Search

Mutations in the gene encoding the RNA-binding protein TDP-43 cause amyotrophic lateral sclerosis (ALS), clinically and pathologically indistinguishable from the majority of ‘sporadic’ cases of ALS, establishing altered TDP-43 function and distribution as a primary mechanism of neurodegeneration. Transgenic mouse models in which TDP-43 is overexpressed only partially recapitulate the key cellular pathology of human ALS, but may also lead to non-specific toxicity. To avoid the potentially confounding effects of overexpression, and to maintain regulated spatio-temporal and cell-specific expression, we generated mice in which an 80 kb genomic fragment containing the intact human TDP-43 locus (either TDP-43WT or TDP-43M337V) and its regulatory regions was integrated into the Rosa26 (Gt(ROSA26)Sor) locus in a single copy. At 3 months of age, TDP-43M337V mice are phenotypically normal but by around 6 months develop progressive motor function deficits associated with loss of neuromuscular junction integrity, leading to a reduced lifespan. RNA sequencing shows that widespread mis-splicing is absent prior to the development of a motor phenotype, though differential expression analysis reveals a distinct transcriptional profile in pre-symptomatic TDP-43M337V spinal cords. Despite the presence of clear motor abnormalities, there was no evidence of TDP-43 cytoplasmic aggregation in vivo at any timepoint. In primary embryonic spinal motor neurons and in embryonic stem cell (ESC)-derived motor neurons, mutant TDP-43 undergoes cytoplasmic mislocalisation, and is associated with altered stress granule assembly and dynamics. Overall, this mouse model provides evidence that ALS may arise through acquired TDP-43 toxicity associated with defective stress granule function. The normal phenotype until 6 months of age can facilitate the study of early pathways underlying ALS.

TDP-43 pathology is a key feature of amyotrophic lateral sclerosis (ALS), but the mechanisms linking TDP-43 to altered cellular function and neurodegeneration remain unclear. We have recently described a mouse model in which human wild-type or mutant TDP-43 are expressed at low levels and where altered stress granule formation is a robust phenotype of TDP-43M337V/− expressing cells. In the present study we use this model to investigate the functional connectivity of human TDP-43 in primary motor neurons under resting conditions and in response to oxidative stress. The interactome of human TDP-43WT or TDP-43M337V was compared by mass spectrometry, and gene ontology enrichment analysis identified pathways dysregulated by the M337V mutation. We found that under normal conditions the interactome of human TDP-43WT was enriched for proteins involved in transcription, translation and poly(A)-RNA binding. In response to oxidative stress, TDP-43WT recruits proteins of the endoplasmic reticulum and endosomal-extracellular transport pathways, interactions which are reduced in the presence of the M337V mutation. Specifically, TDP-43M337V impaired protein-protein interactions involved in stress granule formation including reduced binding to the translation initiation factors Poly(A)-binding protein and Eif4a1 and the endoplasmic reticulum chaperone Grp78. The M337V mutation also affected interactions involved in endosomal-extracellular transport and this this was associated with reduced extracellular vesicle secretion in primary motor neurons from TDP-43M337V/− mice and in human iPSCs-derived motor neurons. Taken together, our analysis highlights a TDP-43 interaction network in motor neurons and demonstrates that an ALS associated mutation may alter the interactome to drive aberrant pathways involved in the pathogenesis of ALS.

Mutations in the ubiquitously expressed survival motor neuron 1 (SMN1) and superoxide dismutase 1 (SOD1) genes are selectively lethal to motor neurons in spinal muscular atrophy (SMA) and familial amyotrophic lateral sclerosis (ALS), respectively. Genetic association studies provide compelling evidence that SMN1 and SMN2 genotypes encoding lower SMN protein levels are implicated in sporadic ALS, suggesting that SMN expression is a potential determinant of ALS severity. We therefore sought genetic evidence of SMN involvement in ALS by generating transgenic mutant SOD1 mice on an Smn deficient background. Partial genetic disruption of Smn significantly worsened motor performance and survival in transgenic SOD1G93A mice. Furthermore, ALS-linked mutant SOD1 expression severely reduced SMN protein levels, but not transcript, in neuronal culture and mouse models from early presymptomatic disease. SMN protein depletion was linked to the nuclear compartment and a physical interaction between SMN and mutant SOD1 was confirmed in mouse spinal cord. Treatment with the environmental toxin paraquat also depleted SMN protein, implicating oxidative stress in the mechanism underlying SMN deficiency in familial ALS and potentially sporadic disease. In contrast, transgenic SOD1WT overexpression in SMA type I mice was incapable of modulating SMN protein levels or disease progression. These data establish that SMN deficiency accelerates phenotypic severity in transgenic familial ALS mice, consistent with an enhancing genetic modifier role. We therefore propose that SMN replacement and upregulation strategies considered for SMA therapy may have protective potential for ALS.