CRITERIA FOR IDENTIFICATION AND PRECISE QUANTIFICATION OF SPINAL MOTOR NEURONS

Motor neuron (MN) loss is a hallmark of development and several neurodegenerative disorders. Since investigation of MN degeneration in human spinal cords is limited, mouse models are indispensable for investigating underlying pathomechanisms and therapeutic efficacy. However, individual studies using identical models often report divergent MN loss, implying methodological sources of variability. To quantify and identify these sources, we performed a systematic literature review and meta-analysis of mouse models of the motor neuron disease spinal muscular atrophy (SMA), revealing nonspecific spinal cord sampling as a major confounder. To address this experimentally, we developed a standardized, manual whole-segment sectioning approach combining ex vivo MN tracing, immunohistochemistry, tissue clearing, and confocal imaging in wild-type mice. We identified SMI-32 and Nissl stains as nonspecific contributors to variability, whereas choline acetyltransferase (ChAT) proved to be the most reliable MN marker in both murine and human spinal cords. Furthermore, we implemented a deep-learning–based segmentation method for intact mouse spinal cords for unbiased quantifications and to validate exact MN counts for individual spinal segments. Detailed lumbar segment analysis in SMA mice, combined with computational modeling, confirmed nonspecific segment sampling as a key driver of variability. Applying this framework across MN disease models revealed reproducible patterns: widespread α-MN loss in amyotrophic lateral sclerosis (ALS) mice, selective proximal α- and γ-MN degeneration in severe SMA, and preserved MNs in mild SMA. These results provide a universal framework for standardized MN identification and quantification, resolving longstanding discrepancies and enabling robust assessment of disease mechanisms and therapeutic interventions.