Rescue of the spinal muscular atrophy phenotype in a mouse model by early postnatal delivery of SMN

Abstract

Spinal muscular atrophy (SMA), the most common autosomal recessive neurodegenerative disease affecting children, results in impaired motor neuron function. Despite knowledge of the pathogenic role of decreased survival motor neuron (SMN) protein levels, efforts to increase SMN have not resulted in a treatment for patients. We recently demonstrated that self-complementary adeno-associated virus 9 (scAAV9) can infect approximately 60% of motor neurons when injected intravenously into neonatal mice. Here we use scAAV9-mediated postnatal day 1 vascular gene delivery to replace SMN in SMA pups and rescue motor function, neuromuscular physiology and life span. Treatment on postnatal day 5 results in partial correction, whereas postnatal day 10 treatment has little effect, suggesting a developmental period in which scAAV9 therapy has maximal benefit. Notably, we also show extensive scAAV9-mediated motor neuron transduction after injection into a newborn cynomolgus macaque. This demonstration that scAAV9 traverses the blood-brain barrier in a nonhuman primate emphasizes the clinical potential of scAAV9 gene therapy for SMA.

Conflict of interest statement

COMPETING INTERESTS STATEMENT

The authors declare no competing financial interests.

Figures

Figure 1

Phenotypic correction of SMA mice injected on P1. (a) Injection of scAAV9-GFP in SMA animals results in GFP expression (green) within dorsal root ganglia and motor neurons (ChAT staining in red) in the lumbar spinal cord 10-d post-injection. (b) Western blots from tissues of control, scAAV9-SMN–treated and untreated SMA animals show elevated levels of SMN expression in SMN-treated animals compared to control animals, although levels are still lower than those of control littermates. Quantifications of western blots are available in the Supplementary Figures 1 and 7. (c) Righting ability shows that SMN-treated animals can right themselves similarly to control animals by P13. (d) SMA animals treated with scAAV9-SMN are larger than GFP-treated animals. (e) scAAV9-SMN treatment of SMA animals results in greatly extended survival over GFP treatment. (f) Body weight assessments show an increase in animals treated with scAAV9-SMN versus those treated with GFP. Scale bars, 200 μm (a); 50 μm (a inset).

Figure 2

Effects of SMN treatment at P1 on NMJs of adult SMA mice. Untreated SMA mice do not survive to adulthood. (a) scAAV9-SMN treatment restores endplate currents (EPC) in ~90-d-old SMA animals. In control mice, the mean EPC amplitude was 82.6 ± 3.5 nA, and in treated SMA mice, it was 83.4 ± 4.1 nA (P = 0.89, n = 4 mice for each group). (b) Affected animals treated with scAAV9-SMN had an increase in miniature endplate currents. (c) Both control and treated SMA endplate currents had a similar degree of depression during 50 Hz nerve stimulation. (d–i) Representative sections from the transverse abdominis (TVA), a proximal muscle with innervation abnormalities in SMA mice, shows normal innervation in both wild-type (d–f) and SMN-treated (g–i) animals. Scale bars, 10 μm.

Figure 3

Systemic injection of scAAV9-GFP into SMA mice of varying ages. (a–c) Animals injected on P2 have a transduction pattern identical to P1-injected animals, with motoneuron transduction in lumbar spinal cord. (d–f) P5-injected animals have more glial transduction and less motoneuron (f inset, arrow) transduction than younger animals in lumbar spinal cord analysis. (g–i) The pattern of increasing glial transduction continues in P10-injected animals. GFP (green), ChAT (red, a motoneuron marker) and merged (yellow). (j–k) scAAV9-SMN injection on P2 in SMA animals rescues life span and increases body weight (n = 6), whereas P5 scAAV9-SMN delivery in SMA animals only partially rescues life span and body weight (n = 4) compared with control scAAV9-GFP–treated (n = 10). No increase in life span or body weight was seen in mice treated with scAAV9-SMN on P10 (n = 4). (l–q) Systemic injection of scAAV9-GFP into a cynomolgus macaque on P1 results in a similar transduction pattern within the spinal cord as previously shown in P1-injected mice. GFP (l,o), ChAT (m,p) and merged (n,q) images from thoracic spinal cord demonstrate motor neuron transduction. A representative longitudinal section is shown in (l–n), indicating transduction along the neuraxis. Transverse sections (o–q) mimic the pattern of dorsal root ganglia and motor neuron transduction seen in P1-injected mice. Inset scale bars, 50 μm; c,f,i, 100 μm; n,q, 200 μm.