Semaphorin 3A (SEMA3A) is a chemorepulsive guidance cue that, during development, plays an important role in guiding neurons to their target. Previous work has identified presymptomatic expression of SEMA3A in terminal Schwann cells (TSCs) at the neuromuscular junction (NMJ) of G93A-hSOD1 ALS mice (De Winter et al., 2006) giving rise to the hypothesis that it may be acting as a chemorepulsive molecule to induce detachment of the motor neuron from its target thereby acting as a protein that induces or facilitates the disease process. In this thesis we investigate this hypothesis by interfering with the signaling of SEMA3A in several ways (outlined below).
Chapter 1 of this thesis provides an extensive review of how alterations in guidance molecules in ALS could be altering the cytoskeleton at the level of the NMJ to initiate the dying-back phenotype in motor neurons, a pathological feature that may be one of the first signs of the development of ALS (Moloney et al., 2014). Several lines of evidence illustrate a potential role for repulsive axon guidance molecules, including NOGO- A (Jokic et al., 2006), EPHA4 (Van Hoecke et al., 2012) and SEMA3A (De Winter et al., 2006) in the selective denervation at the NMJ in ALS.
We used a multi-route approach to study how SEMA3A may fit into the current view of ALS as a distal axonopathy:
SEMA3A mediates its repulsive guidance abilities by interacting with a membrane bound receptor complex composed of the ligand-binding component, neuropilin-1 (NRP1) and the signal-transducing component, plexinA. To manipulate the SEMA3A signaling pathway we generated soluble NRP1 receptors to act as molecular scavengers to interfere with the binding between SEMA3A and the membrane-bound NRP1 receptor. Since NRP1 is also a receptor for Vascular Endothelial Growth Factor (VEGF) we generated various mutant isoforms of the soluble NRP1 receptor to specifically bind to SEMA3A, VEGF, or neither. In Chapter 2 we used a Dorsal Root Ganglion (DRG) growth cone collapse assay to characterize the SEMA3A-neutralizing abilities of the various soluble NRP1 receptors. We observed that receptor isoforms designed to scavenge for SEMA3A were able to significantly inhibit SEMA3A-induced growth cone collapse. The receptor isoform that was designed to bind VEGF only was, as expected, unable to prevent SEMA3A-induced growth cone collapse. Contrary to expected, the receptor isoform designed to bind to neither ligand, was able to inhibit SEMA3A-induced growth cone collapse. In Chapter 3 we used Adeno-associated Viral (AAV) vectors to introduce the soluble NRP1 receptors into the gastrocnemic muscles of G93A-hSOD1 ALS mice. We first observed that unilateral or bilateral administration of soluble NRP1 receptors to the gastrocnemic muscle affects the overall outcome in terms of motor function. Upon bilateral treatment all soluble NRP1 receptor isoforms attenuated the presymptomatic dip in Rotarod performance usually observed in ALS mice. Furthermore, motor performance was significantly altered in the symptomatic phase of the disease. Two NRP1-receptor isoforms (WT, and the VEGF-only binding isoform) improved motor function by delaying the onset of performance decline which manifests in untreated mice in the early symptomatic phase. On the other hand, we observed a severe worsening of motor performance in the remaining two groups (SEMA-only binding isoform, and the "ligand-deficient" isoform). The data from Chapter 2 and Chapter 3 illustrate a clear discrepancy between our in vitro and in vivo results respectively, and we discuss the possible explanations for these paradoxical findings at the end of each of those chapters, as well as in the General Discussion (Chapter 6).
We used an AAV-mediated overexpression of SEMA3A in skeletal muscle to characterize its effects on NMJ stability and/or motor function (Chapter 4). We observed that although we were able to overexpress SEMA3A in the gastrocnemic muscle of WT mice, it failed to cause changes in motor function or anatomical changes at the NMJ. We discuss that muscle-mediated overexpression of SEMA3A alone is unable to cause a progressive pathological-like state at the NMJ.
We used a transgenic mouse line harboring a SEMA3A mutation that decreases the chemorepulsive properties of the protein and characterized the effects of reduced SEMA3A signaling on sprouting after injury, or on ALS-induced motor deficits (Chapter 5). In ALS mice, the presence of mutant SEMA3A does not alter ALS-related motor performance decline. In addition, we did not observe changes in BotoxA-induced sprouting of motor neurons on muscle fiber subtypes that are generally thought to be non-plastic in mice expressing the mutant SEMA3A. However, our BotoxA-experiment did reveal a gender difference in basal sprouting capacities in wild-type mice on what the literature labels as "plastic" and "non-plastic" muscle fiber subtypes, with female motor neurons displaying the opposite sprouting response compared to males after BotoxA-induced paralysis of the gastrocnemic muscle.
Finally in Chapter 6 we discuss the results obtained in the preceding chapters by focusing on four main points: 1. The discrepancies between the in vitro and in vivo effects of the soluble NRP1 receptors; 2. Mimicking the TSC-specific upregulation of SEMA3A may be required to demonstrate its effects, but it poses a significant challenge; 3. The effects of manipulating the SEMA3A-NRP1 pathway are sensitive to acute or chronic changes in signalling capacity; 4. Plasticity at the neuromuscular junction exhibits a gender dimorphism that may be linked to the gender bias of ALS. We conclude by proposing a future experiment whereby SEMA3A can be specifically removed from TSCs of ALS mice. This in itself poses many technical challenges, as currently the best promoter for TSC-directed expression (S100B) has some activity in central nervous system glia, but it may provide more definitive answer to how SEMA3A is involved in the early pathophysiology of ALS.