Knockdown of the survival motor neuron (Smn) protein in zebrafish causes defects in motor axon outgrowth and pathfinding

Abstract

Spinal muscular atrophy (SMA) is an autosomal recessive disorder characterized by a loss of alpha motoneurons in the spinal cord. SMA is caused by low levels of the ubiquitously expressed survival motor neuron (Smn) protein. As it is unclear how low levels of Smn specifically affect motoneurons, we have modeled SMA in zebrafish, a vertebrate model organism with well-characterized motoneuron development. Using antisense morpholinos to reduce Smn levels throughout the entire embryo, we found motor axon-specific pathfinding defects. Reduction of Smn in individual motoneurons revealed that smn is acting cell autonomously. These results show for the first time, in vivo, that Smn functions in motor axon development and suggest that these early developmental defects may lead to subsequent motoneuron loss.

Figure 1.

Motor axons/nerves are abnormal in embryos injected with 6 ng of smn MO. Western blot analysis of WT uninjected (lanes 1 and 3), control MO–injected (lane 2), and smn MO–injected (lane 4) embryos at 36 h (Smn is 37–38 kD). Hu-C, a neuronal marker, is shown as a loading control. Lateral views (anterior to the left; dorsal to the top) of whole-mount embryos labeled with znp1 mAb at 27 (C–E) and 36 h (F–H) in embryos injected with control MO (C and F) or smn MO (D, E, G, and H). Truncated motor axons/nerves (D and G; black arrowheads) and branched motor axons/nerves (E and H; black arrows) occur. White arrows demarcate the proximal axon pathway (C and F). Bars: (C–E) 25 μm; (F–H) 30 μm.

Figure 2.

Motor axons/nerves are abnormal in embryos injected with 9 ng of smn MO. Western blot analysis of WT uninjected (lanes 1 and 3), control MO–injected (lane 2), and smn MO–injected (9 ng) (lane 4) embryos at 36 h. Hu-C, a neuronal marker, is shown as a loading control. Lateral views of whole-mount embryos labeled with znp1 mAb at 27 (C–E) and 36 h (F–H) in embryos injected with control MO (C and F) or smn MO (D, E, G, and H). Truncated motor axons/nerves (D and G; black arrowheads) and branched motor axons/nerves (E and H; black arrows) occur when Smn protein levels are further reduced. Bars: (C–E) 25 μm; (F–H) 30 μm.

Figure 3.

Time-lapse imaging of defective motoneurons suggests that truncation defects precede branching defects. Western blot analysis of WT uninjected (lanes 1 and 3), control MO–injected (lane 2), and smn MO–injected (9 ng) (lane 4) embryos at 74 h. Synaptic vesicle protein 2, SV2, is shown as a loading control. Lateral views of the medial pathway of a GFP-expressing motor nerve of a transgenic gata2–GFP embryo injected with either control MO (C, E, and G; n = 10) or smn MO (D, F, and H; n = 13). Ventral projecting motoneurons are shown at 36 (C and D), 50 (E and F), and 74 h (G and H). White lines demarcate the second intermediate target, the ventral aspect of the notochord. Only one GFP-expressing motor nerve was imaged, compiled, and placed on an artificial black background; the additional trunk nerves have been removed from the image. Bar, 50 μm.

Figure 4.

Dorsally projecting motor nerves are also truncated and branched when Smn protein is decreased. Lateral view of a confocal image of a GFP-expressing, dorsally projecting motor nerves in a transgenic islet1–GFP embryo injected with either 9 ng of control MO (A; n = 50) or smn MO (B and C; n = 89) at 74 h. White arrowhead and asterisk (B) denote a truncated nerve. White arrow (C) indicates a nerve branch extending into the adjacent hemisegment. Bar, 50 μm.

Figure 5.

Decreased levels of Smn do not initially result in motoneuron cell death. Cross section view (dorsal to the top) of TUNEL-stained 50-h control MO–injected (A; n = 3) and smn MO–injected (9 ng) (B; n = 15) embryos. TUNEL-positive (purple) cells are present on the skin of both the smn MO– and control MO–injected embryos, which is not uncommon for this assay. Occasionally, (B) TUNEL-positive cells (black arrowhead) present in the dorsal spinal cord, corresponding to the sensory Rohon-Beard neurons (entire spinal cord demarcated by black dashed lines). Cross section view of zn-5 mAb–stained secondary motoneuron cell bodies (arrows) of 50-h control MO–injected (C; n = 6) and smn MO–injected (9 ng) (D; n = 41) embryos. (E) Motor pool diameter (μm) was measured on medial-lateral (M-L) and dorsal-ventral (D-V) axes (error bars indicate standard error). Anterior trunk was defined as hemisegments 1–5, whereas middle trunk was defined as hemisegments 6–14. t test concluded that P > 0.1 among motor pools of control MO– and smn MO–injected embryos. Bar, 25 μm.

Figure 6.

Other neurons and axon tracts are unaffected upon knockdown of Smn protein. Dorsal view (anterior to the top) of the hindbrain stained with acetylated tubulin mAb in 27-h control MO–injected (A; n = 40) and smn MO–injected (9 ng) (B; n = 39) embryos showing the medial longitudinal fascicle (black arrowheads) and the lateral longitudinal fascicle (black arrows). Dorsal view of the hindbrain Mauthner neuron cell body and axon (arrowhead) stained with 3A10 mAb in 34-h control MO–injected (C; n = 61) and smn MO–injected (D; n = 32) embryos. Lateral views of (E and F) Rohon-Beard (black arrow) and (G and H) lateral line sensory neurons (white arrow) stained with acetylated tubulin mAb in 34-h control MO–injected (E and G; n = 40) and smn MO–injected (F and H; n = 39) embryos. Bars: (A and B) 50 μm; (C and D) 40 μm; (E–H) 50 μm.

Figure 7.

Muscle development is normal when Smn is decreased. Nomarski lateral views of mid-trunk muscle from 50-h gata2–GFP transgenic embryos injected with 9 ng of control MO (A; n = 20) or smn MO (B; n = 20) showing somitic boundaries (black arrows). Dorsal views of 22-h whole-mount in situ hybridization of myoD (purple; black arrowhead) in control MO–injected (C; n = 31) and smn MO–injected (D; n = 18) embryos. Cross section of 27-h znp1 (motor axon; arrows) and F59 (slow muscle; arrowheads) mAb–stained embryos injected with control MO (E; n = 40) or smn MO (F; n = 65). Cross section of 27-h znp1 (motor axon; arrows) and F310 (fast muscle; arrowheads) mAb–stained embryos injected with control MO (G; n = 20) or smn MO (H; n = 41). Bars: (A–D) 75 μm; (E–H) 30 μm.

Figure 8.

AChR clustering is normal when Smn is decreased. Lateral views of znp1 (A and D; motor axons, green) mAb, α-bungarotoxin (B and E; AChR, red) stained, and merge (C and F; yellow) 74-h larvae injected with control MO (A–C; n = 18) or smn MO (D–F; n = 49). Aberrant motor nerve (arrowhead) retains ability to cluster AChR. Bar, 50 μm.

Figure 9.

Smn is acting cell autonomously in CaP motoneurons. Lateral views of psuedocolor images of live CaP motoneurons directly iontophoresed with rhodamine dextran and control MO (A; n = 10) or smn MO (B and C; n = 21) visualized and imaged at 43 h. CaP motoneurons iontophoresed with smn MO exhibit axon branching (B) and truncation (C). (D) Percentage of smn MO–injected CaP motoneurons with defective and normal axons quantitated (error bars represent standard error). HM, nascent horizontal myoseptum; VNC, myotome adjacent to the ventral edge of the notochord. Bar, 50 μm.