Inhibition of Crmp1 Phosphorylation at Ser522 Ameliorates Motor Function and Neuronal Pathology in Amyotrophic Lateral Sclerosis Model Mice

Abstract

Amyotrophic lateral sclerosis (ALS) is a rapidly progressive and fatal neurodegenerative disorder that affects upper and lower motor neurons; however, its pathomechanism has not been fully elucidated. Using a comprehensive phosphoproteomic approach, we have identified elevated phosphorylation of Collapsin response mediator protein 1 (Crmp1) at serine 522 in the lumbar spinal cord of ALS model mice overexpressing a human superoxide dismutase mutant (SOD1^G93A). We investigated the effects of Crmp1 phosphorylation and depletion in SOD1^G93A mice using Crmp1^S522A (Ser522→Ala) knock-in (Crmp1^ki/ki ) mice in which the S522 phosphorylation site was abolished and Crmp1 knock-out (Crmp1^-/-) mice, respectively. Crmp1^ki/ki /SOD1^G93A mice showed longer latency to fall in a rotarod test while Crmp1^-/-/SOD1^G93A mice showed shorter latency compared with SOD1^G93A mice. Survival was prolonged in Crmp1^ki/ki /SOD1^G93A mice but not in Crmp1^-/-/SOD1^G93A mice. In agreement with these phenotypic findings, residual motor neurons and innervated neuromuscular junctions (NMJs) were comparatively well-preserved in Crmp1^ki/ki /SOD1^G93A mice without affecting microglial and astroglial pathology. Pathway analysis of proteome alterations showed that the sirtuin signaling pathway had opposite effects in Crmp1^ki/ki /SOD1^G93A and Crmp1^-/-/SOD1^G93A mice. Our study indicates that modifying CRMP1 phosphorylation is a potential therapeutic strategy for ALS.

Keywords: ALS; CRMP1; SOD1.

Figure 1.

Phosphoproteomic analysis of spinal cords from WT and SOD1^G93A mice at 20 weeks of age. A, Top 10 canonical pathways identified based on the molecules that were differentially expressed (max fold change >1.5, ANOVA p < 0.05) between WT and SOD1^G93A mice in phosphoproteomics. B, A list of phosphoproteins in semaphorin neuronal repulsive signaling pathway with significant expression changes in the lumbar spinal cords of SOD1^G93A mice compared with WT mice (max fold change 1.5; p < 0.05). Phosphorylation sites, fold-change levels, and p-value are also shown. C, Protein–protein interaction (PPI) network of differentially upregulated proteins in SOD1^G93A mice visualized by STRING: https://string-db.org/. Red nodes are phosphoproteins associated with axon guidance in reactome pathway. The red arrow indicates Crmp1.

Figure 2.

Phenotypic comparisons of SOD1^G93A, Crmp1^−/−/SOD1^G93A, and Crmp1^ki/ki/SOD1^G93A mice. This figure is supported by Extended Data Figures 2-1, 2-2. A, Kaplan–Meier survival curves for SOD1^G93A (black), Crmp1^−/−/SOD1^G93A (blue), and Crmp1^ki/ki/SOD1^G93A (red) mice. Median survival duration in SOD1^G93A, Crmp1^−/−/SOD1^G93A, and Crmp1^ki/ki/SOD1^G93A mice was 167, 173, and 180 d, respectively (the log-rank test, p = 0.044). B, The comparative survival durations for three groups (two-way ANOVA, *p < 0.05). C, Kaplan–Meier curves for disease onset. D, Median disease onset in SOD1^G93A, Crmp1^−/−/SOD1^G93A, and Crmp1^ki/ki/SOD1^G93A mice was 112, 102, and 112 d, respectively, and the differences were not significant (the log-rank test, p = 0.797). E, Mean disease duration (days from onset to end stage). Mean disease duration in SOD1^G93A, Crmp1^−/−/SOD1^G93A, and Crmp1^ki/ki/SOD1^G93A mice was 54, 63, and 65 d, respectively. Moreover, the duration was longer in Crmp1^ki/ki/SOD1^G93A mice than in SOD1^G93A mice (one-way ANOVA with Fisher’s LSD test, p = 0.039). F, Rotarod test. Crmp1^ki/ki/SOD1^G93A mice exhibited a significant improvement in motor function at the late stage (21–24 w) compared with SOD1^G93A mice, while Crmp1^−/−/SOD1^G93A showed shorter latency to fall during the rotarod test at 18 w. Values are means ± SD (SOD1^G93A, n = 20; Crmp1^−/−/SOD1^G93A, n = 20; Crmp1^ki/ki/SOD1^G93A, n = 20). *,^†p < 0.05, **p < 0.01 (two-way ANOVA with Fisher’s LSD test). G, Body weight. Significant differences were not observed between the three lines of model mice. N.S. = not significant.

Figure 3.

Motor neuron degeneration and gliosis. This figure is supported by Extended Data Figure 3-1. A, Representative images of ChAT-stained motor neurons in the lumbar spinal cord of mice at 20 w. GFAP-immunostained (B) and Iba1-immunostained (C) images of the lumbar spinal cord. Scale bar: 200 μm. D, Counts of ChAT-positive neurons in hemi sections of the lumbar spinal cord (WT; n = 4, SOD1^G93A; n = 7, Crmp1^−/−/SOD1^G93A; n = 8, Crmp1^ki/ki/SOD1^G93A; n = 7). Significance was determined by one-way ANOVA with Tukey’s multiple comparisons test as follows: WT versus SOD1^G93A; p = 0.003, WT versus Crmp1^−/−/SOD1^G93A; p < 0.001, SOD1^G93A versus Crmp1^−/−/SOD1^G93A; p = 0.871, WT versus Crmp1^ki/ki/SOD1^G93A; p = 0.496, SOD1^G93A versus Crmp1^ki/ki/SOD1^G93A; p = 0.0253. Percentage of GFAP-positive (E) and Iba1-positive (F) area within the ventral horn area (WT; n = 5, SOD1^G93A; n = 5, Crmp1^−/−/SOD1^G93A; n = 5, Crmp1^ki/ki/SOD1^G93A; n = 5). Significance was determined by one-way ANOVA with Dunnett’s multiple comparisons test as follows: (E) SOD1^G93A versus Crmp1^−/−/SOD1^G93A; p = 0.723, SOD1^G93A versus Crmp1^ki/ki/SOD1^G93A; p = 0.292, (F) SOD1^G93A versus Crmp1^−/−/SOD1^G93A; p = 0.457, SOD1^G93A versus Crmp1^ki/ki/SOD1^G93A; p = 0.999. Values are means ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, N.S. = not significant as determined by one-way ANOVA.

Figure 4.

Blocking Crmp1 phosphorylation at S522 delays denervation at NMJs. A, Representative photomicrographs of NMJs in fixed TA muscles. Nerve axons (green) are stained with 2H3 (NF) plus SV2 (synaptic vesicles) and postsynaptic acetylcholine receptors (red) are stained with α-BTX. Scale bar: 20 μm. B, Crmp1 S522A expression increases endplate occupancy in TA muscle of SOD1^G93A mice. The figure shows the percentage of fully innervated (fully occupied), partially denervated (partially occupied), and denervated endplates of the axon terminals in TA muscles from mice with the indicated genotypes. Statistical significance was determined as follows: innervated: WT versus SOD1^G93A; p < 0.001, SOD1^G93A versus Crmp1^−/−/SOD1^G93A; p = 0.999, SOD1^G93A versus Crmp1^ki/ki/SOD1^G93A; p = 0.038, denervated: WT versus SOD1^G93A; p < 0.001, SOD1^G93A versus Crmp1^−/−/SOD1^G93A; p = 0.955, SOD1^G93A versus Crmp1^ki/ki/SOD1^G93A; p = 0.01, by one-way ANOVA with Dunnett’s multiple comparisons test. Values are means ± SEM (n = 3); *,^†p < 0.05, ^††p < 0.01, ***p < 0.001 by one-way ANOVA with Dunnett’s multiple comparisons test.