast two-hybrid studies could be conducted to identify interacting partners with the genes identified by the screen. In this way, one could identify which molecules are required to position, build, and maintain these synapses. Cell adhesion molecules, scaffolding proteins, proteins of the extracellular matrix, canonical axon guidance cues, or even LRP family members might be some genes identified in such a screen. If individual labs studied specific synapses in a focused and systematic way, then we might soon reach a better understanding of how synapses are positioned. In the distant future, it will be fascinating to understand how synapses in the cortex, hippocampus, or other regions of the brain are positioned, or even how they change over time. The typical CNS synapse is small, and synapses in many regions of the brain change dynamically with experience, so it is a challenging question to pursue. Perhaps we will learn that typical CNS synapses are not positioned in the same way as large stable synapses like the neuromuscular junction, inner hair cells, or olfactory bulb glomeruli; I think it is more likely we will learn that CNS synapses are positioned more randomly during development, and they are refined and shaped predominantly through experience.

4.2 Lrp4 is a multi-functional cell-surface receptor and ligand

In 2006, a group of researchers discovered that loss of Lrp4 causes dramatic neuromuscular synaptic defects in mice (Weatherbee et al., 2006), and then Lrp4 was characterized as the long-sought-after Agrin receptor, the link between Agrin and MuSK to build neuromuscular synapses (Kim et al., 2008; Zhang et al., 2008). Since its characterization as the

Agrin receptor, Lrp4 has been shown to play many roles within and outside the nervous system.

The Agrin-Lrp4-MuSK pathway is one application of Lrp4 that was of great impact to the field, because the identity of Lrp4 in this pathway was fervently sought for many years. However, if we step back and think about the other biological processes in which Lrp4 functions and its other 76

binding partners, which I will discuss below, we reach the conclusion that Lrp4 can act in many cell types as a multi-functional cell-surface receptor, and that it can also act as a ligand.

First, let us consider Lrp4 in the nervous system. In the PNS, Lrp4 is required to build and maintain neuromuscular junctions by binding Agrin and signaling through MuSK (Kim et al., 2008;

Zhang et al., 2008; Zong et al., 2012). In a paracrine role, the Lrp4 ectodomain can be cleaved from the surface of muscle cells and bind directly to an as-yet-unidentified receptor on motor axon terminals, thereby promoting nerve terminal differentiation (Wu et al., 2012; Yumoto et al., 2012).

My results reveal that it is required for peripheral nerve regeneration, likely acting in Schwann cells to promote axon regrowth across the injury gap (though whether it acts as a receptor on the surface of Schwann cells and/or as a secreted factor remains unanswered). In the CNS, Lrp4 is expressed in many areas of mouse brain and localizes at post-synaptic sites (Tian et al., 2006; Gomez et al.,

2014) and pre-synaptic active zones (Sun et al., 2016), and is required for excitatory synapse formation, number, and architecture in Drosophila by signaling through the kinase SRPK79D

(Mosca et al., 2017). Loss of Lrp4 causes cognitive and memory defects in mice (Gomez et al.,

2014) and impairments in olfactory attraction in Drosophila (Mosca et al., 2017). Outside the nervous system, Lrp4 is required for bone and tooth formation (Weatherbee et al., 2006), and point mutations in human LRP4 cause Cenani-Lenz syndrome, a form of congenital syndactyly with other defects in bone length and organization and kidney development (Li et al., 2010); and missense mutations cause Sclerosteosis 2, a syndrome of severe bone dysplasia and skeletal overgrowth

(Leupin et al., 2011). Clearly, Lrp4 acts in many cell types in the body, and perhaps others that have yet to be discovered, and it can bind to many different classes of molecules to promote these functions.

Since Lrp4 has evolved mechanisms to transduce signal in several different ways, I conclude that it acts as a multi-functional cell-surface receptor. Lrp4 can bind many ligands, modulate several signaling pathways, and itself act as a ligand. I believe in the next decade it will

be found to play more roles in the development and function of the nervous system, likely through novel mechanisms of action. Lrp4 is expressed in hippocampus, olfactory bulb, cerebellum, and neocortex, and Lrp4 is known to be involved in learning and memory in mice. I hypothesize that it will be required for synapse formation in the olfactory bulb, and perhaps even required for olfactory sensory neuron targeting to their glomeruli. Just as Lrp4 has many biological functions in the body, it also has diverse extracellular binding partners that contribute to its multi-functionality. The Lrp4 ectodomain is composed of LDLR repeats, four YWTD β-propeller domains, and five EGF-like domains. The Lrp4 ectodomain binds MuSK’s Ig-like domain (Zhang et al., 2011); Agrin’s laminin globular domain (Zong et al., 2012); Wnts and Wnt modulators like Sclerostin and Dickkopf 1 (Choi et al., 2009; Ahn et al., 2013); APP (Choi et al., 2013); Apolipoprotein E (Lu et al., 2007); and the

BMP antagonist Wise (Ohazama et al., 2008). Curiously, Lrp4 does not bind its namesake ligand, low-density lipoprotein. But the number and diversity of pathways that Lrp4 influences is remarkable, and the list continues to grow. Lrp4 is a multi-functional receptor with diverse functions, and I believe that continuing to study it will reveal many important mechanisms of nervous system development and function.

Interestingly, all seven members of the LRP family, except the original LDL receptor, are also multi-functional receptors expressed in many tissues. For example, Lrp1 interacts with 40 distinct ligands in many diverse tissues ranging from the nervous system to vasculature to lung.

And Lrp2 binds many ligands including vitamins, steroid hormones, BMPs, and a variety of lipoproteins and is expressed throughout the body (May and Herz, 2003; for review see May et al.,

2007). An interesting difference between Lrp4 and its other family members is the C-terminus of

Lrp4 appears not to signal (Pohlkamp et al., 2015), however it has a few noteworthy domains: an

NPXY motif that in other contexts can bind adaptor proteins including PTB-domain containing proteins like Dok-7 (Chen et al., 1990; Till et al., 2002; Okada et al., 2006); a YXXL motif that signals endocytosis of other LRP receptors (Li et al., 2000); and a terminal SESQV motif that can bind

PDZ-domain containing proteins (Johnson et al., 2006). While the other LRPs bind ligands and 78

then signal intracellularly, Lrp4 presently appears only to bind ligands, and has evolved to transduce its signal through co-receptors like MuSK or by acting as a ligand itself. Future studies should aim to determine whether the intracellular domain of Lrp4 is required for aspects of nervous system development or function. Does Lrp4 homodimerize across synaptic clefts to stabilize synapses; or perhaps does Lrp4 heterodimerize with other LRP family members or other proteins to stabilize synapses? Are there any synaptic phenotypes in the olfactory bulb or cerebellum of

Lrp4 mutants? If so, are there any associated behavioral phenotypes that originate from these areas of the brain? Many questions remain about Lrp4 and other LRPs in synaptic development and nervous system function, and studying this family of receptors will hopefully shed light on different aspects of nervous system development and function.

4.3 Neuron-extrinsic cues to promote axon regeneration

Many studies about axon regeneration in the nervous system revolve around the neuron- intrinsic factors that drive a neuron to re-form growth cones and regenerate axons, and thus many neuron-intrinsic molecules that regulate regeneration have been discovered, such as DLK, PTEN,

SOCS3, and the mTOR pathway among others (Park et al., 2008; Hammarlund et al., 2009; Yan et al., 2009; Sun et al., 2011; Duan et al., 2015; for further review see He and Jin, 2016). On the other hand, what is known about extrinsic or environmental factors that promote regeneration? We have a solid understanding of the cellular events that promote peripheral nervous system regeneration. For example, macrophages and Schwann cells phagocytose axonal and myelin debris after axon degeneration, thus clearing the path for regenerating axons (Scherer and Easter,

1984; Stoll et al., 1989; Fernandez-Valle et al., 1995; Perry et al., 1995; for review see Vargas and

Barres, 2007). And Schwann cells and perineural glia are required for axon regeneration (Ide et al.,

1983; Hall, 1986; Lewis and Kucenas, 2014), presumably providing both trophic and structural support for regenerating axons. It is less clear to what extent Schwann cells provide molecular cues to regenerating axons, but I hypothesize that they provide many molecular cues that have yet to be identified. 79

Recent work in the Granato lab suggests that Schwann cells harness at least two molecular pathways to promote axon regeneration. In one example, we showed that an extracellular matrix collagen acts in Schwann cells to promote axon regeneration by stabilizing the repulsive guidance cue Slit1a, which thereby preferentially directs a class of dorsal-projecting motor axons toward their original path (Isaacman-Beck et al., 2015). In addition, my thesis work suggests that Schwann cells present Lrp4 to injured motor axons, thereby promoting axon regeneration through interaction with an unknown receptor localized on a subset of regenerating follower axons (Gribble et al., under revision; see Chapter 3). An obvious future experiment would be to selectively delete lrp4 from

Schwann cells and observe whether axon regeneration is impaired. In addition, it will be interesting to determine whether Lrp4 acts as a receptor or a ligand in regeneration. If it acts as a ligand, which implies that it is secreted from Schwann cells, then what factor(s) enable its secretion? Are there metalloproteinases upregulated in denervated Schwann cells that enable Lrp4 cleavage; or perhaps there are metalloproteinases on regenerating axons that cleave Lrp4 from the surface of

Schwann cells? Many future questions remain about the precise role of Lrp4 in promoting motor axon regeneration. These two studies suggest that Schwann cells provide discrete molecular cues to regenerating axons, but surely this represents only a fraction of what Schwann cells do during degeneration and regeneration. I believe that the roles Schwann cells play are underappreciated, and that their basic cell biology during development and regeneration should be studied extensively. To me this is an area of neurobiology that is ready for deeper exploration.

Zebrafish provide an excellent model to study Schwann cells during de- and regeneration, because researchers can use live-cell imaging to capture the in vivo dynamics of Schwann cell biology during axon degeneration and regeneration (Rosenberg et al., 2012; 2014). In the future, researchers could use antibodies to stain for transcription factors that define Schwann cell states— i.e. Sox2 for immature Schwann cells; Sox10 for all Schwann cells; MBP to identify myelin; or Krox-

20 for myelinating Schwann cells (Topilko et al., 1994; Le et al., 2005; Jessen and Mirsky, 2008)— to determine temporally when Schwann cells change states in developing zebrafish. One could 80

then observe how these markers change during axon degeneration and regeneration, and when they exist in a particular state, one could capture Schwann cells and do mRNA profiling or proteomics to identify which genes are or are not expressed during Schwann cell de- and re- differentiation. I think doing this kind of systematic expression profiling of Schwann cells will greatly increase our understanding of the neuron-extrinsic, Schwann cell-derived factors that regulate axon regeneration.

In conclusion, many questions remain about the extrinsic cues that promote axon regeneration, and the cues that promote synapse positioning during development. The target muscle cells involved in neuromuscular synapse formation and the glial cells that myelinate peripheral nerves and promote their regeneration are both fascinating cell types that are critical for nervous system formation and function. I believe the future of neurobiology lies in persistently studying the cellular and molecular mechanisms that allow these and other non-neuronal cells to direct nervous system generation and regeneration.

CHAPTER 5: APPENDIX

5.1 Generation of MuSKΔKringle mutant zebrafish

5.1.1 sgRNA design and current results

The MuSK Kringle domain spans amino acids 454-536, and falls entirely and exclusively within exons 12-14. I originally designed two sgRNAs to target introns 11 and 14 to selectively excise the exons that encode the Kringle domain: the intron 11 sgRNA works well, but I tried three possible sgRNAs in intron 14 and none of them work (this intron is very short and the sites were predicted to be poor). I designed a second sgRNA in exon 14 that works, but this leaves a few

Kringle amino acids. Please see below for a table of the sgRNAs that I designed and tested for their efficacy:

Figure 5-1: Sequences of MuSK Kringle sgRNAs and whether sgRNAs worked

Location Sequence Did it work?

MuSK intron 11 #1 5’ AGCAGCTAATTAGCAGAAAGGGG 3’ Yes – HRMA and PCR verified

MuSK intron 14 #2 5’ CCATACTGAGTGCTAATCCTTTA 3’ No – HRMA / PCR

MuSK intron 14 #3 5’ CCAAATAATCTGAGTTAATATTT 3’ No – HRMA / PCR

MuSK intron 14 #4 5’ CCTTTATTCAGATAAATGTCATT 3’ No – HRMA / PCR

MuSK exon 14 5’ CCCTGGTGCTACACTACAAACCC 3’ Yes – HRMA / PCR

When I co-injected the sgRNAs targeting intron 11 and exon 14, I used a PCR assay to determine whether the Cas9 cut at both sites and repair mechanisms removed the intervening sequence. For further reference, including HRMA melt curves and PCR results, please see the following pages in

Katherine’s Notebook #7: 21, 22, 28, 38, 39, 40, 43, 50, 51. Happy reading!

Figure 5-2: Zebrafish MuSK domain structure

Domain structure of zebrafish MuSK (zebrafish musk originally named unplugged; FL refers to the full-length splice variant; “WT” means wild-type). The extracellular portion contains an amino- terminal signal sequence, three Ig-like domains, the Frizzled-like Cysteine-rich domain (CRD), and

Kringle domain; followed by a single-pass transmembrane domain, and the intracellular Tyrosine kinase domain.

5.1.2 MuSK Kringle domain next steps

There are G0 adult fish that were co-injected with intron 11 and exon 14 sgRNAs in a tank in the Blue system (injected November 2015; for injection details see Katherine’s Notebook #7 page 50). Some embryos from these injections were lysed to verify that both sgRNA target sites were cut, and they were (Katherine’s Notebook #7 page 51). These fish are mosaic for the genetic lesions, so perform random incrosses of the G0s and raise the resultant F1 embryos. When the F1 fish are old enough, clip their fins and use the PCR assay I established (for details see Katherine’s

Notebook #7 pages 43 and 51) to determine whether any of your F1 adults harbor mutations that are predicted to eliminate the Kringle domain but leave the rest of the open reading frame intact

(since the second sgRNA targets exon 14, you will need to verify that the mutation present in this exon is a non-frame-shifting mutation, i.e. an insertion or deletion of x basepairs, where x is a multiple of 3).

5.2 Juvenile Nerve Transection Protocol

I developed this protocol to transect peripheral motor nerves and monitor their regeneration in one month old juvenile zebrafish.

5.2.1 Part 1: Raising the fish (the first 5 days)

Day 1: bleach the embryos and sort them for Tg(mnx1:GFP). Make sure they spend the night at 25 degrees C, to delay them for mutant sorting the next morning.

Day 2: first thing in the morning, sort the embryos for the lrp4 mutant swimming phenotype

(it is best to screen for the mutant swimming phenotype around 36 hpf; by 48 hpf the phenotype disappears). Look for buzzing tails that don’t articulate in a full sinusoid shape, thereby preventing the larvae from swimming very far at all. Also on day 2, sort for the casper mutation that eliminates melanophores. I recommend raising only the larvae lacking melanophores — this is much easier and will yield higher N than if you also sort for absence of iridiphores (screening for iridiphores is difficult and must be done on day 3-4 on the fluorescent dissecting scope in the red channel). Keep picking embryos until you have a dish of 20 siblings and a second dish of 20 mutants. Try to pick

20 of each genotype with no melanophores, but it’s okay if you don’t have enough – in this case just add extra embryos with the same genotype to your dish. It’s important to raise the embryos in equal numbers so they grow to approximately the same size. N=20 is a good way to ensure that by 4 weeks old, the fish are the ideal size for cutting (between 1.5 and 1.8cm long).

Day 5: bring your sibling and mutant dishes down to the baby farm for raising.

5.2.2 Part 2: Nerve transection (days 28-34)

Plan your experiment for the week that the fish turn 4 weeks old. Early in the week, check them out in the baby farm to see how big they are. You can even Tricaine a few to measure them and make sure they are between 1.5 cm and 1.8 cm long. Prepare the following supplies:

• Tungsten needle in needle holder (0.37mm sharpened using KOH electrode bath in the

Bashaw lab // Melissa Baxter is the expert on doing this)

• 0.2% Tricaine in brown bottle

• 2.5% low-melt agarose in Ringer’s (aliquot ~700µl in an eppendorf tube; 4 tubes total);

melted and stored in a 42ºC heat block

You will need to do work on the following days during this experiment:

• Day 0: precut imaging and cutting

• Day 1: 24 hour post-cut imaging (to make sure the nerves are fully cut)

• Day 4: 4 day post-cut imaging (to observe regeneration)

Protocol

Day 0

• Put fish to sleep in 0.2% Tricaine diluted 1:10 in system water (final concentration 0.02%)

• Transfer fish using spoon to a slide that has a border of vacuum grease (or the new acrylic

slide that Owen made on the laser cutter!); mount the fish so it is lying on its right side, with

anterior to the left and dorsal up.

• Measure the fish and record your measurement (ideal size for a fish is between 1.5 cm and

1.8 cm).

• Remove as much liquid from around the fish as possible, and add a few drops of 2.5%

agarose to the tail and snout to immobilize the fish. Let dry for 10 seconds and then add

back a few drops of system water + Tricaine so the fish can breathe. You should be able

to see the fish breathing regularly or else it will die!

• Put the fish under the fluorescent dissecting scope and turn on the 488 lamp. Focus on the

pectoral fin nerve. Change the viewfinder so the camera is receiving the light. Focus the

image on the screen and take a picture - 83ms exposure works well. This is your precut

image. 85

• Change the viewfinder back to the microscope. Grab your tungsten needle and insert it just

above the left pectoral fin, where the adductor superficialis branch extends. Pull the needle

across the nerve, in a roughly-vertical SSW to NNE direction. See Figure

• The nerve is quite stretchy, so pulling on it with the tungsten needle might not initially sever

it. Stay persistent and keep pulling! The agarose should keep the fish mounted so you can

apply enough force to sever the nerve.

• When it is transected, you should be able to see the distal part degenerate almost

immediately.

• When you think the nerve is cut, put the whole fish on the slide into its single box so it starts

to wake up, then gently unmount it from the agarose. The fish should wake up within 2

minutes (one time it took 5 minutes, but we thought for sure the fish was dead).

• Clean and dry your slide with a kimwipe, and put the next fish to sleep to start the process

again.

Day 1 (24 hours post-transection)

• At 24 hours post-transection, verify that the nerve was fully cut. Repeat steps 1 and 2 from

Day 0, putting the fish to sleep and mounting it on a slide in the same way. Note: you do

not need to immobilize the fish with agarose, since you won’t be cutting anything!

• Once the fish is mounted on the slide, take a picture of its adductor superficialis nerve. It

should look like a nerve stump, with no obvious branches distal to the cut site. See Figure

• Put the fish back in its single box and feed it a little bit of brine shrimp.

Day 4 (4 days post-transection)

• Repeat the same protocol as on Day 1. The fish doesn’t need to survive any longer, so it’s

okay to wait until the fish stops breathing to capture the best image!

Troubleshooting

The biggest hurdle is capturing a good image on a slow camera with a fish that is actively breathing.

If you are having trouble getting a good image using the camera, you can try taking a screenshot of the nerve (keyboard shortcut Fn + PrintScreen) and pasting the image into Microsoft Paint (Ctrl

+ V).

Figure 5-3: Peripheral motor nerves in 4-week-old fish regenerate after transection injury

(A) Lateral, precut view of the zebrafish fin and gill area, with all motor nerves expressing

Tg(mnx1:GFP). In the field of view is the motor nerve that innervates the adductor superficialis muscle. Gill and fin are indicated. Transection site and direction of cut are indicated with a white arrow. (B) Immediately after cutting, the distal part of the nerve degenerates while the proximal part of the nerve retracts (arrowheads). (C) 1 day post-cut, the proximal part of the nerve has 87

begun extending branches that explore the environment (arrowheads). (D) By 4 days post-cut, the proximal part of the nerve has re-extended axons that grow across the injury gap and return to the original muscle target (arrowheads).

BIBLIOGRAPHY

Ahn Y, Sims C, Logue JM, Weatherbee SD, Krumlauf R (2013) Lrp4 and Wise interplay controls the formation and patterning of mammary and other skin appendage placodes by modulating Wnt signaling. Development 140:583–593.

Al-Majed AA, Neumann CM, Brushart TM, Gordon T (2000) Brief electrical stimulation promotes the speed and accuracy of motor axonal regeneration. Journal of Neuroscience 20:2602– 2608.

Amores A, Force A, Yan YL, Joly L, Amemiya C (1998) Zebrafish hox clusters and vertebrate genome evolution. Science 282:1711–1714.

Apel ED, Glass DJ, Moscoso LM, Yancopoulos GD, Sanes JR (1997) Rapsyn is required for MuSK signaling and recruits synaptic components to a MuSK-containing scaffold. Neuron 18:623–635.

Arthur-Farraj PJ, Latouche M, Wilton DK, Quintes S, Chabrol E, Banerjee A, Woodhoo A, Jenkins B, Rahman M, Turmaine M, Wicher GK, Mitter R, Greensmith L, Behrens A, Raivich G, Mirsky R, Jessen KR (2012) c-Jun Reprograms Schwann Cells of Injured Nerves to Generate a Repair Cell Essential for Regeneration. Neuron 75:633–647.

Bate CM (1976) Pioneer neurones in an insect embryo. Nature 260:54–56.

Bergamin E, Hallock PT, Burden SJ, Hubbard SR (2010) The Cytoplasmic Adaptor Protein Dok7 Activates the Receptor Tyrosine Kinase MuSK via Dimerization. Molecular cell 39:100–109.

Bevan S, Steinbach JH (1977) The distribution of α-bungarotoxin binding sites on mammalian skeletal muscle developing in vivo. J Physiol (Lond) 267:195–213.

Bezakova G, Helm JP, Francolini M, Lømo T (2001) Effects of purified recombinant neural and muscle agrin on skeletal muscle fibers in vivo. J Cell Biol 153:1441–1452.

Burden SJ, Depalma RL, Gottesman GS (1983) Crosslinking of proteins in acetylcholine receptor- rich membranes: Association between the β-subunit and the 43 kd subsynaptic protein. Cell 35:687–692.

Burgess RW, Nguyen QT, Son Y-J, Lichtman JW, Sanes JR (1999) Alternatively Spliced Isoforms of Nerve- and Muscle-Derived Agrin. Neuron 23:33–44.

Campanelli JT, Hoch W, Rupp F, Kreiner T, Scheller RH (1991) Agrin mediates cell contact- induced acetylcholine receptor clustering. Cell 67:909–916.

Castellino FJ, Beals JM (1987) The genetic relationships between the kringle domains of human plasminogen, prothrombin, tissue plasminogen activator, urokinase, and coagulation factor XII. J Mol Evol 26:358–369.

Chang CC, Lee CY (1963) Isolation of neurotoxins from the venom of Bungarus multicinctus and their modes of neuromuscular blocking action. Archives Internationales de Pharmacodynamie 144.

Changeux JP, Devillers-Thiery A, Chemouilli P (1984) Acetylcholine receptor: an allosteric protein. Science 225:1335–1345.

Chen WJ, Goldstein JL, Brown MS (1990) NPXY, a sequence often found in cytoplasmic tails, is required for coated pit-mediated internalization of the low density lipoprotein receptor. J Biol Chem 265:3116–3123.

Chen YY, McDonald D, Cheng C, Magnowski B, Durand J, Zochodne DW (2005) Axon and Schwann cell partnership during nerve regrowth. J Neuropathol Exp Neurol 64:613–622.

Cheng C, Zochodne DW (2002) In vivo proliferation, migration and phenotypic changes of Schwann cells in the presence of myelinated fibers. Neuroscience 115:321–329.

Choi HY, Dieckmann M, Herz J, Niemeier A (2009) Lrp4, a novel receptor for Dickkopf 1 and sclerostin, is expressed by osteoblasts and regulates bone growth and turnover in vivo. PLoS ONE 4:e7930.

Choi HY, Liu Y, Tennert C, Sugiura Y, Karakatsani A, Kröger S, Johnson EB, Hammer RE, Lin W, Herz J, Buckingham M (2013) APP interacts with LRP4 and agrin to coordinate the development of the neuromuscular junction in mice. eLife 2:e00220.

Coate TM, Kelley MW (2013) Making connections in the inner ear: recent insights into the development of spiral ganglion neurons and their connectivity with sensory hair cells. Seminars in Cell and Developmental Biology 24:460–469.

Dahlem TJ, Hoshijima K, Jurynec MJ, Gunther D, Starker CG, Locke AS, Weis AM, Voytas DF, Grunwald DJ (2012) Simple Methods for Generating and Detecting Locus-Specific Mutations Induced with TALENs in the Zebrafish Genome. PLoS Genet 8:e1002861.

DeChiara TM, Bowen DC, Valenzuela DM, Simmons MV, Poueymirou WT, Thomas S, Kinetz E, Compton DL, Rojas E, Park JS, Smith C, DiStefano PS, Glass DJ, Burden SJ, Yancopoulos GD (1996) The receptor tyrosine kinase MuSK is required for neuromuscular junction formation in vivo. Cell 85:501–512.

Dietrich MF, van der Weyden L, Prosser HM, Bradley A, Herz J, Adams DJ (2010) Ectodomains of the LDL receptor-related proteins LRP1b and LRP4 have anchorage independent functions in vivo. PLoS ONE 5:e9960.

Downes GB, Waterbury JA, Granato M (2002) Rapid in vivo labeling of identified zebrafish neurons. genesis 34:196–202.

Duan X, Qiao M, Bei F, Kim I-J, He Z, Sanes JR (2015) Subtype-Specific Regeneration of Retinal Ganglion Cells following Axotomy: Effects of Osteopontin and mTOR Signaling. Neuron 85:1244–1256.

Dyck PJ, Hopkins AP (1972) Electron microscopic observations on degeneration and regeneration of unmyelinated fibres. Brain 95:233–234.

Eaton RC, Bombardieri RA, Meyer DL (1977) The Mauthner-initiated startle response in teleost fish. J Exp Biol 66:65–81.

Eaton RC, Lee RK, Foreman MB (2001) The Mauthner cell and other identified neurons of the brainstem escape network of fish. Progress in Neurobiology 63:467–485.

Eisen JS, Myers PZ, Westerfield M (1986) Pathway selection by growth cones of identified motoneurones in live zebra fish embryos. Nature 320:269–271.

Fernandez-Valle C, Bunge RP, Bunge MB (1995) Schwann cells degrade myelin and proliferate in the absence of macrophages: evidence from in vitro studies of Wallerian degeneration. J Neurocytol 24:667–679.

Ferns M, Hoch W, Campanelli JT, Rupp F, Hall ZW, Scheller RH (1992) RNA splicing regulates agrin-mediated acetylcholine receptor clustering activity on cultured myotubes. Neuron 8:1079–1086.

Ferns MJ, Campanelli JT, Hoch W, Scheller RH, Hall Z (1993) The ability of agrin to cluster AChRs depends on alternative splicing and on cell surface proteoglycans. Neuron 11:491– 502.

Fischer LR, Culver DG, Tennant P, Davis AA, Wang M, Castellano-Sanchez A, Khan J, Polak MA, Glass JD (2004) Amyotrophic lateral sclerosis is a distal axonopathy: evidence in mice and man. Exp Neurol 185:232–240.

Flanagan-Steet H, Fox MA, Meyer D, Sanes JR (2005) Neuromuscular synapses can form in vivo by incorporation of initially aneural postsynaptic specializations. Development 132:4471– 4481.

Foreman MB, Eaton RC (1993) The direction change concept for reticulospinal control of goldfish escape. J Neurosci 13:4101–4113.

Frey D, Schneider C, Xu L, Borg J, Spooren W, Caroni P (2000) Early and selective loss of neuromuscular synapse subtypes with low sprouting competence in motoneuron diseases. Journal of Neuroscience 20:2534–2542.

Funakoshi H, Frisén J, Barbany G, Timmusk T, Zachrisson O, Verge VM, Persson H (1993) Differential expression of mRNAs for neurotrophins and their receptors after axotomy of the sciatic nerve. J Cell Biol 123:455–465.

Gautam M, Noakes PG, Moscoso L, Rupp F, Scheller RH, Merlie JP, Sanes JR (1996) Defective Neuromuscular Synaptogenesis in Agrin-Deficient Mutant Mice. Cell 85:525–535.

Gautam M, Noakes PG, Mudd J, Nichol M, Chu GC, Sanes JR, Merlie JP (1995) Failure of postsynaptic specialization to develop at neuromuscular junctions of rapsyn-deficient mice. Nature 377:232–236.

Gesemann M, Denzer AJ, Ruegg MA (1995) Acetylcholine receptor-aggregating activity of agrin isoforms and mapping of the active site. J Cell Biol 128:625–636.

Gillespie SK, Balasubramanian S, Fung ET, Huganir RL (1996) Rapsyn clusters and activates the synapse-specific receptor tyrosine kinase MuSK. Neuron 16:953–962.

Glass DJ, Bowen DC, Stitt TN, Radziejewski C, Bruno J, Ryan TE, Gies DR, Shah S, Mattsson K,

Burden SJ, DiStefano PS, Valenzuela DM, DeChiara TM, Yancopoulos GD (1996) Agrin acts via a MuSK receptor complex. Cell 85:513–523.

Godfrey EW, Nitkin RM, Wallace BG (1984) Components of Torpedo electric organ and muscle that cause aggregation of acetylcholine receptors on cultured muscle cells. J Cell Biol 99:615–627.

Gomez AM, Burden SJ (2011) The extracellular region of Lrp4 is sufficient to mediate neuromuscular synapse formation. Dev Dyn 240:2626–2633.

Gomez AM, Froemke RC, Burden SJ (2014) Synaptic plasticity and cognitive function are disrupted in the absence of Lrp4. eLife 3:583.

Gordon LR, Gribble KD, Syrett CM, Granato M (2012) Initiation of synapse formation by Wnt- induced MuSK endocytosis. Development 139:1023–1033.

Granato M, van Eeden FJ, Schach U, Trowe T, Brand M, Furutani-Seiki M, Haffter P, Hammerschmidt M, Heisenberg CP, Jiang YJ, Kane DA, Kelsh RN, Mullins MC, Odenthal J, Nüsslein-Volhard C (1996) Genes controlling and mediating locomotion behavior of the zebrafish embryo and larva. Development 123:399–413.

Gross-Thebing T, Paksa A, Raz E (2014) Simultaneous high-resolution detection of multiple transcripts combined with localization of proteins in whole-mount embryos. BMC Biology 12:55.

Hall AC, Lucas FR, Salinas PC (2000) Axonal remodeling and synaptic differentiation in the cerebellum is regulated by WNT-7a signaling. Cell 100:525–535.

Hall SM (1986) The effect of inhibiting Schwann cell mitosis on the re-innervation of acellular autografts in the peripheral nervous system of the mouse. Neuropathol Appl Neurobiol 12:401–414.

Hall ZW, Sanes JR (1993) Synaptic structure and development: The neuromuscular junction. Cell 72:99–121.

Hammarlund M, Nix P, Hauth L, Jorgensen EM, Bastiani M (2009) Axon Regeneration Requires a Conserved MAP Kinase Pathway. Science 323:802–806.

Han B, Zhao J-Y, Wang W-T, Li Z-W, He A-P, Song X-Y (2017) Cdc42 Promotes Schwann Cell Proliferation and Migration Through Wnt/β-Catenin and p38 MAPK Signaling Pathway After Sciatic Nerve Injury. Neurochem Res 42:1317–1324.

He X, Semenov M, Tamai K, Zeng X (2004) LDL receptor-related proteins 5 and 6 in Wnt/beta- catenin signaling: arrows point the way. Development 131:1663–1677.

He Z, Jin Y (2016) Intrinsic Control of Axon Regeneration. Neuron 90:437–451.

Heumann R, Korsching S, Bandtlow C, Thoenen H (1987) Changes of nerve growth factor synthesis in nonneuronal cells in response to sciatic nerve transection. J Cell Biol 104:1623– 1631.

Hidalgo A, Brand AH (1997) Targeted neuronal ablation: the role of pioneer neurons in guidance and fasciculation in the CNS of Drosophila. Development 124:3253–3262.

Hoch W, Ferns M, Campanelli JT, Hall ZW, Scheller RH (1993) Developmental regulation of highly active alternatively spliced forms of agrin. Neuron 11:479–490.

Hollis ER II, Zou Y (2012) Reinduced Wnt signaling limits regenerative potential of sensory axons in the spinal cord following conditioning lesion. Proceedings of the National Academy of Sciences USA 109:14663–14668.

Holtzman E, Novikoff AB (1965) Lysosomes in the rat sciatic nerve following crush. J Cell Biol 27:651–669.

Hong W, Mosca TJ, Luo L (2012) Teneurins instruct synaptic partner matching in an olfactory map. Nature 484:201–207.

Höke A, Gordon T, Zochodne DW, Sulaiman OAR (2002) A decline in glial cell-line-derived neurotrophic factor expression is associated with impaired regeneration after long-term Schwann cell denervation. Exp Neurol 173:77–85.

Hu X, Hicks CW, He W, Wong P, Macklin WB, Trapp BD, Yan R (2006) Bace1 modulates myelination in the central and peripheral nervous system. Nat Neurosci 9:1520–1525.

Hwang WY, Fu Y, Reyon D, Maeder ML, Tsai SQ, Sander JD, Peterson RT, Yeh J-RJ, Joung JK (2013) Efficient genome editing in zebrafish using a CRISPR-Cas system. Nat Biotechnol 31:227–229.

Ide C, Tohyama K, Yokota R, Nitatori T, Onodera S (1983) Schwann cell basal lamina and nerve regeneration. Brain Res 288:61–75.

Ikeo K, Takahashi K, Gojobori T (1991) Evolutionary origin of numerous kringles in human and simian apolipoprotein (a). FEBS Lett 287:146–148.

Inoue A, Setoguchi K, Matsubara Y, Okada K, Sato N, Iwakura Y, Higuchi O, Yamanashi Y (2009) Dok-7 activates the muscle receptor kinase MuSK and shapes synapse formation. Sci Signal 2:ra7.

Isaacman-Beck J, Schneider V, Franzini-Armstrong C, Granato M (2015) The lh3 Glycosyltransferase Directs Target-Selective Peripheral Nerve Regeneration. Neuron 88:691–703.

Jennings CG, Dyer SM, Burden SJ (1993) Muscle-specific trk-related receptor with a kringle domain defines a distinct class of receptor tyrosine kinases. Proc Natl Acad Sci USA 90:2895–2899.

Jessen KR, Mirsky R (2008) Negative regulation of myelination: Relevance for development, injury, and demyelinating disease. Glia 56:1552–1565.

Jing L, Gordon L, Shtibin E, Granato M (2010) Temporal and spatial requirements of unplugged/MuSK function during zebrafish neuromuscular development. PLoS ONE 5:e8843.

Jing L, Lefebvre JL, Gordon LR, Granato M (2009) Wnt signals organize synaptic prepattern and axon guidance through the zebrafish unplugged/MuSK receptor. Neuron 61:721–733.

Johnson EB, Steffen DJ, Lynch KW, Herz J (2006) Defective splicing of Megf7/Lrp4, a regulator of distal limb development, in autosomal recessive mulefoot disease. Genomics 88:600–609.

Jones G, Meier T, Lichtsteiner M, Witzemann V, Sakmann B, Brenner HR (1997) Induction by agrin of ectopic and functional postsynaptic-like membrane in innervated muscle. Proc Natl Acad Sci USA 94:2654–2659.

Karlsson E, Heilbronn E, Widlund L (1972) Isolation of the nicotinic acetylcholine receptor by biospecific chromatography on insolubilized Naja naja neurotoxin. FEBS Lett 28:107–111.

Kim M, Roesener AP, Mendonca PRF, Mastick GS (2011) Robo1 and Robo2 have distinct roles in pioneer longitudinal axon guidance. Dev Biol 358:181–188.

Kim N, Burden SJ (2008) MuSK controls where motor axons grow and form synapses. Nat Neurosci 11:19–27.

Kim N, Stiegler AL, Cameron TO, Hallock PT, Gomez AM, Huang JH, Hubbard SR, Dustin ML, Burden SJ (2008) Lrp4 is a receptor for Agrin and forms a complex with MuSK. Cell 135:334–342.

Klassen MP, Shen K (2007) Wnt signaling positions neuromuscular connectivity by inhibiting synapse formation in C. elegans. Cell 130:704–716.

Klose M, Bentley D (1989) Transient pioneer neurons are essential for formation of an embryonic peripheral nerve. Science 245:982–984.

Korn H, Faber DS (2005) The Mauthner Cell Half a Century Later: A Neurobiological Model for Decision-Making? Neuron 47:13–28.

Kuan Y-S, Roberson S, Akitake CM, Fortuno L, Gamse J, Moens C, Halpern ME (2015) Distinct requirements for Wntless in habenular development. Dev Biol 406:117–128.

Kucenas S, Takada N, Park H-C, Woodruff E, Broadie K, Appel B (2008) CNS-derived glia ensheath peripheral nerves and mediate motor root development. Nat Neurosci 11:143–151.

Le N, Nagarajan R, Wang JYT, Araki T, Schmidt RE, Milbrandt J (2005) Analysis of congenital hypomyelinating Egr2Lo/Lo nerves identifies Sox2 as an inhibitor of Schwann cell differentiation and myelination. Proc Natl Acad Sci USA 102:2596–2601.

Lefebvre JL, Jing L, Becaficco S, Franzini-Armstrong C, Granato M (2007) Differential requirement for MuSK and dystroglycan in generating patterns of neuromuscular innervation. Proc Natl Acad Sci USA 104:2483–2488.

Leupin O et al. (2011) Bone overgrowth-associated mutations in the LRP4 gene impair sclerostin facilitator function. Journal of Biological Chemistry 286:19489–19500.

Lewis GM, Kucenas S (2014) Perineurial Glia Are Essential for Motor Axon Regrowth following Nerve Injury. Journal of Neuroscience 34:12762–12777. 94

Li Y, Marzolo MP, van Kerkhof P, Strous GJ, Bu G (2000) The YXXL Motif, but Not the Two NPXY Motifs, Serves as the Dominant Endocytosis Signal for Low Density Lipoprotein Receptor-related Protein. J Biol Chem 275:17187–17194.

Li Y, Pawlik B, Elcioglu N, Aglan M, Kayserili H (2010) LRP4 mutations alter Wnt/β-catenin signaling and cause limb and kidney malformations in Cenani-Lenz syndrome. The American Journal of Human Genetics 86:696–706.

Lin W, Burgess RW, Dominguez B, Pfaff SL, Sanes JR, Lee KF (2001) Distinct roles of nerve and muscle in postsynaptic differentiation of the neuromuscular synapse. Nature 410:1057–1064.

Liu HM, Yang LH, Yang YJ (1995) Schwann cell properties: 3. C-fos expression, bFGF production, phagocytosis and proliferation during Wallerian degeneration. Journal of Neuropathology & Experimental Neurology 54:487–496.

Liu Y, Wang X, Lu C-C, Kerman R, Steward O, Xu X-M, Zou Y (2008) Repulsive Wnt signaling inhibits axon regeneration after CNS injury. Journal of Neuroscience 28:8376–8382.

Lu Y, Tian QB, Endo S, Suzuki T (2007) A role for LRP4 in neuronal cell viability is related to apoE-binding. Brain Res 1177:19–28.

Luo ZG, Wang Q, Zhou JZ, Wang J, Luo Z, Liu M, He X, Wynshaw-Boris A, Xiong WC, Lu B, Mei L (2002) Regulation of AChR clustering by Dishevelled interacting with MuSK and PAK1. Neuron 35:489–505 Available at: http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&id=12165471&retmode =ref&cmd=prlinks.

May P, Herz J (2003) LDL Receptor-Related Proteins in Neurodevelopment. Traffic 4:291–301.

May P, Woldt E, Matz RL, Boucher P (2007) The LDL receptor-related protein (LRP) family: An old family of proteins with new physiological functions. Ann Med 39:219–228.

McConnell SK, Ghosh A, Shatz CJ (1989) Subplate neurons pioneer the first axon pathway from the cerebral cortex. Science 245:978–982.

McDonald D, Cheng C, Chen Y, Zochodne D (2005) Early events of peripheral nerve regeneration. Neuron Glia Biol 2:139–139.

Messéant J, Dobbertin A, Girard E, Delers P, Manuel M, Mangione F, Schmitt A, Le Denmat D, Molgó J, Zytnicki D, Schaeffer L, Legay C, Strochlic L (2015) MuSK frizzled-like domain is critical for mammalian neuromuscular junction formation and maintenance. Journal of Neuroscience 35:4926–4941.

Messéant J, Ezan J, Delers P, Glebov K, Marchiol C, Lager F, Renault G, Tissir F, Montcouquiol M, Sans N, Legay C, Strochlic L (2017) Wnt proteins contribute to neuromuscular junction formation through distinct signaling pathways. Development 144:1712–1724.

Meyer M, Matsuoka I, Wetmore C (1992) Enhanced synthesis of brain-derived neurotrophic factor in the lesioned peropheral nerve. The Journal of Cell Biology 119:45–54.

Michailov GV, Sereda MW, Brinkmann BG, Fischer TM, Haug B, Birchmeier C, Role L, Lai C, 95

Schwab MH, Nave K-A (2004) Axonal neuregulin-1 regulates myelin sheath thickness. Science 304:700–703.

Mosca TJ, Hong W, Dani VS, Favaloro V, Luo L (2012) Trans-synaptic Teneurin signalling in neuromuscular synapse organization and target choice. Nature 484:237–241.

Mosca TJ, Luginbuhl DJ, Wang IE, Luo L (2017) Presynaptic LRP4 promotes synapse number and function of excitatory CNS neurons. eLife 6:e27347.

Muller JS, Jepson CD, Laval SH, Mei L (2010) Dok-7 promotes slow muscle integrity as well as neuromuscular junction formation in a zebrafish model of congenital myasthenic syndromes. Hum Mol Genet 19:1726–1740.

Mullins MC, Hammerschmidt M, Haffter P, Nüsslein-Volhard C (1994) Large-scale mutagenesis in the zebrafish: in search of genes controlling development in a vertebrate. Current Biology 4:189–202.

Mullins MC, Hammerschmidt M, Kane DA, Odenthal J, Brand M, van Eeden FJ, Furutani-Seiki M, Granato M, Haffter P, Heisenberg CP, Jiang YJ, Kelsh RN, Nüsslein-Volhard C (1996) Genes establishing dorsoventral pattern formation in the zebrafish embryo: the ventral specifying genes. Development 123:81–93.

Neubig RR, Krodel EK, Boyd ND, Cohen JB (1979) Acetylcholine and local anesthetic binding to Torpedo nicotinic postsynaptic membranes after removal of nonreceptor peptides. Proc Natl Acad Sci USA 76:690–694.

Nguyen QT, Sanes JR, Lichtman JW (2002) Pre-existing pathways promote precise projection patterns. Nature Neuroscience 5:861–867.

Nitkin RM, Smith MA, Magill C, Fallon JR, Yao YM, Wallace BG, McMahan UJ (1987) Identification of agrin, a synaptic organizing protein from Torpedo electric organ. J Cell Biol 105:2471–2478.

Ohazama A, Johnson EB, Ota MS, Choi HJ, Porntaveetus T, Oommen S, Itoh N, Eto K, Gritli- Linde A, Herz J, Sharpe PT (2008) Lrp4 Modulates Extracellular Integration of Cell Signaling Pathways in Development. PLoS ONE 3:e4092.

Okada K, Inoue A, Okada M, Murata Y, Kakuta S, Jigami T, Kubo S, Shiraishi H, Eguchi K, Motomura M, Akiyama T, Iwakura Y, Higuchi O, Yamanashi Y (2006) The muscle protein Dok-7 is essential for neuromuscular synaptogenesis. Science 312:1802–1805.

Ono F, Higashijima S, Shcherbatko A, Fetcho JR, Brehm P (2001) Paralytic zebrafish lacking acetylcholine receptors fail to localize rapsyn clusters to the synapse. Journal of Neuroscience 21:5439–5448.

Ono F, Shcherbatko A, Higashijima S-I, Mandel G, Brehm P (2002) The Zebrafish motility mutant twitch once reveals new roles for rapsyn in synaptic function. Journal of Neuroscience 22:6491–6498.

Packard M, Koo ES, Gorczyca M, Sharpe J, Cumberledge S, Budnik V (2002) The Drosophila Wnt, wingless, provides an essential signal for pre- and postsynaptic differentiation. Cell

111:319–330.

Panzer JA, Song Y, Balice-Gordon RJ (2006) In vivo imaging of preferential motor axon outgrowth to and synaptogenesis at prepatterned acetylcholine receptor clusters in embryonic zebrafish skeletal muscle. Journal of Neuroscience 26:934–947.

Parikh P, Hao Y, Hosseinkhani M, Patil SB, Huntley GW, Tessier-Lavigne M, Zou H (2011) Regeneration of axons in injured spinal cord by activation of bone morphogenetic protein/Smad1 signaling pathway in adult neurons. Proc Natl Acad Sci USA 108:E99–E107.

Park J-Y, Ikeda H, Ikenaga T, Ono F (2012) Acetylcholine receptors enable the transport of rapsyn from the Golgi complex to the plasma membrane. Journal of Neuroscience 32:7356– 7363.

Park KK, Liu K, Hu Y, Smith PD, Wang C, Bin Cai, Xu B, Connolly L, Kramvis I, Sahin M, He Z (2008) Promoting Axon Regeneration in the Adult CNS by Modulation of the PTEN/mTOR Pathway. Science 322:963–966.

Parkinson DB, Bhaskaran A, Arthur-Farraj P, Noon LA, Woodhoo A, Lloyd AC, Feltri ML, Wrabetz L, Behrens A, Mirsky R, Jessen KR (2008) c-Jun is a negative regulator of myelination. J Cell Biol 181:625–637.

Parrinello S, Napoli I, Ribeiro S, Digby PW, Fedorova M, Parkinson DB, Doddrell RDS, Nakayama M, Adams RH, Lloyd AC (2010) EphB Signaling Directs Peripheral Nerve Regeneration through Sox2-Dependent Schwann Cell Sorting. Cell 143:145–155.

Patthy L (1985) Evolution of the proteases of blood coagulation and fibrinolysis by assembly from modules. Cell 41:657–663.

Patthy L, Trexler M, Váli Z, Bányai L, Váradi A (1984) Kringles: modules specialized for protein binding. Homology of the gelatin-binding region of fibronectin with the kringle structures of proteases. FEBS Lett 171:131–136.

Perry VH, Tsao JW, Fearn S, Brown MC (1995) Radiation-induced reductions in macrophage recruitment have only slight effects on myelin degeneration in sectioned peripheral nerves of mice. Eur J Neurosci 7:271–280.

Pérez-García MJ, Burden SJ (2012) Increasing MuSK Activity Delays Denervation and Improves Motor Function in ALS Mice. Cell Rep 2:497–502.

Pike SH, Melancon EF, Eisen JS (1992) Pathfinding by zebrafish motoneurons in the absence of normal pioneer axons. Development 114:825–831.

Pinson KI, Brennan J, Monkley S, Avery BJ, Skarnes WC (2000) An LDL-receptor-related protein mediates Wnt signalling in mice. Nature 407:535–538.

Pittman AJ, Law M-Y, Chien C-B (2008) Pathfinding in a large vertebrate axon tract: isotypic interactions guide retinotectal axons at multiple choice points. Development 135:2865–2871.

Pohlkamp T, Durakoglugil M, Lane-Donovan C, Xian X, Johnson EB, Hammer RE, Herz J (2015) Lrp4 Domains Differentially Regulate Limb/Brain Development and Synaptic Plasticity

Phillips W, ed. PLoS ONE 10:e0116701–e0116711.

Raper J, Mason C (2010) Cellular strategies of axonal pathfinding. Cold Spring Harbor Perspectives in Biology 2:a001933.

Raper JA, Bastiani M, Goodman CS (1983) Pathfinding by neuronal growth cones in grasshopper embryos. II. Selective fasciculation onto specific axonal pathways. J Neurosci 3:31–41.

Remédio L, Gribble KD, Lee JK, Kim N, Hallock PT, Delestrée N, Mentis GZ, Froemke RC, Granato M, Burden SJ (2016) Diverging roles for Lrp4 and Wnt signaling in neuromuscular synapse development during evolution. Genes Dev 30:1058–1069.

Rosenberg AF, Isaacman-Beck J, Franzini-Armstrong C, Granato M (2014) Schwann cells and deleted in colorectal carcinoma direct regenerating motor axons towards their original path. Journal of Neuroscience 34:14668–14681.

Rosenberg AF, Wolman MA, Franzini-Armstrong C, Granato M (2012) In Vivo Nerve-Macrophage Interactions Following Peripheral Nerve Injury. Journal of Neuroscience 32:3898–3909.

Ruegg MA, Tsim KW, Horton SE, Kröger S, Escher G, Gensch EM, McMahan UJ (1992) The agrin gene codes for a family of basal lamina proteins that differ in function and distribution. Neuron 8:691–699.

Saint-Amant L, Sprague SM, Hirata H, Li Q, Cui WW, Zhou W, Poudou O, Hume RI, Kuwada JY (2007) The zebrafish ennui behavioral mutation disrupts acetylcholine receptor localization and motor axon stability. Devel Neurobio 68:45–61.

Salpeter MM, Loring RH (1985) Nicotinic acetylcholine receptors in vertebrate muscle: Properties, distribution and neural control. Progress in Neurobiology 25:297–325.

Sanes JR, Lichtman JW (2001) Induction, assembly, maturation and maintenance of a postsynaptic apparatus. Nat Rev Neurosci 2:791–805.

Scarlato M, Xu T, Bannerman P, Beesley J, Reddy UR, Rostami A, Scherer SS, Pleasure D (2001) Axon-Schwann cell interactions regulate the expression of fibroblast growth factor-5 (FGF-5). J Neurosci Res 66:16–22.

Schaefer AM, Sanes JR, Lichtman JW (2005) A compensatory subpopulation of motor neurons in a mouse model of amyotrophic lateral sclerosis. J Comp Neurol 490:209–219.

Scherer SS (1997) The biology and pathobiology of Schwann cells. Curr Opin Neurol 10:386– 397.

Scherer SS, Easter SS (1984) Degenerative and regenerative changes in the trochlear nerve of goldfish. J Neurocytol 13:519–565.

Schmidt J, Raftery MA (1972) Use of affinity chromatography for acetylcholine receptor purification. Biochem Biophys Res Commun 49:572–578.

Schmidt J, Raftery MA (1973) Purification of acetylcholine receptors from Torpedo claifornica electroplax by affinity chromatography. Biochemistry 12:852–856. 98

Silver A (1963) A Histochemical Investigation of Cholinesterases at Neuromuscular Junctions in Mammalian and Avian Muscle. J Physiol (Lond) 169:386–393.

Steinbach JH (1981) Developmental changes in acetylcholine receptor aggregates at rat skeletal neuromuscular junctions. Dev Biol 84:267–276.

Stoll G, Griffin JW, Li CY, Trapp BD (1989) Wallerian degeneration in the peripheral nervous system: participation of both Schwann cells and macrophages in myelin degradation. J Neurocytol 18:671–683.

Strand NS, Hoi KK, Phan TMT, Ray CA, Berndt JD, Moon RT (2016) Wnt/β-catenin signaling promotes regeneration after adult zebrafish spinal cord injury. Biochem Biophys Res Commun 477:952–956.

Strochlic L, Falk J, Goillot E, Sigoillot S, Bourgeois F, Delers P, Rouvière J, Swain A, Castellani V, Schaeffer L, Legay C (2012) Wnt4 Participates in the Formation of Vertebrate Neuromuscular Junction Klymkowsky M, ed. PLoS ONE 7:e29976.

Sun F, Park KK, Belin S, Wang D, Lu T, Chen G, Zhang K, Yeung C, Feng G, Yankner BA, He Z (2011) Sustained axon regeneration induced by co-deletion of PTEN and SOCS3. Nature 480:372–375.

Sun X-D et al. (2016) Lrp4 in astrocytes modulates glutamatergic transmission. Nat Neurosci 19:1010–1018.

Tamai K, Semenov M, Kato Y, Spokony R, Liu C, Katsuyama Y, Hess F, Saint-Jeannet JP, He X (2000) LDL-receptor-related proteins in Wnt signal transduction. Nature 407:530–535.

Taniuchi M, Clark HB, Johnson EM Jr (1986) Induction of nerve growth factor receptor in Schwann cells after axotomy. Proceedings of the National Academy of the Sciences USA 83:4094–4098.

Taveggia C, Zanazzi G, Petrylak A, Yano H, Rosenbluth J, Einheber S, Xu X, Esper RM, Loeb JA, Shrager P, Chao MV, Falls DL, Role L, Salzer JL (2005) Neuregulin-1 Type III Determines the Ensheathment Fate of Axons. Neuron 47:681–694.

Thermes V, Grabher C, Ristoratore F, Bourrat F, Choulika A, Wittbrodt J, Joly J-S (2002) I-SceI meganuclease mediates highly efficient transgenesis in fish. Mech Dev 118:91–98.

Tian QB, Suzuki T, Yamauchi T, Sakagami H, Yoshimura Y, Miyazawa S, Nakayama K, Saitoh F, Zhang J-P, Lu Y, Kondo H, Endo S (2006) Interaction of LDL receptor-related protein 4 (LRP4) with postsynaptic scaffold proteins via its C-terminal PDZ domain-binding motif, and its regulation by Ca/calmodulin-dependent protein kinase II. Eur J Neurosci 23:2864–2876.

Till JH, Becerra M, Watty A, Lu Y, Ma Y, Neubert TA, Burden SJ, Hubbard SR (2002) Crystal Structure of the MuSK Tyrosine Kinase. Structure 10:1187–1196.

Topilko P, Schneider-Maunoury S, Levi G, Baron-Van Evercooren A, Chennoufi ABY, Seitanidou T, Babinet C, Charnay P (1994) Krox-20 controls myelination in the peripheral nervous system. Nature 371:796–799.

Trevarrow B, Marks DL, Kimmel CB (1990) Organization of hindbrain segments in the zebrafish embryo. Neuron 4:669–679.

Valenzuela DM, Stitt TN, DiStefano PS, Rojas E, Mattsson K, Compton DL, Nuñez L, Park JS, Stark JL, Gies DR (1995) Receptor tyrosine kinase specific for the skeletal muscle lineage: expression in embryonic muscle, at the neuromuscular junction, and after injury. Neuron 15:573–584.

Vargas ME, Barres BA (2007) Why is Wallerian degeneration in the CNS so slow? Annu Rev Neurosci 30:153–179.

Waller A (1850) Experiments on the Section of the Glossopharyngeal and Hypoglossal Nerves of the Frog, and Observations of the Alterations Produced Thereby in the Structure of Their Primitive Fibres. Phil Trans R Soc Lond 140:423–429.

Weatherbee SD, Anderson KV, Niswander LA (2006) LDL-receptor-related protein 4 is crucial for formation of the neuromuscular junction. Development 133:4993–5000.

Willem M, Garratt AN, Novak B, Citron M, Kaufmann S, Rittger A, DeStrooper B, Saftig P, Birchmeier C, Haass C (2006) Control of Peripheral Nerve Myelination by the -Secretase BACE1. Science 314:664–666.

Wolman MA, Jain RA, Liss L, Granato M (2011) Chemical modulation of memory formation in larval zebrafish. Proc Natl Acad Sci USA 108:15468–15473.

Woodhoo A, Alonso MBD, Droggiti A, Turmaine M, D'Antonio M, Parkinson DB, Wilton DK, Al- Shawi R, Simons P, Shen J, Guillemot F, Radtke F, Meijer D, Feltri ML, Wrabetz L, Mirsky R, Jessen KR (2009) Notch controls embryonic Schwann cell differentiation, postnatal myelination and adult plasticity. Nat Neurosci 12:839–847.

Wu H, Lu Y, Shen C, Patel N, Gan L, Xiong WC, Mei L (2012) Distinct roles of muscle and motoneuron LRP4 in neuromuscular junction formation. Neuron 75:94–107.

Wu H, Xiong WC, Mei L (2010) To build a synapse: signaling pathways in neuromuscular junction assembly. Development 137:1017–1033. y Cajal SR (1928) Degeneration & regeneration of the nervous system. Oxford University Press, Humphrey Milford.

Yan D, Wu Z, Chisholm AD, Jin Y (2009) The DLK-1 kinase promotes mRNA stability and local translation in C. elegans synapses and axon regeneration. Cell 138:1005–1018.

Yang X, Arber S, William C, Li L, Tanabe Y, Jessell TM, Birchmeier C, Burden SJ (2001) Patterning of muscle acetylcholine receptor gene expression in the absence of motor innervation. Neuron 30:399–410.

Yang X, Li W, Prescott ED, Burden SJ, Wang JC (2000) DNA topoisomerase IIbeta and neural development. Science 287:131–134.

Yumoto N, Kim N, Burden SJ (2012) Lrp4 is a retrograde signal for presynaptic differentiation at neuromuscular synapses. Nature 489:438–442.

100

Zhang B, Liang C, Bates R, Yin Y, Xiong W-C, Mei L (2012) Wnt proteins regulate acetylcholine receptor clustering in muscle cells. Mol Brain 5:7.

Zhang B, Luo S, Wang Q, Suzuki T, Xiong WC, Mei L (2008) LRP4 serves as a coreceptor of agrin. Neuron 60:285–297.

Zhang J, Granato M (2000) The zebrafish unplugged gene controls motor axon pathway selection. Development 127:2099–2111.

Zhang J, Lefebvre JL, Zhao S, Granato M (2004) Zebrafish unplugged reveals a role for muscle- specific kinase homologs in axonal pathway choice. Nat Neurosci 7:1303–1309.

Zhang JY, Luo XG, Xian CJ, Liu ZH, Zhou XF (2000) Endogenous BDNF is required for myelination and regeneration of injured sciatic nerve in rodents. Eur J Neurosci 12:4171– 4180.

Zhang W, Coldefy A-S, Hubbard SR, Burden SJ (2011) Agrin binds to the N-terminal region of Lrp4 protein and stimulates association between Lrp4 and the first immunoglobulin-like domain in muscle-specific kinase (MuSK). Journal of Biological Chemistry 286:40624–40630.

Zhong J, Zou H (2014) BMP signaling in axon regeneration. Curr Opin Neurobiol 27:127–134.

Ziskind-Conhaim L, Bennett JI (1982) The effects of electrical inactivity and denervation on the distribution of acetylcholine receptors in developing rat muscle. Dev Biol 90:185–197.

Ziskind-Conhaim L, Dennis MJ (1981) Development of rat neuromuscular junctions in organ culture. Dev Biol 85:243–251.

Zong Y, Zhang B, Gu S, Lee K, Zhou J, Yao G, Figueiredo D, Perry K, Mei L, Jin R (2012) Structural basis of agrin-LRP4-MuSK signaling. Genes Dev 26:247–258.

101