Postsymptomatic restoration of SMN rescues the disease phenotype in a mouse model of severe spinal muscular atrophy

Abstract

Spinal muscular atrophy (SMA) is a common neuromuscular disorder in humans. In fact, it is the most frequently inherited cause of infant mortality, being the result of mutations in the survival of motor neuron 1 (SMN1) gene that reduce levels of SMN protein. Restoring levels of SMN protein in individuals with SMA is perceived to be a viable therapeutic option, but the efficacy of such a strategy once symptoms are apparent has not been determined. We have generated mice harboring an inducible Smn rescue allele and used them in a model of SMA to investigate the effects of turning on SMN expression at different time points during the course of the disease. Restoring SMN protein even after disease onset was sufficient to reverse neuromuscular pathology and effect robust rescue of the SMA phenotype. Importantly, our findings also indicated that there was a therapeutic window of opportunity from P4 through P8 defined by the extent of neuromuscular synapse pathology and the ability of motor neurons to respond to SMN induction, following which restoration of the protein to the organism failed to produce therapeutic benefit. Nevertheless, our results suggest that even in severe SMA, timely reinstatement of the SMN protein may halt the progression of the disease and serve as an effective postsymptomatic treatment.

Figure 1. Construction of the inducible Smn rescue allele.

The structure of the Smn inducible rescue allele prior to and following Cre-mediated recombination, depicting the expected transcripts from the 2 different conformations. The WT murine Smn allele and its single FL-SMN transcript are included for reference. Drawing is not to scale.

Figure 2. In vivo characterization of the inducible Smn rescue allele.

(A) Recombination of the rescue allele is observed in the presence but not in the absence of Cre recombinase in P18 mice. WT murine Smn-specific primers were used to amplify samples in lanes numbered 1. Pre- and postrecombination-specific primers were used to amplify DNA in lanes numbered 2 and 3, respectively. (B) In the absence of Cre recombinase, the rescue allele fails to express FL-SMN transcript in mouse tissue. Perdurance of the SMNΔ7 transcript in EIIa-Cre^+/–;Smn^Res/+ mice indicates incomplete recombination, perhaps due to mosaic expression of Cre recombinase. The first 2 lanes are amplified products from bacterial clones into which were cloned the SMNΔ7 and FL-SMN fragments, respectively. (C) Human SMN2 rescues Smn^Res/Res embryonic lethality but does not prevent appearance of SMA symptoms in neonates (arrows). (D) Cre-mediated recombination of the rescue allele in the germline results in near normal levels of SMN protein by Western blotting and prevents (E) muscle pathology and (F) loss of spinal motor neurons in EIIaCre;Smn^Res/Res adult (P60-70) mice. n ≥ 2 for each genotype, P > 0.05, Student’s t test. Scale bars: 170 μm (E); 117 μm (F). Data are represented as mean ± SEM.

Figure 3. Restoration of the SMN protein in symptomatic SMA model mice rescues the disease phenotype.

(A) Western blotting reveals disease state levels of spinal cord SMN protein in nontreated P4 animals homozygous for the hybrid allele. (B) NMJ pathology in untreated P4 mutants manifests as poor terminal arborization and NF engorged nerve terminals (arrowheads). (C) Postsymptomatic TM treatment of Cre-ER–positive but not Cre-ER–negative mutant mice overcomes disease-related downturn in growth at P13 and accounts for a significant increase in weight by P17 (n ≥ 16, **P < 0.01, Student’s t test). (D) Improved motor performance in a righting reflex assay (n ≥ 4) following TM treatment at P4. (E) Mutant SMN2;Δ7;Cre-ER;Smn^Res/Res mice treated at P4 with TM and sacrificed at P16–P17 express increased SMN protein in nervous as well as nonnervous tissue. Controls were WT (Smn^+/+) at the murine Smn locus. (F) NMJs of gastrocnemius muscle of TM-treated mutants exhibit a reduction but not complete absence of pathology at P17 (solid arrowhead), whereas those of untreated mutants become progressively worse, displaying NF swollen nerve terminals (arrows) and immature, plaque-like motor endplates (open arrowheads). (G) Significant increase in the size of the motor endplate of TM-treated mutants compared with those of untreated mutants (100 endplates assessed in each of 3 mice; *P < 0.05; **P < 0.01, Student’s t test). (H) An increase in SMN-mediated snRNP assembly is observed in TM-treated mutants at 4 days after treatment. n = 4, *P < 0.05, Student’s t test. Scale bars: 10 μm (B); 40 μm and 10 μm (enlarged images of NMJs). (F). Data are represented as mean ± SEM.

Figure 4. Postsymptomatic restoration of SMN enhances survival of SMA model mice.

(A) Restoration of SMN at P4 results in a 23-fold increase in survival of approximately 50% of mutant mice. Control animals were also treated with TM at P4 and are WT for murine Smn on at least 1 allele. χ² (treated versus untreated mutants) = 48.65, P < 0.0001, log-rank test. (B) Adult TM-treated mutants have approximately 60% the weight of WT controls (**P < 0.01) but are phenotypically indistinguishable when tested in (C) grip strength (corrected for body weight), (D) rotarod, and (E) open-field assays. n ≥ 5, P > 0.05 in each case for C–E, Student’s t test. Data are represented as mean ± SEM.

Figure 5. A limited therapeutic window in SMA model mice harboring the rescue allele.

SMN restoration at earlier time points generally correlated with (A) greater increases in body weight (n ≥ 4 for each genotype except P10 mice = 3; *P < 0.05, Student’s t test) and (B) enhanced motor performance assessed in righting ability tests (n ≥ 4 for each genotype, **P < 0.01; ***P < 0.001; 1-way ANOVA). (C) Equivalent increases in spinal cord SMN protein expression are observed in mutants bearing the rescue allele treated between P4 and P10. (D) Earlier restoration of SMN protein translates into greater phenotypic rescue in Kaplan-Meier survival curves. (E) Loss of spinal motor neurons and (F) gems is significantly mitigated following treatment only at the earliest (P4) time point (n = 3 for each cohort, ***P < 0.001; 1-way ANOVA). Data are represented as mean ± SEM.

Figure 6. Level of phenotypic rescue is inversely correlated with the extent of neuromuscular synapse pathology.

(A) Immunohistochemistry of cervical spinal cord sections from mice administered TM at various time points during the course of the disease revealed gems (arrows) and normal cytoplasmic staining only in P4-treated mutants. Scale bars: 125 μm and 5 μm (insets). The timing of SMN restoration correlates closely with (B) AChR cluster development as well as (C) the ability to clear presynaptic NF aggregates by P17. Early treatment resulted in greater sophistication of the postsynapse and reduced NF in nerve terminals. *P < 0.05; **P < 0.01; ***P < 0.001; 1-way ANOVA. Note: more than 500 AChR clusters (for postsynaptic specialization) and more than 50 endplates (for NF accumulation) were examined in at least 3 mice per time point. Data are represented as mean ± SEM.