Survival motor neuron protein in motor neurons determines synaptic integrity in spinal muscular atrophy

Abstract

The inherited motor neuron disease spinal muscular atrophy (SMA) is caused by deficient expression of survival motor neuron (SMN) protein and results in severe muscle weakness. In SMA mice, synaptic dysfunction of both neuromuscular junctions (NMJs) and central sensorimotor synapses precedes motor neuron cell death. To address whether this synaptic dysfunction is due to SMN deficiency in motor neurons, muscle, or both, we generated three lines of conditional SMA mice with tissue-specific increases in SMN expression. All three lines of mice showed increased survival, weights, and improved motor behavior. While increased SMN expression in motor neurons prevented synaptic dysfunction at the NMJ and restored motor neuron somal synapses, increased SMN expression in muscle did not affect synaptic function although it did improve myofiber size. Together these data indicate that both peripheral and central synaptic integrity are dependent on motor neurons in SMA, but SMN may have variable roles in the maintenance of these different synapses. At the NMJ, it functions at the presynaptic terminal in a cell-autonomous fashion, but may be necessary for retrograde trophic signaling to presynaptic inputs onto motor neurons. Importantly, SMN also appears to function in muscle growth and/or maintenance independent of motor neurons. Our data suggest that SMN plays distinct roles in muscle, NMJs, and motor neuron somal synapses and that restored function of SMN at all three sites will be necessary for full recovery of muscle power.

Figure 1.

Cre is expressed in tissue-specific patterns in ChAT^Cre, MyoD^iCre, and Myf5^Cre mice. A, Representative images of lacZ staining of whole mount E15.5 embryos in MyoD^iCre+ lacZ (n = 2) and Myf5^Cre+ lacZ lines (n = 5). Scale bar, 2.5 mm. B, Representative images of lacZ staining of spinal cord (top row) and skeletal muscle cross sections (bottom row) from P7 ChAT^Cre+ lacZ, MyoD^iCre+ lacZ, and Myf5^Cre+ lacZ mice (n = 2 each). *Indicates a DRG. Arrow indicates neurons in ventral horn of spinal cord. Scale bar, 500 μm. C, Representative images of ChAT/lacZ IHC in spinal cords from ChAT^Cre+ lacZ (n = 3 lacZ-positive), MyoD^iCre+ lacZ (n = 3 lacZ-positive, 3 lacZ-negative) and Myf5^Cre+ lacZ mice (n = 4 lacZ-positive, 3 lacZ-negative examined). Punctate lacZ staining is evident in the cytoplasm and diffuse staining is visible in the nuclei of all motor neurons in ChAT^Cre mice. LacZ staining is present in some, but not all motor neurons in Myf5^Cre mice (indicated by arrows) and in other cells of the spinal cord (indicated by arrowheads). Scale bar, 50 μm.

Figure 2.

SMN is expressed in a tissue-specific manner in conditional SMA mice. A, Schematic demonstrating the exonic structures of the possible transcript products of the Smn^Res hybrid allele. In the absence of Cre recombinase, truncated hybrid transcript (SMN 68) is made containing upstream mouse Smn (exons 1–6), human SMN2 exon 8, and no exon 7 sequence. In the presence of Cre recombinase, a full-length hybrid transcript (SMN 678) containing mouse Smn exons 1–7 and human SMN2 exon 8 sequence may be made. The positions of qRT-PCR primers used are indicated with arrows indicating forward and reverse primers and asterisks indicating the positions of the probes. SMN6m8h primers detect both truncated and full-length hybrid transcripts, the SMN67m8h primers only detect full-length transcripts, and the SMN12m primers were used as an endogenous control. B, SMN6m8h primers detect hybrid transcripts in both spinal cord and muscle of P10 Cre+ Het, Cre− SMA, and Cre+ SMA mice, but not in WT mice. C, SMN67m8-h primers detect SMN 678 transcripts in spinal cord only in ChAT^Cre+ and to a lesser extent, in Myf5^Cre+ Het and SMA mice. SMN 678 transcripts are expressed in muscle only in MyoD^iCre+ and Myf5^Cre+ Het and SMA mice (n = 3–8 per genotype). D, SMN protein levels were determined in spinal cord (top row) and muscle (bottom row) tissues of P10 WT, Cre− SMA, and Cre+ SMA mice from each conditional SMA line. Both full-length (FL) and truncated (Δ7) SMN protein isoforms are visible and are compared with the endogenous controls, actin or GAPDH. E, Densitometry analysis of FL SMN protein expression shows reduced FL SMN expression in Cre− SMA mice compared with WT mice in all three lines in both spinal cord and muscle. FL-SMN expression is significantly increased in the muscles of MyoD^iCre+ and Myf5^Cre+ SMA mice compared with Cre− SMA mice. *p < 0.01. F, SMN6m8h primers detect hybrid transcripts in the DRGs of all lines of Cre+ SMA mice, but (G) SMN67m8 h primers detect SMN 678 transcripts only in DRGs of Myf5^Cre+ SMA mice, not ChAT^Cre+ or MyoD^iCre+ SMA mice (n = 2–3 per genotype). H, SMN6m8-h primers and I, SMN67m8h primers detect hybrid transcripts that are expressed in the appropriate tissue-specific patterns at early developmental time points. (All tissues isolated from mice at P2 except 2 of 3 Myf5^Cre+ SMA mice, which were isolated at P5). n = 2–4 per genotype.

Figure 3.

Increased SMN expression in motor neurons or muscle increases survival and weights of SMA mice. A, Survival curves of ChAT^Cre (n = 15 Cre− SMA, 28 Cre+ SMA mice), MyoD^iCre (n = 16 Cre− SMA, 28 Cre+ SMA mice), and Myf5^Cre (n = 13 Cre− SMA, 15 Cre+ SMA mice) mice indicate increased survival of Cre+ SMA mice in all 3 lines (log rank p < 0.0005). B, Weight curves are improved in all three lines, p < 0.001 (ChAT^Cre: n = 11 WT, 12 Cre− SMA, and 27 Cre+ SMA, MyoD^iCre: n = 11 WT, 12 Cre− SMA, and 18 Cre+ SMA, Myf5^Cre: n = 10 WT, 14 Cre− SMA, and 15 Cre+ SMA mice).

Figure 4.

Motor behavior is improved in conditional SMA mice. A, Righting latency is reduced in Cre+ SMA mice compared with Cre− SMA mice in all 3 lines of mice, p < 0.0001 (ChAT^Cre: n = 11 WT, 12 Cre− SMA, and 27 Cre+ SMA, MyoD^iCre: n = 11 WT, 12 Cre− SMA, and 18 Cre+ SMA, Myf5^Cre: n = 10 WT, 14 Cre− SMA, and 15 Cre+ SMA mice). B, Cumulative tube test scores are increased in MyoD^iCre+ SMA mice compared with Cre− SMA mice, p = 0.01 (ChAT^Cre: n = 5 WT, 6 Cre− SMA, and 13 Cre+ SMA, MyoD^iCre: n = 5 WT, 6 Cre− SMA, and 7 Cre+ SMA, Myf5^Cre: n = 5 WT, 6 Cre− SMA, and 6 Cre+ SMA mice). C, Ambulation indices including time upright (left), numbers of grids crossed (middle), and number of rearings (right) is equivalent between ChAT^Cre+ SMA mice and WT mice between P13 and P20 (n = 3–6 WT, 5–6 Cre+ SMA mice).

Figure 5.

Increased SMN expression in motor neurons, but not muscle results in increased motor neuron number. A, Representative images of L1 spinal cord ventral horn and ChAT-positive motor neurons in ChAT^Cre, MyoD^iCre, and Myf5^Cre WT, Cre− SMA, and Cre+ SMA mice. Scale bar, 50 μm. B, Quantification of total number of L1 motor neurons (ChAT^Cre: n = 4 WT, 4 Cre− SMA, and 4 Cre+ SMA, MyoD^iCre: n = 4 WT, 3 Cre− SMA, and 4 Cre+ SMA, Myf5^Cre: n = 4 WT, 6 Cre− SMA, and 5 Cre+ SMA mice). *p = 0.01.

Figure 6.

SMN expression in muscle restores myofiber size. A, Representative images of TA muscle cross sections from ChAT^Cre, MyoD^iCre, and Myf5^Cre WT, Cre− SMA, and Cre+ SMA mice. Scale bar, 50 μm. B, Quantification of total muscle area and myofiber diameter (ChAT^Cre: n = 3 WT, 3 Cre− SMA, and 3 Cre+ SMA, MyoD^iCre: n = 3 WT, 4 Cre− SMA, and 4 Cre+ SMA, Myf5^Cre: n = 3WT, 3 Cre− SMA, and 3 Cre + SMA mice). *p = 0.01, **p = 0.001.

Figure 7.

NMJ physiology and structure is improved by expression of SMN in motor neurons. A, Representative EPC traces from ChAT^Cre WT (black), Cre− SMA (red), and Cre+ SMA mice (blue). B, Quantification of EPC amplitudes, MEPC amplitudes, quantal content, and facilitation in the three lines of mice (n = 16 WT, 10 Cre− SMA, 6 ChAT^Cre+ SMA, 6 MyoD^iCre+ SMA, and 6 Myf5^Cre+ SMA mice). *p ≤ 0.05, **p < 0.001. C, Representative images of NMJs from the splenius capitis muscle in ChAT^Cre, MyoD^iCre, and Myf5^Cre WT, Cre− SMA, and Cre+ SMA mice (red, α-bungarotoxin; green, NF; blue, synaptophysin). Arrows indicate presynaptic terminals with NF accumulations. Asterisks indicate fully or partially denervated NMJs. Scale bar, 10 μm. D, Quantification of TA muscle presynaptic terminal NF accumulation, TA muscle endplate size, and innervation status of the splenius capitis muscle in the 3 lines of mice (ChAT^Cre: n = 4 WT, 4 Cre− SMA, 4 Cre+ SMA, MyoD^iCre: n = 3 WT, 4 Cre− SMA, and 4 Cre+ SMA, Myf5^Cre: n = 3 WT, 3 Cre− SMA, 3 Cre+ SMA mice). *p < 0.05, **p < 0.001.

Figure 8.

Motor neuron expression of SMN improves motor neuron somal synapses. A, Representative images of VGluT1+ synapses on L1 motor neurons in ChAT^Cre, MyoD^iCre, and Myf5^Cre WT, Cre− SMA, and Cre+ SMA mice (red, ChAT; green, VGluT1). Scale bar, 10 μm. B, Quantification of VGluT1+ synapses in the three lines [ChAT^Cre: n = 3 WT (17 neurons), 4 Cre− SMA (20 neurons), 3 Cre+ SMA (22 neurons), MyoD^iCre: n = 3 WT (16 neurons), 4 Cre− SMA (17 neurons), and 3 Cre+ SMA (21 neurons), Myf5^Cre: n = 3 WT (14 neurons), 3 Cre− SMA (21 neurons) and 4 Cre+ SMA (15 neurons)]. *p ≤ 0.5. C, Representative EM images of motor neuron somal synapses in ChAT^Cre WT, Cre− SMA, and Cre+ SMA mice. S, soma; B, bouton; arrowheads indicate PSDs. Scale bar, 1 μm. D, Quantification of synapse number, PSD length, overall synaptic vesicle density, and active zone synaptic vesicle density in ChAT^Cre WT, Cre− SMA, and Cre+ SMA mice (motor neuron#: n = 14 WT, 18 Cre− SMA, and 14 Cre+ SMA). *p < 0.05, **p < 0.001.