Where nerve meets muscle in ALS: Insights from a human NMJ-on-a-chip model

Amyotrophic Lateral Sclerosis (ALS) is a fatal neurodegenerative disease characterized by the progressive degeneration of motor neurons, leading to muscle atrophy and ultimately death of the patient. A central and unresolved debate in ALS research concerns the site of disease initiation. The "dying-forward" hypothesis posits that degeneration begins in the central nervous system and spreads anterogradely. Conversely, the "dying-back" hypothesis suggests that pathology originates distally at the neuromuscular junction (NMJ), progressing retrogradely towards the cell body. Understanding the cellular and molecular mechanisms driving the pathology is of great importance for identifying early therapeutic targets. This thesis investigates the "dying-back" hypothesis by dissecting the contributions of motor neurons and skeletal muscle to NMJ dysfunction.

To overcome the interspecies differences limiting the translational value of animal models, we developed a novel, fully human in vitro model: an induced pluripotent stem cell (iPSC)-derived NMJ-on-a-chip. This microfluidic platform allows for the co-culture of patient-specific motor neurons and skeletal muscle in compartmentalized microenvironments, enabling the formation of functional synaptic connections and the independent analysis of cell-type-specific pathologies. Using this platform, we investigated two major genetic forms of ALS: FUS and C9ORF72.

In motor neurons carrying FUS mutations, we identified a cell-autonomous pathology localized to the axon. By sequencing the axonal transcriptome, we discovered dysregulation of RNA metabolism, including altered alternative polyadenylation of SNCA and a pathogenic upregulation of SLITRK2. Functional analysis in our NMJ-on-a-chip demonstrated that elevated SLITRK2 levels directly drive aberrant axonal outgrowth and impair neuromuscular transmission efficacy, which could be rescued by knockdown of SLITRK2. This supports a neuron-intrinsic mechanism for distal axonopathy in FUS-ALS.

Our study of C9ORF72-ALS highlighted an active pathogenic role for skeletal muscle. C9-ALS myotubes exhibited intrinsic defects, including the accumulation of RNA foci and dipeptide repeat proteins (DPRs), cellular hyperactivity, and metabolic dysregulation. Importantly, these diseased myotubes exerted a non-cell autonomous toxic effect on healthy motor neurons, impairing NMJ integrity. Furthermore, we identified extracellular vesicles (EVs) as a potential mediator of this toxicity. C9-ALS muscle secretes EVs with an altered proteomic cargo, including reduced levels of the axon-guidance protein CRMP1, suggesting disrupted intercellular signaling as a driver of denervation.

In conclusion, this thesis reveals that while both mutations lead to NMJ dysfunction, distal pathology in ALS is a complex interplay of mutation-specific triggers and cell-type-specific vulnerabilities. The pathology can be initiated intrinsically within the motor neuron axon or driven extrinsically by diseased skeletal muscle. By establishing the NMJ as a convergence point for these diverse pathogenic mechanisms, this work provides a refined model of the "dying-back" hypothesis and highlights the potential of the NMJ as an early target for therapeutic interventions in ALS.