Fig 1.
Developmental synaptic degeneration in the absence of Schwann cells.
(A) Genetic ablation of Schwann cells, similar to erbB inactivation, leads to developmental synaptic degeneration of the neuromuscular junction (NMJ). Diaphragms were dissected at E15.5 from Wnt1-Cre:Rosa26LoxSTOPLox Diptheria Toxin A Chain (Wnt1-DTA; right column) and wild-type mice (+/+; left columns), and stained with antibodies against S100 to label Schwann cells (green) and vesicular acetylcholine transporter (VAChT) to label presynaptic motor axon terminals (red). Scale bar = 70 μm. Representative example of n = 4. (B) Low- and high-power images (top and bottom panels, respectively) of diaphragm muscle at E14.25 show a strikingly higher percentage of α-bungarotoxin (α-BTX)-labeled postsynaptic nicotinic acetylcholine receptors (AChRs; red in bottom panels) receiving contact from synaptophysin-positive (Syp) motor axon terminals (green in top and bottom panels) in erbB3 mutant (erbB3-/-; right panels) vs. wild-type (erbB3+/+; left panels) mice; arrows denote unapposed AChRs, arrowheads denote apposed or innervated AChRs. Scale bar in top panels = 250 μm; in lower panels = 50 μm.
Fig 2.
Evoked activity through muscle-derived AChRs is required for developmental synaptic degeneration induced by Schwann cell ablation.
(A) E15.5 diaphragms from the indicated genotypes were dissected and immunostained with antibodies against neurofilament. In contrast to the absence of motor innervation observed in erbB3 mutant (erbB3-/-) mice, erbB3 mutant mice lacking ChAT, Snap25, or AChRα1 exhibit a complete lack developmental synaptic degeneration. The rescued axons are spread out as a consequence of the absence of Schwann cell-mediated fasciculation. Scale bar = 1000 μm. At least n = 3 for every genotype. (B) The motor endplate band is correctly positioned in the central region of Snap25; erbB3 double mutant diaphragm. E15.5 diaphragms were labeled with antibodies against motor nerve (anti-neurofilament; NF; green), the presynaptic marker vesicular acetylcholine transporter (VAChT; red), and Cy5-α-BTX (blue). Scale bar = 150 μm. (C) Rescued synapses in E18.5 Snap25; erbB3 double mutant diaphragm show apposition of VAChT-rich nerve terminals to postsynaptic, α-BTX-labeled AChRs. Scale bar = 100 μm, n = 3. (D) The relative number of VAChT-positive innervated AChRs is higher in E15.5 erbB3; Snap25 double mutant vs. wild-type mice (similar to the increased innervation observed in single erbB3 mutant vs. wild-type mice at E14.25), but is equal between these genotypes by E18.5 (n = 3 diaphragms for wild-type and double mutant mice, 50 NMJs counted per diaphragm).
Fig 3.
Functional genomic analysis of genes differentially regulated in diaphragm muscle containing (wild-type; erbB3+/+) or lacking (mutant; erbB3-/-) peripheral Schwann cells.
(A) Gene Ontology (GO) term networks from the results of GO analysis of the set of genes significantly upregulated in erbB3 wild-type vs. mutant muscle were overlapped in Cytoscape. Two individual comparisons (each one between erbB3 wild-type- and mutant-derived muscle samples) were performed and are represented by the blue and green lines. The number of gene members of each term, and degree of overlap, or genes common, between multiple terms are represented by node and edge attributes. Highly interconnected nodes with overlap from both comparisons are particularly noteworthy. (B) Heatmap showing fold-changes in gene expression of some individual members of the GO terms differentially regulated in erbB3 wild-type and mutant muscle. Expression values were determined by the number of mapped reads normalized to gene length and depth of sequencing.
Table 1.
Functional categories of genes upregulated in muscle in the presence of Schwann cells.
Fig 4.
Thrombin causes motor axon degeneration in vitro and is blocked by pre-incubation with glia- but not muscle-conditioned medium.
E12.5 cervical spinal explants from HB9:GFP mice were grown on laminin, treated at plating with 5 nM GDNF in B27-containing neurobasal (B27-NB) medium, and re-imaged 24 hours after specific treatments, and the number of GFP-positive motor axons with pathological swelling or other signs of degeneration were quantified at pre- and post-treatment intervals. (A) Representative images of explants each treated with GDNF at plating and then treated with GDNF (top panels) or 200 nM recombinant thrombin (lower panels) one day after plating. Images were captured one day after plating (left column), and one and two days after treatment (middle and right panels, respectively). Scale bar = 200 μm (B) Quantification of axon degeneration, represented by the percentage of degenerating motor axons observed one day after vs. before treatment. Thrombin exerted a dose-dependent increase in the number of degenerating motor axons, which was significantly different from that in control, GDNF-treated, muscle-conditioned medium (MCM)-, Schwann cell-conditioned medium (SCCM)-, and astrocyte-conditioned medium (ACM)-treated explants (***, P<0.005, n = 3). Pre-incubation for 15 minutes of 200 nM thrombin with hirudin (at 500 μg/mL), SCCM, or ACM, but not MCM (not significant; ns), blocked the degenerative effects of thrombin on motor axons (***, P<0.005, **, P<0.001). Treatment with MCM alone produced a significantly higher number of degenerating axons vs. CTL or GDNF treatment (cross, P<0.005).
Fig 5.
PAR-1 mediates thrombin-induced developmental synaptic degeneration caused by Schwann cell ablation.
(A) E15.5 diaphragms from the indicated genotypes were dissected and immunostained with antibodies against synaptophysin (green). Note the retention of motor innervation of NMJs in prothrombin (FII), erbB3;FII double mutant (erbB3-/-; FII-/-; bottom right panel) vs. erbB3 single mutant diaphragm (erbB3-/-; bottom left panel). In contrast, there is no difference in the motor innervation between prothrombin wild-type (FII+/+) and mutant (FII-/-) diaphragm (top panels). (B) Muscle-specific elimination of prothrombin in erbB3 mutants (erbB3-/-; FIIFlox/Flox; Myf5-Cre; bottom panel) results in the rescue of presynaptically innervated NMJs. Scale bar in A-C = 100 μm. (C) PAR1; erbB3 double mutant diaphragm (erbB3-/-; PAR1-/-; bottom panel) also exhibits a rescue of motor innervation, whereas PAR1 single mutant diaphragm is similar to that of PAR1 wild-type (top panel). (D) Quantification of NMJs. The percentage of α-BTX-labeled AChRs apposed to synaptophysin-immunoreactive presynaptic terminals in diaphragm muscle is significantly higher in erbB3 mutants lacking FII, PAR1 or muscle-derived FII when compared to single erbB3 mutants alone (erbB3-/-; FIIFlox/Flox; Myf5-Cre vs. erbB3-/-; ***P<0.0005, n = 3 diaphragms for each genotype). (E) Lack of muscle-derived prothrombin/FII expression in FIIFlox/Flox; Myf5-Cre mice, 1,2 = muscle from FIIFlox/Flox; Myf5-Cre mice (+,- reverse transcriptase; RT); Lanes 3,4 = muscle from Myf5-Cre and FIIFlox/+; Myf5-Cre mice (+,- RT); Lanes 5,6 = muscle from FII-/- mice (+,- RT); Lane 7 = liver from wild-type mice (+RT). β-actin expression from same samples is shown below.
Fig 6.
Thrombin fails to cause motor axon degeneration in spinal explants derived from PAR1 mutant mice.
(A) Explants obtained from PAR1 wild-type; HB9:GFP (PAR1+/+, top panels) and PAR1 mutant; HB9:GFP mice (PAR1-/-; bottom panels) were plated with 10 nM GDNF and imaged the next day (left panels). Explants were then treated with 200 nM thrombin and imaged a second time the following day (right panels). In contrast to explants of PAR1 wild-type mice, which exhibit axonal degeneration in response to 200 nM thrombin, those of PAR1 mutant mice are largely unaffected. Scale bar = 100 μm. (B) The protective effect of PAR-1 deletion on thrombin-mediated axonal degeneration is dose-dependent, because higher concentrations of thrombin (400 nM) elicit motor axon pathology. CTL refers to GDNF treatment at plating and again at 1 day after plating. Each value reflects the percentage of healthy motor axons at 2 vs. 1 day after plating, and represents the mean of 3 samples. *P<0.0005, thrombin-treated vs. control; **P<0.0005, thrombin-treated PAR1-/- vs. thrombin-treated PAR1+/+explants; ***P< 0.0005, 200 nM vs. 400 nM thrombin-treated PAR1-/- explants. Student’s t with Bonferroni correction.
Fig 7.
Model for the interplay of neural activity and Schwann cells in maintaining the motor innervation of the NMJ.
(A) Summary of the presence or absence of Schwann cells, neuromuscular activity (+,–), or developmental synaptic degeneration in 19 mouse lines. erbB2-/- mice refer to erbB2 null mutants crossed to transgenic mice overexpressing erbB2 in the heart, which survive until birth[20]. (B) (Left panel) Presynaptic nerve terminals (green) release ACh onto muscle-derived AChRs, resulting in the release of muscle-derived thrombin (factor II; FII), whose activity is normally opposed by serpins C1 and D1 released from Schwann cells. (Right panel) In mice lacking Schwann cells, thrombin causes developmental synaptic degeneration (red X) because of the absence of these antithrombins.