Close interaction between oligodendrocytes and axons is essential to initiate myelination and to form specialised domains along myelinated fibres. These domains are characterised by the assembly of protein complexes at the axon-glia interface and key components of these complexes are the Neurofascins. Neurofascins are transmembrane glycoproteins belonging to the L1 subgroup of the Immunoglobulin (Ig) superfamily of cell adhesion molecules. The Neurofascin (Nfasc) gene is subject to extensive alternative splicing. Two of the best characterised isoforms are Nfasc155 and Nfasc186, which are expressed in glia and neurons respectively. In myelinated fibres, Nfasc186 is the predominant isoform expressed at nodes of Ranvier and axon initial segments (AIS) in both the central and peripheral nervous system (CNS and PNS), whereas Nfasc155 resides on the glial side of the paranodal axoglial junction. The Neurofascin gene has been inactivated by homologous recombination and Neurofascin-null mice die within the first week of postnatal life. The main focus of this work was to investigate the role of the Neurofascins in the developing CNS. Similarly to what has been previously observed in the PNS, this study shows that in myelinated fibres of the spinal cord, nodal and paranodal markers are mislocalised and axoglial junctions do not form in the absence of the Neurofascins. In contrast to the PNS, where ensheathment of axons is unaffected, myelin proteins in the CNS are greatly reduced in the mutant. This appears to be due to the reduced ability of oligodendrocyte myelinating processes to extend along axons. This work also shows that the role of Nfasc186 is to maintain the long term stability of the AIS rather than its assembly. In the PNS, Nfasc186 was found to play an essential role in node assembly. However, PNS and CNS nodes are likely to assemble by different mechanisms. To investigate the relative contribution of the Neurofascin isoforms in CNS node assembly, this work made use of transgenic lines in which either neuronal Nfasc186 or glial Nfasc155 was expressed on a Neurofascin null background. Expression of either isoform was found to independently rescue the nodal complex and a model of how the Neurofascins cooperate in the assembly of the CNS node of Ranvier is proposed.

The mouse nervous system undergoes a vast remodelling of synaptic connections postnatally, resulting in a reduced number of innervating axons to target cells within the first few weeks of life. This extensive loss of connections is known as synapse elimination and it plays a critical role in sculpting and refining neural connectivity throughout the nervous system, resulting in a finely tuned and well-synchronised network of innervation. This process has been well characterised at the mouse neuromuscular junction (NMJ), where synapse elimination takes place postnatally in all skeletal muscles. It has been well studied for the reasons that it is easily accessible for live imaging and post-mortem experimental analysis. Studies utilising this synapse to uncover regulators of synapse elimination have mainly focused on the importance of glial cell lysosomal activity, nerve conduction and target-derived growth factor supply. It is clear that non-axonal cell types play key roles in the success of developmental axon retraction at the NMJ, however the role of glial cells in the regulation of this process has not been fully explored, as lysosomal activity is thought of as a consequence of axon pruning rather than a molecular driver. Previous studies have shown that signals emanating from myelinating glial cells can modulate neurofilament composition and transport within the underlying axons. We know that these changes in neurofilament composition and transport are underway during developmental synapse elimination at the NMJ, so it seems logical to predict that myelinating glial cells may play a role in the regulation of axonal pruning. Myelinating glial cells are found along the entire length of lower motor neurons and form physical interactions with the underlying axons at regions known as paranodes. At the paranode, Neurofascin155 (Nfasc155: expressed by the myelinating glial cell) interacts with a Caspr/contactin complex (expressed by the axon). This site has been proposed as a likely site for axon-glial signalling due to the close apposition of the cell membranes. The main focus of this PhD project was to study the potential role of myelinating glial cells in the success of synapse elimination at the NMJ, using a mouse model of paranodal disruption (Nfasc155-/-). Chapters 3 and 4 show the results of this work. This work has revealed a novel role for glia in the modulation of synapse elimination at the mouse neuromuscular junction, mediated by Nfasc155 in the myelinating Schwann cell. Synapse elimination was profoundly delayed in Nfasc155-/- mice and was found to be associated with a non-canonical role for Nfasc155, as synapse elimination occurred normally in mice lacking the axonal paranodal protein Caspr. Loss of Nfasc155 was sufficient to disrupt axonal proteins contributing to cytoskeletal organisation and trafficking pathways in peripheral nerve of Nfasc155-/- mice and lower levels of neurofilament light (NF-L) protein in maturing motor axon terminals. Synapse elimination was delayed in mice lacking NF-L, suggesting that Nfasc155 influences neuronal remodelling, at least in part, by modifying cytoskeletal dynamics in motor neurons. This work provides the first clear evidence for myelinating Schwann cells acting as drivers of synapse elimination, with Nfasc155 playing a critical role in glial cell-mediated postnatal sculpting of neuronal connectivity in the peripheral nervous system. A small section of the results within this thesis are devoted to the study of axon-glial interactions in a mouse model of childhood motor neuron disease, otherwise known as spinal muscular atrophy (SMA). In SMA, there are reduced levels of the ubiquitously expressed survival motor neuron (SMN) protein. The NMJ is a particularly vulnerable target in SMA, manifesting as a breakdown of neuromuscular connectivity and progressive motor impairment. Recent studies have begun to shed light on the role of non-neuronal cell types in the onset and progression of the disease, presenting SMA as a multi-system disease rather than a purely neuronal disorder. Recent evidence has highlighted that myelinating glial cells are significantly affected in a mouse model of SMA, manifesting as an impaired ability to produce key myelin proteins, resulting in deficient myelination. The final results chapter of this thesis (Chapter 5) is focussed on further exploring the effects that loss of SMN has in Schwann cells including their interactions with underlying axons. This work reveals a disruption to axon-glial interaction, shown by a delay in the development of paranodes, supporting the idea that non-neuronal cell types are also affected in SMA. The results within this thesis reveal a novel role for a glial cell protein, Nfasc155, in the modulation of synapse elimination at the NMJ. Mechanistic insight in to Nfasc155’s role in this process is also uncovered and likely involves axonal cytoskeletal transport systems and the filamentous protein NF-L, which have not previously been implicated in the process of synapse elimination. This work highlights an important role for axon-glial interactions during normal postnatal development of the mouse NMJ. This work also highlights a role for axon-glial interactions in disease states of the NMJ. Using a mouse model of SMA, axon-glial interaction was assessed with the finding of a delay in paranodal maturation due to loss of SMN.

Willin is a novel member of the Four-point-one Ezrin Radixin Moesin (FERM) protein superfamily, containing an N-terminal FERM domain most like the Ezrin-Radixin-Moesin (ERM) family but also the closely related protein Merlin. Willin was initially discovered as a yeast two-hybrid binding partner of neurofascin155, and this interaction has now been confirmed by both co-localisation studies and the use of two different biochemical methods. Like neurofascin155, Willin also localises to detergent resistant membranes, and like the ERM family, it is able to bind to phospholipids. The expression of Willin appears to be toxic as the production of cell-lines stably expressing Willin proved to be not possible and this appears to be because it induces apoptosis in cultured cells. This is a proliferation control function consistent with the suggestion that Willin is the human homologue of the Drosophila tumour suppressor ‘Expanded’. Three antibodies to Willin were also characterised and a novel splice variant, Willin2, subcloned into a GFP-tagged plasmid for comparison with the original form.

This thesis contains the findings from four years of research regarding an oligodendrocyte protein named neurofascin 155. The role of this protein in maintaining adhesion between the myelin sheath of oligodendrocytes and the axons of neurons has become well established in recent years and the research presented here has revealed that while western blots have previously shown one protein/band representing neurofascin 155, there are two proteins/bands. These two proteins have been named neurofascin 155 high and neurofascin 155 low due to their previous inclusion in the single band. The work leading up to their discovery, findings, and the relevance of these two proteins will be discussed in animal models with disrupted myelin:axon adhesion and in the human disease multiple sclerosis.

Axon initial segment (AIS) disruption has been described in a number of pathological environments where neuroinflammation is a contributing factor; however, whether this disruption is reversible in unknown. To address the principle of AIS structural recovery, we employed an acute neuroinflammatory model. Acute neuroinflammation induced disruption of AIS structural and functional domains and, importantly, upon resolution of neuroinflammatory conditions, was reversed. Consistent with other studies, we observed a close interaction of microglia with AISs, and utilized this acute neuroinflammatory model to investigate the relationship between reactive microglia and AIS integrity. Gene expression analysis of microglial transcription profiles identified reactive oxygen species (ROS)-producing enzymes as candidates in AIS pathogenesis. Experiments employing mice lacking the major ROS-producing enzyme NOX2, identified ROS as mediators of AIS disruption. Furthermore, we established calcium-dependent protease calpain as a disruptor of AIS protein clustering in inflammation-induced disruption. Since we observed an intimate interaction between microglia and the AIS, we conducted studies designed to identify a candidate in microglia that regulates microglial-AIS contact. During chronic inflammatory conditions, microglia enhance contact with AISs often completely surrounding the domain. Concomitant with this morphological change, neurofascin (Nfasc) expression increased in microglia. Nfasc is a cell adhesion molecule with cell-specific isoforms known to mediate glial-neuronal interactions, but until now, was not reported to be expressed by microglia. Here, I characterize the unique Nfasc isoform expressed by microglia and present evidence that suggests that microglial Nfasc may mediate microglial-AIS contact, a potentially pivotal interaction in the induction of AIS disruption by pro-inflammatory factors.

Background: Both the innate and adaptive immune systems contribute to autoimmune injury in multiple sclerosis (MS). We have been particularly interested in elucidating the role of the innate γδ T-cell population in MS pathogenesis. In particular, some γδ T-cells that express Fc receptors (FcR), such as CD16, that bind antibody are more prominent with MS disease progression and have been shown to exert cytolysis via antibody-dependent cellular cytotoxicity (ADCC). We postulated that if there were also relevant and detectable antibodies in MS patients that might engage these FcR-bearing γδ T-cells then this might be a purported mechanism of neuro-axonal injury. A search for antibodies specific to axonal elements in MS revealed the presence of antibodies to neurofascin (Nfasc). Methods: Anti-Nfasc antibody titres, and concentrations of the light and heavy chains of neurofilament (NfL and NfH, respectively), markers of neuro-axonal injury, were measured in the sera and cerebrospinal fluid (CSF) of MS patients using enzyme-linked immunosorbent assays (ELISA), including those that underwent autologous hematopoietic stem cell transplantation (aHSCT), both prior to and yearly for 3 years thereafter. HeLa cells were transfected with the axonal variant of Nfasc, Nfasc-186, and were utilized as targets in ADCC assays involving γδ T-cells as the effectors, and anti-Nfasc antibodies that were enriched from MS patient sera. Results: Positive anti-Nfasc antibody titres were detected in of 22% and 25% of MS patient sera and CSF, respectively. The most elevated serum titres were in secondary progressive MS (SPMS), and highest CSF titres in relapsing-remitting MS (RRMS) (p<0.05 and p<0.0001, respectively, vs. other neurological disease [OND] controls). Patient serum and CSF antibody titres correlated and, in the CSF, the titres correlated positively with the concentration of NfL. Though NfL and NfH concentrations declined markedly following aHSCT in the CSF, anti-Nfasc antibody titres failed to decline. When co-cultured with CD16+ γδ T-cells in the presence of MS patient-derived anti-Nfasc antibodies, the percent specific cytolysis of the Nfasc-transfected HeLa cells was significantly greater than that of the non-transfected control HeLa cells, at 18% and 1%, respectively, indicating cytolytic kill via ADCC. Summary: Anti-Nfasc antibodies were detectable in the sera and CSF of MS patients, and rarely in OND controls, suggesting they are relevant to MS. Higher titres in the serum support peripheral synthesis, while higher CSF titres in the relapsing phase, that correlate with serum titres, imply that antibodies access the CNS during periods of active inflammation that are associated with disruption of the blood-CSF barrier. CSF anti-Nfasc antibody titres correlated strongly with the release of NfL, suggesting that axonal injury could be related to the presence of Nfasc-specific antibodies. Following aHSCT, CSF NfL and NfH release were reduced without concomitant CSF anti-Nfasc antibody reductions, suggesting that the presence alone of anti-Nfasc antibodies is not enough to cause axonal injury. Indeed, when co-cultured with CD16+ γδ T-cells in the presence of MS patient-derived anti-Nfasc antibodies, the percent specific cytolysis of the Nfasc-transfected HeLa cells was significantly greater than that of the non-transfected control HeLa cells, proving that FcR-bearing γδ T-cells can cause axonal damage by lysing axonal membranes via ADCC, when armed with axon-specific antibodies such as anti-Nfasc. This is the first report of γδ T-cell-mediated cytolysis by ADCC using both γδ T-cells and antibodies derived from MS patients.