Skip to main content
PMC home page
eLife logoLink to eLife
. 2018 Aug 16;7:e34625. doi: 10.7554/eLife.34625

Motoneuron Wnts regulate neuromuscular junction development

Chengyong Shen 1,†,✉, Lei Li 2,†, Kai Zhao 3,†, Lei Bai 1,†, Ailian Wang 1, Xiaoqiu Shu 1, Yatao Xiao 1, Jianmin Zhang 1, Kejing Zhang 1, Tiankun Hui 4, Wenbing Chen 2,4, Bin Zhang 5, Wei Hsu 6, Wen-Cheng Xiong 2,7, Lin Mei 2,7,✉
Editors: Jeremy Nathans8, Marianne E Bronner9
PMCID: PMC6128691  PMID: 30113308

Abstract

The neuromuscular junction (NMJ) is a synapse between motoneurons and skeletal muscles to control motor behavior. Unlike extensively investigated postsynaptic differentiation, less is known about mechanisms of presynaptic assembly. Genetic evidence of Wnt in mammalian NMJ development was missing due to the existence of multiple Wnts and their receptors. We show when Wnt secretion is abolished from motoneurons by mutating the Wnt ligand secretion mediator (Wls) gene, mutant mice showed muscle weakness and neurotransmission impairment. NMJs were unstable with reduced synaptic junctional folds and fragmented AChR clusters. Nerve terminals were swollen; synaptic vesicles were fewer and mislocated. The presynaptic deficits occurred earlier than postsynaptic deficits. Intriguingly, these phenotypes were not observed when deleting Wls in muscles or Schwann cells. We identified Wnt7A and Wnt7B as major Wnts for nerve terminal development in rescue experiments. These observations demonstrate a necessary role of motoneuron Wnts in NMJ development, in particular presynaptic differentiation.

Research organism: Mouse

Introduction

The neuromuscular junction (NMJ) is a peripheral synapse between motoneurons and skeletal muscle fibers that is critical to control muscle contractions. Dysfunction of this synapse has been implicated in various neurological disorders including congenital myasthenia syndrome (CMS), myasthenia gravis (MG), and amyotrophic lateral sclerosis (ALS) (Li et al., 2018; Sanes and Lichtman, 2001; Shen et al., 2015; Wu et al., 2010). During NMJ formation, acetylcholine receptors (AChRs) become concentrated at the postjunctional membrane whereas axon terminals differentiate to form active zones that align precisely to AChR clusters on junctional folds. NMJ development requires intimate interaction between motor nerve terminals and muscle fibers. For example, motoneurons produce agrin that binds its receptor LRP4 (low-density lipoprotein receptor-related protein 4) in muscle fibers and activates the muscle specific kinase MuSK (Bezakova and Ruegg, 2003; DeChiara et al., 1996; Kim et al., 2008; McMahan, 1990; Zhang et al., 2008).The agrin-LRP4-MuSK signaling is necessary for NMJ formation, in particular postsynaptic AChR clustering. In contrast to postsynaptic differentiation, much less is known about mechanisms of presynaptic assembly of the NMJ.

Recent evidence suggests that Wnt ligands may regulate synapse development in mammals. Regarding NMJ development, Wnts can induce AChR clusters by themselves and can promote agrin-induced AChR clustering in cultured C2C12 myotubes (Barik et al., 2014b; Zhang et al., 2012). These effects seem to require the extracellular cysteine-rich domain (CRD) of MuSK, which is homologous to that in the Wnt receptor Frizzled (Hubbard and Gnanasambandan, 2013). Deleting the CRD was shown to impair in vivo NMJ formation (Messéant et al., 2015; Strochlic et al., 2012) although this notion was challenged by another study (Remédio et al., 2016). Intracellularly, MuSK interacts with disheveled (Dvl1), the key scaffold protein in the Wnt signal pathway (Luo et al., 2002). In Zebrafish, Wnt binding to the CRD of MuSK and MuSK interaction with Dvl are necessary for guiding axons to the middle region of muscle fibers (Jing et al., 2009). Mutant mice lacking or overexpressing β-catenin, a key mediator of the Wnt canonical pathway, in muscle fibers display both pre- and postsynaptic deficits (Li et al., 2008; Liu et al., 2012; Wu et al., 2012a). In terms of synapses in the brain, Dvl1 expresses higher in axons than dendrites, especially in growth cones of cultured hippocampus neurons, suggesting a role of Wnt pathway in presynaptic differentiation (Zhang et al., 2007). Cerebella granule cells secret Wnt7A to induce axon and growth cone remodeling, including Synapsin 1 puncta clustering (Hall et al., 2000). Inducing the expression of Dickkopf1, a secreted Wnt antagonist, decreases the number of cortico-striatal synapses in adult mice. The remaining excitatory terminals contain fewer synaptic vesicles (Galli et al., 2014).

There are 19 different Wnts in rodents and humans (Barik et al., 2014b), many of which have redundant functions, making it difficult to reveal convincingly the function of Wnt ligands in vivo. In this study, we investigated the role of Wnt ligands in NMJ formation by taking the advantage of Wnt ligand secretion mediator (Wls). Wls is a seven-transmembrane protein that is necessary for sorting and transporting Wnts from the endoplasmic reticulum to the cell surface (Hausmann et al., 2007). In Drosophila Wls mutants, Wingless (the homolog of vertebrate Wnt1) is retained intracellularly and cannot be secreted (Bänziger et al., 2006; Bartscherer et al., 2006; Goodman et al., 2006). Wls null mouse embryos exhibit defects in establishment of the body axis (Fu et al., 2009). We generated mutant mice that lacked Wls specifically in motorneurons, skeletal muscles, or Schwann cells and characterized NMJ development by using a combination of morphological and functional approaches. We have also identified the Wnt ligands whose secretion requires Wls and determined whether the NMJ deficits of HB9-Wls-/- mice could be ameliorated by supplying candidate Wnts. Our results support a model where motoneurons secret Wnts such as Wnt7A and Wnt7B for nerve terminal differentiation, suggesting dysregulation of Wnt signaling as a pathological mechanism of neuromuscular disorders.

Results

Unique Wnt expression patterns in motoneurons, Schwann cells and muscles

The NMJ is a tripartite synapse consisting of muscle fibers, Schwann cells, and motoneuron terminals (Darabid et al., 2014; Wu et al., 2010). We determined Wnt expression patterns in these cells. To isolate motoneurons, spinal cords were dissected out from P0 HB9-tdTomato mice where motoneurons were labeled by tdTomato (Figure 1A). Dissociated motoneurons were subjected to FACS analysis and RNA extraction. Real-time PCR analysis identified many Wnts in motoneurons. Noticeably, Wnt7A appeared to be the most abundant, followed by Wnt3, Wnt4, and Wnt7B. Expression of other Wnts appeared to be low in motoneurons. Dominant Wnts in Schwann cells appeared to be Wnt5A and Wnt9A whereas Wnt4, Wnt6 and Wnt9A were the most abundant in muscle cells (Figure 1B and C).

Figure 1. Wnt expression in motoneurons, Schwann cells, and muscle cells.

(A) Wnt expression in motoneurons. Motoneurons in HB9-tdTomato mice (P0) were isolated by FACS. mRNAs were subjected to real-time PCR with GAPDH as internal control. (B) Wnt expression in primary Schwann cells. Sciatic nerves were isolated from P3 mice for primary Schwann cell culture. mRNAs were subjected to real-time PCR with GAPDH as internal control. (C) Wnt expression in C2C12 muscle cells. mRNAs were isolated from C2C12 myotubes and subjected to real-time PCR with GAPDH as internal control.

Figure 1.

Figure 1—figure supplement 1. Frizzled expression in spinal ventral horn, Schwann cells, and muscle cells.

Figure 1—figure supplement 1.

(A) Expression of Frizzled mRNA in spinal ventral cord. ChAT-EGFP mice were dissected to separate the spinal ventral horn. mRNA was subjected for real-time PCR analysis of Frizzled expression. GAPDH was set as internal control. (B) Expression of Frizzled mRNA in primary Schwann cells. Sciatic nerve in wide type mice (P3) was isolated and disassociated for primary Schwann cells culture, and their mRNA were subjected for real-time PCR analysis. GAPDH was set as internal control. (C) Expression of Frizzled mRNA in muscle cells. C2C12 myoblast was fused into myotubes and their mRNA was subjected for real-time PCR analysis. GAPDH was set as internal control.
Open in a new tab

No apparent NMJ deficits by Wls mutation in Schwann cells or muscles

We determined whether mutation of Wls in muscles, Schwann cells, or motoneurons alters the NMJ formation. HSA::Cre mice express Cre specifically in myotomal regions of somites beginning at E9.5, and Cre protein is detectable in almost all skeletal muscle fibers in P0 mice (Figure 2—figure supplement 1A) (Escher et al., 2005; Li et al., 2008). Cre in Wnt1::Cre mice is expressed in neural crest cells and their derivatives including Schwann cells (Figure 2—figure supplement 1E) (Fu et al., 2011). We crossed Wlsf/f mice with HSA::Cre and Wnt1::Cre mice, respectively. Resulting HSA-Wls-/- mice and Wnt1-Wls-/- mice were produced at expected ratios following the Mendel rule with comparable body size to control (HSA::Cre or Wlsf/f) mice. Western blot showed that, most, if not all of Wls proteins were lost in skeletal muscle homogenates in HSA-Wls-/- mice (Figure 2—figure supplement 1B). To characterize NMJs, muscles were stained whole-mount with antibodies against neurofilament (anti-NF) and synaptophysin (anti-Syn) to label axons and nerve terminal synaptic vesicles, and rhodamine-conjugated α-bungarotoxin (R-BTX) to visualize postsynaptic AChR clusters. HSA-Wls-/- mice appeared to be normal at birth without apparent deficits in AChR clusters (number and location), motor nerve terminals (arborization) and innervation of AChR clusters (Figure 2—figure supplement 1C and D). These results were in agreement with a previous report (Remédio et al., 2016) and indicate that muscle Wls is dispensable for NMJ formation.

Western blotting of sciatic nerve homogenates revealed a reduction of Wls protein in Wnt1-Wls-/- mice (Figure 2—figure supplement 1F). The residual level of Wls protein was likely to be from nerve axons. Wnt1-Wls-/- mice died soon after birth with craniofacial skeleton defects, in accord with the notion that Wnt signaling in neural crests is critical for craniofacial skeleton development (Fu et al., 2011). However, NMJs in Wnt1-Wls-/- mice were similar to those of control mice in terms of number, location, innervation of AChR clusters and motor nerve arborization (Figure 2—figure supplement 1G and H). These results suggest that Wls in Schwann cells is not necessary for NMJ formation.

Reduced body weight and muscle strength in Wls motoneuron knockout mice

To determine whether Wnts from motoneurons are involved in NMJ formation, we crossed HB9::Cre mice with Wlsf/f mice. HB9 is a motoneuron-specific transcription factor that begins to express at E9.5 and is critical for motoneuron differentiation (Arber et al., 1999; Thaler et al., 1999). Cre expression in HB9::Cre mice was verified by crossing with Rosa-tdTomato reporter mice. As shown in Figure 2A, the Tomato fluorescence protein was specifically restricted in motor nerves that were located in the middle of diaphragm in HB9-tdTomato mice, but not in control mice (Figure 2A). Western blotting of sciatic nerve homogenates revealed a reduction of Wls protein (Figure 2B). The residual Wls was likely to due to Wls from sensory axons and Schwann cells. To determine that Wls was truly ablated from motoneurons, we stained spinal cord sections with anti-Wls antibody. As shown in Figure 2C, Wls was readily detectable in ChAT positive motoneurons in ventral horns of the spinal cord of control mice, but not those of HB9-Wls-/- mice. These results indicate that Wls was ablated from motoneurons in HB9-Wls-/- mice.

Figure 2. Ablation of Wls expression in motoneuron causes muscle weakness.

(A) Up: NMJ structure. Cre targets motoneuron (HB9::cre). Down: Cre expression was verified by crossing HB9::cre mice with Rosa-tdTomato reporter mice (HB9-tdTomato). HB9::cre drives tdTomato expression in the middle line of diaphragm where motoneuron innervates in the muscle (HB9-tdTamato, (P7). MN, motoneuron; SC, Schwann cells; SM, skeletal muscle. (B) Immunoblot of Wls in sciatic nerve from 2-month-old control and HB9-Wls-/- mutant. Notice that the reduced Wls protein expression in the mutant. (C) Immunostaining of Wls in spinal cord from 2-month-old control and HB9-Wls-/- mutant. Notice that Wls staining signal is lost in the ChAT positive cells in the spinal ventral horn in the HB9-Wls-/-. Anti-Wls (Red); Anti-ChAT (Green). Arrows indicate colocalization of Wls and ChAT protein. (D) Two genotypes at P0 and P7. Notice that apparent normal size at P0, and reduced body size at P7. (E) Body weight of two genotypes at P7. Control, n = 20; mutant, n = 11; Unpaired t-test, **p<0.01. (F) Grip strength of two genotypes at P60. Control, n = 5; mutant, n = 5. Unpaired t-test, **p<0.01. (G) Grip strength of two genotypes with same body weight at P60. Control, n = 5; mutant, n = 5. Unpaired t-test, **p<0.01.

Figure 2.

Figure 2—figure supplement 1. Ablation of Wls in skeletal muscle or Schwann cell has no apparent NMJ morphological defects.

Figure 2—figure supplement 1.

(A) Up: NMJ structure. Cre targets skeletal muscle (HSA::cre). Down: Cre expression was verified by crossing HSA::cre mice with Rosa-tdTomato reporter mice (HSA-tdTomato, (P7). HSA::cre drives tdTomato protein expression in each muscle fiber of diaphragm. (B) Immunoblot of Wls protein in gastrocnemius of P5 control and HSA-Wls-/- mice. (C) Immunostaining of NMJs in diaphragms of P5 control and HSA-Wls-/- mice with anti-NF and anti-Syn (Green). R-BTX stains postsynaptic AChR (Red). Control: Cre or Wlsf/f littermates. (D) Statistical results of endplate width (Left), AChR number (Right) in B. n = 5 mice per group; Unpaired t-test. p>0.05. (E) Up: NMJ structure. Cre targets Schwann cells (Wnt1::cre). Down: Cre expression was verified by crossing Wnt1::cre mice with Rosa-Laz reporter mice (Wnt1-Laz, (P7). Wnt1::cre drives laz expression in middle line of diaphragm where Schwann cells surround the innervated motor axons (Wnt1-Laz). (F) Immunoblot of Wls protein in sciatic nerve of E18.5 control and Wnt1-Wls-/- mice. (G) Immunostaining of NMJs in diaphragms of E18.5 control and Wnt1-Wls-/-with anti-NF and anti-Syn (Green), R-BTX (Red). Control: Cre or Wlsf/f littermates. (H) The statistics of endplate width (Left), AChR number (Right) in E. n = 5 mice per group; Unpaired t-test. p>0.05.
Figure 2—figure supplement 2. NMJ formation appears grossly normal in HB9-Wls-/- mice.

Figure 2—figure supplement 2.

(A) Immunostaining of NMJs in diaphragm of E18.5 control and HB9-Wls-/- embryos with anti-NF and anti-Syn to label motoneuron axons and synaptic vesicles (Green). R-BTX stains postsynaptic AChR (Red). (B) Statistical results of endplate band width of two genotypes. n = 3 mice per group; Unpaired t-test, p>0.05. (C) Statistical results of AChR cluster size. n = 3 mice per group; Unpaired t-test, p>0.05. (D) Statistical results of AChR number. n = 3 mice per group; Unpaired t-test, p>0.05.
Figure 2—figure supplement 3. Comparable NMJ formation between HSA-Wls-/- and HB9/HSA-Wls-/-, Wnt1-Wls-/- and HB9/Wnt1-Wls-/- mice.

Figure 2—figure supplement 3.

(A) Immunostaining of NMJs in diaphragms of HSA-Wls-/- and HB9/HSA-Wls-/- E18.5 embryos with anti-NF/Syn (Green). R-BTX stains postsynaptic AChR (Red). Scale bar, 100 μm. (B) The statistics of endplate width and AChR cluster number in A. n = 3 mice per group; Unpaired t-test. p>0.05. (C) Immunostaining of NMJs in diaphragms of Wnt1-Wls-/- and HB9/Wnt1-Wls-/- E18.5 embryos with anti-NF/Syn (Green). R-BTX stains postsynaptic AChR (Red). Scale bar, 100 μm. (D) The statistics of endplate width and AChR cluster number in C. n = 3 mice per group; Unpaired t-test. p>0.05.
Open in a new tab

HB9-Wls-/- pups were born at expected Mendelian ratio and were similar in body size at birth, compared to control pups. However, as mice developed, HB9-Wls-/- pups showed growth retardation with smaller body sizes (100.0 ± 2.2% in control vs 68.6 ± 9.4% in HB9-Wls-/-, p<0.01, Figure 2D and E). One third of HB9-Wls-/- mice died prematurely before adulthood. At 2 months old, the body weight of surviving HB9-Wls-/- mice was 31% less than control. In particular, grip strength was reduced by 69%, compared with control (113.4 ± 5.8 g in control mice to 35.2 ± 7.9 g in HB9-Wls-/- mice; p<0.01) (Figure 2F). Because body weights were reduced in HB9-Wls-/- mice, we also compared grip strength between two groups of mice with similar body weights (29.3 ± 1.1 g vs 27.7 ± 1.6 g for control and HB9-Wls-/-, respectively, p>0.05). As shown in Figure 2G, HB9-Wls-/- mice had reduced grip strength (129 ± 9.7 g in control mice; 71.8 ± 7.9 g in HB9-Wls-/- mice, p<0.01). These results indicate that motoneuron Wls is required for development of body weight and muscle strength.

Age-dependent nerve terminal swellings and poor innervation

To determine whether loss of Wls in motoneurons alters the NMJ formation, we first characterized NMJs of mouse embryos (E18.5). As shown in suppl Figure 2A, the primary branches of phrenic nerves are localized in central regions of muscle fibers; the numbers of secondary and tertiary branches were similar between control and HB9-Wls-/- mice. The morphology of nerve terminals was comparable between the two genotypes and each of them was associated with an AChR cluster (Figure 2—figure supplement 2A). These results indicate that nerve terminal pathfinding to the middle region of muscle fibers and defasciculation or arborization appears grossly normal. Postsynaptically, AChR clusters appeared as opaque ovals and were enriched in the middle region of muscle fibers, outlining a central band in control mice (Wu et al., 2010) (Figure 2—figure supplement 2A). The width of the central region that occupied by AChR clusters was similar in control and in HB9-Wls-/- diaphragms (172 ± 25.4 µm vs 198 ± 25.5 µm, p>0.05). The density and size of AChR clusters in HB9-Wls-/- mice were also similar to those in control mice (AChR size: 100 ± 5.0% vs 87.9 ± 4.3%, p>0.05; AChR density: 537 ± 28.0 vs 556 ± 22.2 per µm2, p>0.05) (Figure 2—figure supplement 2B–2D). These results suggest that HB9-Wls-/- mice were able to form the NMJ in the absence of motoneuron Wls, in agreement with the finding of normal body weights of P0 HB9-Wls-/- mice.

Normal NMJs at P0 pups were unable to explain muscle weakness at age of 2 months. Therefore, we isolated gastrocnemius from 2-month-old mice and analyzed NMJ morphology by whole-mount staining. To avoid potential bias during imaging process, NMJs were viewed from three different axes and only largest images of AChR clusters were captured for analysis (Shen et al., 2013). Mature NMJs had well developed into characteristic pretzel-like structures with complex continuous arrays that were innervated by motor nerve terminals (Shen et al., 2013) (Figure 3A). The arrays of HB9-Wls-/- mice appeared to be as complex as control mice, but less continuous (Figure 3A and B). There were small pieces of AChR clusters (arrow, Figure 3A and B), and the number of AChR cluster fragments were increased to 3.0 ± 0.3 fragments per NMJ in HB9-Wls-/- mice from 1.4 ± 0.2 fragments per NMJ in control mice (p<0.01) (Figure 3A–3C).

Figure 3. Abnormal pre- and postsynaptic NMJ structure in HB9-Wls-/- mutant mice.

(A) Immunostaining of NMJs in gastrocnemius of 2-month-old control and HB9-Wls-/- mice with anti-NF/anti-Syn (Green) and R-BTX (Red). Nuclei were labeled with DAPI (Blue). In mutant, some nerve terminals are swollen (white arrowhead) and AChR clusters are broken (white arrow). (B) The high magnification of boxed area in A. White arrowhead indicates swollen nerve terminals. Blue arrowhead indicates motor nerve detaching from AChR clusters; White arrow indicates broken pieces of AChR. (C) Statistical results of AChR fragmentation per NMJ. n = 3 mice per group; Unpaired t-test, **p<0.01. (D) Percentage of AChR cluster innervated with nerve terminals. n = 3 mice per group; Unpaired t-test, *p<0.05. (E) Percentage of nerve terminal swellings. n = 3 mice per group; Unpaired t-test, ***p<0.001. (F) The diaphragm of 2-month-old HB9-tdTomato and HB9-WLS-/- tdTomato mice. Nerve terminal swellings are indicated with white arrowhead. Normal nerve terminals are indicated with white arrow. Axon branch are indicated with blue arrow. Boxes indicated by dash line were enlarged in the right.

Figure 3.

Figure 3—figure supplement 1. Wls loss in motoneuron does not enhance synapse number in single muscle fiber.

Figure 3—figure supplement 1.

(A) Representative images of individual gastrocnemius muscle fibers of 2-month-old control and HB9-Wls-/- mice. To visualize AChR clusters on single muscle fibers, gastrocnemius was fixed in 4% PFA and stained with R-BTX (Red) and DAPI (Blue) to show AChR clusters and myonuclei. Muscles were washed three times in PBS and teased into single fibers and mounted in Vectashield mounting medium. We counted the number of AChR cluster in each single muscle fiber and defined that synapse elimination was impaired when more than one AChR cluster was found after P15. (B) Quantitative analysis of NMJ number per muscle fiber. NMJ number per muscle fiber was comparable between control and HB9-Wls-/- mice at P0, P7, P15, and P60. n = 3 mice per group. Unpaired t-test, p>0.05.
Open in a new tab

A characteristic abnormality of HB9-Wls-/- mice was the appearance of abnormal swellings of nerve terminals. The swellings appeared to be at axon terminals (arrowhead, Figure 3A and B) although we could not exclude the possibility that the swellings may bulge from axons en passant to form a different NMJ (arrowhead, Figure 3A). Axonal bulbs or shedding were observed during synapse elimination (Tapia et al., 2012). We did not find the excessive NMJ number in each single muscle fiber at different ages of HB9-Wls-/- mice, suggesting the normal synapse elimination in HB9-Wls-/- mice (Figure 3—figure supplement 1). Unlike axonal shedding which was observed in a small number of NMJs, axonal swellings in HB9-Wls-/- mice were observed in 53.5 ± 3.2% of nerve terminals (Figure 3E). Some swellings were co-stained with dispersed AChR signal, suggesting a problem of mutant axons in inducing AChR clusters. In addition to axonal swellings, nerve terminals in HB9-Wls-/- mice became detached from AChR arrays, causing misalignment between nerve terminals and AChR clusters (blue arrowhead, Figure 3B). Consequently, the area of AChR arrays that were covered by nerve terminals was reduced in HB9-Wls-/- mice (Figure 3D) (73.6 ± 7.1% in HB9-Wls-/- vs 100 ± 3.7% in control, p<0.05).

To further characterize axonal swellings, we utilized Rosa-tdTomato reporter mice where tdTomato expression is controlled by a loxP-stop-loxP cassette. When crossed with HB9-Wls-/- mice, HB9::Cre thus drove the expression of tdTomato and at the same time mutated the Wls gene in motoneurons. In diaphragms of control mice, tdTomato-labeled axons were transparent in secondary and tertiary branches, and each terminal ended with a pretzel-like enlargement. However, in HB9-Wls-/- mice, in addition to these transparent axon branches and terminals, many branches appeared to contain aggregates. Axon terminals in HB9-Wls-/- mice appeared as individual swellings without characteristic pretzel-like structure (Figure 3F). Quantitatively, 53% of axon terminals lost the pretzel-like structure (Figure 3F), in agreement with the result from immunostaining (Figure 3A and B). Together, these morphological results demonstrate that motoneuron-specific Wls mutant mice were able to form NMJs, but displayed axonal deficits at adult, indicating an age-dependent requirement of Wls in NMJ maintenance.

Reduced synaptic vesicles, junctional folds and increased synaptic cleft width

To further characterize NMJ defects, we performed EM analysis. In control NMJs, nerve terminals are filled with synaptic vesicles and covered with processes of terminal Schwann cells. Membranes of muscle fibers opposing axon terminals invaginate to form junctional folds where AChRs are enriched (Figure 4A). These unique structures ensure the neurotransmitter ACh, released from axonal terminals, to be efficiently captured by AChRs on the crest of junctional folds (Li et al., 2018; Sanes and Lichtman, 2001; Wu et al., 2010). However, these NMJ structures were largely distorted in HB9-Wls-/- mice, which made it difficult to identify characteristic NMJs. Noticeably, the number of synaptic vesicles was reduced in nerve terminals (107 ± 14.6 per µm2 in control vs 25.6 ± 3.0 in HB9-Wls-/-, p<0.01, black arrowhead) (Figure 4A and B). The number of junctional folds was reduced (1.2 ± 0.3 in HB9-Wls-/- vs 3.0 ± 0.5 in control per µm, p<0.05, black arrow) (Figure 4). Synaptic cleft was uneven in HB9-Wls-/- mice and appeared to be larger than that of control mice (2.1 ± 0.3 folds in HB9-Wls-/- mice, compared with that in control mice, p<0.01) (Figure 4A and D). In control NMJs, basal lamina was visible in synaptic cleft and between junctional folds, with highest electron density in the middle. However, in HB9-Wls-/- mice, basal lamina density was reduced without a high-density midline (empty arrowheads, Figure 4A). These EM data are in accord with those from light microscopic studies (Figure 3) and demonstrate that Wls in motoneurons is critical for the integrity of NMJ structure.

Figure 4. Disruption of NMJ ultrastructure in HB9-Wls-/- mutant mice.

Figure 4.

Open in a new tab

(A) Representative EM images of diaphragm NMJs in P15 control (a, a’, a’’) and HB9-Wls-/- mice (b, b’, b’’). Notice that there is a dramatic reduction of synaptic vesicles number in the HB9-Wls-/- (black arrowhead, (a’’ and b’’) and loss of synaptic junctional folds (black arrow, (a’’ and b’’) and basal lamina (empty arrowhead, (a’’ and b’’) in HB9-Wls-/-. Boxed regions are shown at higher magnification immediately below. NT, nerve terminal; MF, muscle fiber; SC, Schwann cell; SV, synaptic vesicle; JF, junctional fold. Scale bars: 2.0 μm (top); 0.5 μm (middle); 0.1 μm (bottom). (B) Synaptic vesicle density in two genotypes. n = 3 mice per group; Unpaired t-test, **p<0.01. (C) Number of synaptic junctional folds. n = 3 mice per group; Unpaired t-test, *p<0.05. (D) Synaptic craft width. n = 3 mice per group; Unpaired t-test, **p<0.01.

Impaired neuromuscular transmission in HB9-Wls-/- mice

To investigate pathophysiological mechanisms of muscle weakness in HB9-Wls-/- mice, we measured compound muscle action potentials (CMAPs) to determine if there is impaired neuromuscular transmission. Gastrocnemius was recorded in response to repetitive nerve stimuli (Shen et al., 2013). In control mice, CMAPs showed little change after 10 consecutive nerve stimuli at frequencies from 2 Hz to 40 Hz (Figure 5A and B). While in HB9-Wls-/- mice, CMAPs recorded at 40 Hz began to decrease at the second stimulus and significantly decreased from the fourth; the decrement of CMAP at the tenth stimulus was about 10% (Figure 5B). The reduction of CMAPs in HB9-Wls-/- mice was frequency-dependent (Figure 5C), which indicates progressive loss of successful neuromuscular transmission after repeated stimulation.

Figure 5. Wls loss in motoneuron impairs neurotransmission initially from presynapse.

(A) CMAPs were recorded in gastrocnemius from P30 mice in response to a train of 10 submaximal stimuli at different frequencies. The first stimulus response in control mice was designated as 100%. Representative 10 CMAP traces, shown stacked in succession for better comparison. (B) Reduced CMAP amplitudes at 40 Hz in HB9-Wls-/- mice. n = 4 mice per group. Two-way ANOVA, *p<0.05. (C) CMAP amplitudes of the tenth stimulation at different stimulation frequencies. n = 4 mice per group. Two-way ANOVA, *p<0.05. (D) Representative mEPP traces from P30 mice. mEPPs were recorded from hemidiaphragms. Traces underlined on the left are enlarged on the right. (E) Reduced mEPP amplitudes in mutant mice and CCh effects at P30. F(2, 9)=13.56, ns, p>0.05, **p<0.01. n = 4 mice per group, 5–6 muscle fibers per mouse; One-way ANOVA. (F) Reduced mEPP frequencies in mutant mice and CCh effect at P30. F(2, 9)=14.2, ns, p>0.05, **p<0.01, *p<0.05, n = 4 mice per group, 5–6 muscle fibers per mouse; One-way ANOVA. (G) Representative mEPP traces from P1 mice. (H) Comparable mEPP amplitude between two genotypes and CCh effect at P1. F(2, 9)=63.07, ns, p>0.05, ***p<0.001. n = 4 mice per group, 5–6 muscle fibers per mouse; One-way ANOVA. (I) Reduced mEPP frequency in mutant mice and CCh effect at P1. F(2, 9)=12.87, ns, p>0.05, **p<0.01. n = 4 mice per group, 5–6 muscle fibers per mouse; One-way ANOVA.

Figure 5.

Figure 5—figure supplement 1. Impairment of pre- and postsynaptic prosperity in P15 HB9-Wls-/- mice.

Figure 5—figure supplement 1.

(A) Representative mEPP traces from P15 mice. mEPPs were recorded from hemidiaphragms. Traces underlined on the left are enlarged on the right. (B) Reduced mEPP amplitudes in diaphragms of HB9-Wls-/- mice at P15. n = 4 mice per group, 5–6 muscle fibers per mouse; Unpaired t-test, ***p<0.001. (C) Reduced mEPP frequencies in in diaphragms of HB9-Wls-/- mice at P15. ***p<0.001. n = 4 mice per group, 5–6 muscle fibers per mouse; Unpaired t-test, ***p<0.001. (D) Cumulative plots of mEPP events against amplitudes. (E) Cumulative plots of mEPP events intervals.
Open in a new tab

To further investigate whether the neurotransmission deficits result from pre- and/or postsynaptic impairment, we recorded miniature endplate potentials (mEPPs), events generated by spontaneous vesicle release, in the single muscle fiber of hemidiaphragms of P30 HB9-Wls-/- mice. As shown in Figure 5D and E, mEPP amplitudes were reduced in HB9-Wls-/- mice, compared with controls (1.20 ± 0.08 mV in control vs 0.76 ± 0.06 mV in HB9-Wls-/-; p<0.01), indicative of reduced AChR density. mEPP frequencies were also decreased by ~40% in HB9-Wls-/- diaphragms, compared with control (0.78 ± 0.03 Hz in HB9-Wls-/- vs 1.20 ± 0.08 Hz in control, p<0.01, Figure 5D and F). This result demonstrates deficits in spontaneous ACh release from motor nerve terminals, consistent with morphological presynaptic defects in mice lacking Wls in motoneurons. We observed similar electrophysiological defects at P15 HB9-Wls-/- mice (Figure 5—figure supplement 1). The result was consistent with that showed by EM, in which postsynaptic as well as presynaptic membranes were disorganized (Figure 4).

Because Wls was specifically mutated in motoneurons, it was interesting that HB9-Wls-/- mice at P30 and P15 displayed postsynaptic deficits. Such deficits may result as a consequence of deteriorated presynaptic terminals. This hypothesis would predict that mutant mice at earlier age display only presynaptic, but not postsynaptic deficits. To test this hypothesis, mEPPs were measured in diaphragm at neonatal pups (P1) when the NMJ morphology appeared normal. mEPP frequency in HB9-Wls-/- NMJs was reduced by 42% (0.042 ± 0.003 Hz in control vs 0.022 ± 0.002 Hz in mutant, p<0.01) (Figure 5G and I), indicative of presynaptic deficits. In contrast, there was no significant difference in mEPP amplitudes between control and HB9-Wls-/- mice (1.98 ± 0.10 mV in control vs 1.88 ± 0.09 mV in HB9-Wls-/-, p>0.05) (Figure 5G and H). These functional results demonstrate that presynaptic deficits occurred in advance of postsynaptic deficits, agreeing the notion that EM and electrophysiological analysis are more sensitive in revealing NMJ deficits than light microscope analysis.

To further test this model, we studied effects of carbochol (CCh), an ACh agonist, on neurotransmission in HB9-Wls-/- mice. CCh (10 µM) increased both mEPP frequency and amplitude at P1 and P30 mice (Figure 5D–5I), in agreement with previous reports (Katz and Miledi, 1972; Miyamoto and Volle, 1974). Interestingly, CCh also elevated mEPP frequency and amplitude in Wls mutant diaphragms to similar or higher levels of controls in the absence of CCh (Figure 5D–5I). Nevertheless, mEPP frequency and amplitude in the presence of CCh were lower in mutant, compared with control at age of P30. These results indicate that transmission impairment in mutant mice could be attenuated by ACh agonist.

SV2-positive aggregates in HB9-Wls-/- axons

To determine molecular components of axon swellings, muscles were stained with antibody against NF. As shown in Figure 6A, the swellings were positive for NF, indicating that they may contain neurofilaments. They could also be stained with antibody against S100, a marker of Schwann cells, suggesting that the swellings were covered by Schwann cells (Figure 6A).

Figure 6. Clustered synaptic vesicles in HB9-Wls-/- axons.

Figure 6.

Open in a new tab

(A) Immunostaining of axonal terminal swelling in gastrocnemius of P21 HB9-Wls-/- mice with anti-NF (Green), anti-S100 (Purple), R-BTX (Red). (B) Immunostaining of NMJs in gastrocnemius of 2-month-old control and mutant mice with anti-NF or anti-SV2 to label synaptic vesicles (Green). R-BTX indicates the AChR (Red). Cell nucleus is labeled with DAPI (Blue). Dashline indicates motor axon. Notice that in the mutant NMJ, there are aberrant deposits of synaptic vesicles in the motor axon (bold white arrow in mutant and empty arrowhead in control), and reduced overlap between pre and post-synapse (white arrowhead). Nucleuses around motor axons are from Schwann cells. (C) Immunoblot of SV2 and Synapsin 1 in phrenic nerve at 2-month-old control and mutant mice.

Both EM and electrophysiological analyses revealed presynaptic deficits. In particular, vesicle release was reduced and there were fewer synaptic vesicles in HB9-Wls-/- mice, compared with control mice (Figures 4 and 5). We posited that it may result from a problem in vesicle transportation to nerve terminals. To test this hypothesis, gastrocnemius of 2-month-old mice was immunostained with antibody against SV2, a marker of synaptic vesicles. In control NMJs, SV2-positive vesicles were enriched at NMJs and showed puncta-like structures, which were distributed evenly in arrays of AChR clusters (Figure 6B). In HB9-Wls-/- NMJs, SV2 staining signal was spotty and discontinuous, faint or missing in many areas where AChR staining was strong. Noticeably, SV2-positive aggregates were seen in motoneuron axons in HB9-Wls-/- mice (white arrow, Figure 6B). Such SV2 aggregates were not observed in control mice. These results suggest abnormal accumulation of synaptic vesicles in axons. To test this further, we isolated phrenic nerves which contains purer motoneuron axons than sciatic nerve and performed immunoblotting to check the protein levels of synaptic vesicles. As shown in Figure 6C, compared with control, SV2 and Synapsin 1 protein level in phrenic nerve was increased in HB9-Wls-/- mutant mice. Together, these results identify a cellular mechanism for reduced synaptic vesicles at NMJs in mice lacking Wls in motoneurons.

Motoneuron Wnts for nerve terminal development

Wls is critical for sorting and transporting Wnts from the endoplasmic reticulum to the cell surface (Yang et al., 2008). NMJ deficits in mutant mice lacking Wls in motoneurons suggest a role of Wnts from motoneurons for NMJ development. Motoneurons express many Wnts as shown in Figure 1A. We attempted to identify which Wnts were vulnerable to Wls deficiency. We co-transfected Wls-shRNA that reduced Wls expression (data not shown) and individual Wnt constructs in HEK293 cells and analyzed Wnts level in condition medium. Each Wnt ligand was tagged with a Flag epitope. We focused on Wnts that were relatively abundant in motoneurons. As shown in Figure 7A and B, Wls knockdown reduced most of Wnts level including Wnt5A, Wnt7A, Wnt7B, Wnt9A, and Wnt9B in the medium, in agreement with previous reports.

Figure 7. Wnt7A partially rescues axonal terminal swelling in HB9-Wls-/- mice.

Figure 7.

Open in a new tab

(A) Wls-dependent Wnts secretion. HEK293T cells were co-transfected with shRNA of Wls and motoneuron Wnts cDNA with a Flag tag. Conditioned medium was collected for testing secreted Wnt. Ponceau S red indicates the loading. (B) Statistical results of secreted Wnt protein levels in conditioned medium in A. *p<0.05, **p<0.01, ***p<0.001. Unpaired t-test. Three independent experiments were performed. (C) Immunostaining of mutant TA muscles injected with Wnt recombinant proteins. Wnt recombinant proteins (10 ng/µl, 20 µl) were injected into TA muscles of right leg in HB9-Wls-/- mice at every three days. TA muscle of left leg in the same mouse was injected with vesicle as control. Thirty days later, TA muscle was isolated for NMJ staining. Compared with control, Wnt5A or Wnt9B, Wnt7A and Wnt7B partially rescued the axon terminal swelling in HB9-Wls-/- mice. Anti-NF/Syn: Green; R-BTX: Red; DAPI: Blue. (D) Statistical results of F. Control group: n = 5 mice; HB9-Wls-/- group: n = 6 mice; HB9-Wls-/- + Wnt5A group: n = 6 mice; HB9-Wls-/- + Wnt7A group: n = 5 mice. HB9-Wls-/- + Wnt7B group: n = 5 mice. HB9-Wls-/- + Wnt9B group: n = 5 mice. Unpaired t-test, *p<0.05, **p<0.01. ns, non-significant. (E) Increased CMAP amplitudes in HB9-Wls-/- mice by Wnt7A at 30 Hz stimulation. Control group: Black; HB9-Wls-/- group: Red; Control +Wnt7A group: Gray; HB9-Wls-/- + Wnt7A group: Blue. n = 4 mice per group; Two-way ANOVA, *p<0.05.

The observations that Wnt5A, Wnt7A, Wnt7B, and Wnt9B were expressed in the spinal cord (Figure 1A) and their levels in the medium were largely reduced by Wls knockdown (Figure 7A and B) suggest that these Wnts may be utilized by motoneurons to regulate the NMJ formation. To test this hypothesis, we determined whether Wnt5A, Wnt7A, Wnt7B, and Wnt9B recombinant proteins were able to diminish NMJ deficits in HB9-Wls-/- mice. TA muscles of mutant pups were injected with respective Wnts or vehicle (as control) 10 times, once every three days for a period of 30 days. Remarkably, muscles injected with Wnt7A and Wnt7B displayed fewer axon swellings, compared with muscles injected with vehicle (Figure 7C and D). This effect appeared to be specific because muscles injected Wnt5A or Wnt9B displayed similar number of axon swellings (NMJs with swellings: 22.7 ± 3.0% in HB9-Wls-/-; 17.9 ± 3.9% in HB9-Wls-/- + Wnt5A; 10.1 ± 2.7% in HB9-Wls-/- + Wnt7A; 12.2 ± 1.5 in HB9-Wls-/- + Wnt7B; 19.4 ± 2.4% in HB9-Wls-/- + Wnt9B). To further test this, we determined whether neuromuscular transmission impairment in HB9-Wls-/- mice could be attenuated by Wnt7A administration. As shown in Figure 7E, CMAPs of HB9-Wls-/- mice were resuced by Wnt7A administration. Together, these results support a model where motoneurons release Wnts to regulate NMJ development.

Discussion

The roles of Wnt in NMJ presynaptic differentiation remain elusive. Here we found that blocking Wnt secretion from motoneurons causes NMJ degeneration evidenced by muscle weakness, impaired neurotransmission, distorted NMJ structure, motor axonal terminal swelling, and synaptic vesicle reduction. Rescue experiments identified Wnt7A and Wnt7B as key Wnts secreted from motoneurons for NMJ development. Our results provide a novel role of Wnt ligand secretion from motoneurons in NMJ development and maintenance.

NMJ deficits were observed in mice lacking Wls in motoneurons. The 19 Wnts have different expression patterns in motoneurons, Schwann cells, and muscles (Figure 1). Dominant Wnts in motoneurons are Wnt7A, Wnt3, Wnt7B, and Wnt4 (Figure 1), of which Wnt7A and Wnt7B are sensitive to Wls regulation (Figure 7). In addition, these two Wnts are able to diminish NMJ deficits in HB9-Wls-/- mice (Figure 7). A parsimonious explanation of these results is that motoneurons release Wnt7A and Wnt7B to regulate NMJ development. Knockout Wls in Schwann cells or muscles alone or together with motoneurons seemed to have little effect on NMJ morphology (Figure 2—figure supplements 1–3). In agreement, no apparent NMJ deficits were detected when Wls was knocked out in muscles by Pax3::Cre (Remédio et al., 2016). These results could suggest that motoneuron Wnts cannot be compensated by Wnts from skeletal muscle fibers or Schwann cells. A more detailed analysis of NMJs in the triple knockout mice will help to dissect whether motoneuron Wnts compensate Wnts from Schwann cells or muscles during NMJ formation.

How does motoneuron Wnts regulate NMJ development? In vitro, Wnt7A was shown to inhibit agrin-induced AChR clustering (Barik et al., 2014b). However, we did not observe a change in AChR cluster size or intensity after Wnt7A injection (data not shown). Rather, Wnt7A reduced the number of axonal swellings and functionally rescued neurotransmission impairment in the HB9-Wls-/-, suggesting that Wnt7A mainly acts on nerve terminals in the mutant. This notion is supported following two lines of evidence. First, the unique presynaptic phenotypes of HB9-Wls-/- mice - swollen axon terminals and enlarged varicosities were not observed in MuSK or Rapsyn mutant mice where AChR clusters were abolished and nerve terminals arborize extensively (without swellings, however) (DeChiara et al., 1996; Gautam et al., 1995; Li et al., 2016). Second, they were not observed in muscle-specific β-catenin knockout or transgenic mice overexpressing β-catenin in muscles (Li et al., 2008; Liu et al., 2012; Wu et al., 2015b; Wu et al., 2012a). Wnt7A has been reported to initiate both canonical and non-canonical pathway (Heasley and Winn, 2008; Posokhova et al., 2015; von Maltzahn et al., 2011; Winn et al., 2005; Yang et al., 2012). However, motoneuron-specific deletion of β-catenin does not impair NMJ development (Li et al., 2008; Wu et al., 2015a), excluding a role of canonical pathway in motoneurons. Interestingly, Wnt non-canonical JNK pathway has been reported to regulate axonal transport of vesicles in C. elegans (Byrd et al., 2001). Many Frizzled were detectable in the ventral horns of the spinal cord, Schwann cells, and muscle cells (Figure 1—figure supplement 1). Future experiments are warranted to determine whether Wnt non-canonical pathways and which wnt receptor(s) regulate NMJ development in mammals.

Presynaptic defects in HB9-Wls-/- mice occur prior to postsynaptic defects evidenced by the reduction of mEPP frequency, but not amplitude, in P1 mutant (Figure 5). The cause to postsynaptic deficits could be complex. Because Wnts can directly regulate postsynaptic AChR assembly (Barik et al., 2014b; Zhang et al., 2012), lack of Wnts from motoneurons may thus impair postsynaptic assembly. When N-box of the AChR ε-subunit promoter is mutated, this leads to AChR deficiency and severely impaired transmission at the NMJ, but postsynaptic membrane appears normal (Nichols et al., 1999). On the other hand, neurotransmission blockade by botulinum toxin disrupts NMJ morphology (Rogozhin et al., 2008). Problems with presynaptic assembly may impair synaptic basal lamina. In fact, in mutant mice synaptic basal lamina was disorganized without a high-density midline (Figure 4). This could be a mechanism of loss of postsynaptic structures.

SV2 staining is progressively restricted from axon to the nerve terminal during NMJ development (Lupa and Hall, 1989). This is thought to indicate active vesicle transportation during neural development. In agreement, SV2 signal was low and sparse in wild type mice at two months of age (Figure 6B). However, SV2 and Synapsin 1 another marker of synaptic vesicle, were increased in axons of adult mutant mice by staining and Western blot analysis (Figure 6B and C), which is similar to the phenotype when autophagy is dysregulated in axon (Lüningschrör et al., 2017). This could be a reason to reduced synaptic vesicles at mutant NMJs. At the same, increased vesicle proteins in axons may be indicative of axon regeneration in the absence of Wls. In addition to increased SV2 levels, SV2-positive aggregates were detected in Wls mutants (Figure 6B). However, the nature of these aggregates is unknown. They may be inclusion bodies containing SV2 proteins for degradation. Nevertheless, such axon terminals swellings and varicosities are observed in NMJs recovering from Botulinum toxin poisoning (Rogozhin et al., 2008) and resemble those of nerve degeneration during aging (Bhattacharya et al., 2016; Lüningschrör et al., 2017; Valdez et al., 2010). Together, these observations are indicative of reinnervation in response to functional denervation. It would be interesting to determine whether Wnts from motoneurons regulate synapse regeneration and their loss of function precipitates the NMJ senescence.

It is worthy pointing out that the above conclusion is based on an assumption that Wls is a cargo receptor for all 19 Wnts’ secretion (Gross et al., 2012). Apparently, this might not be the case because individual Wnt has different sensitivities to Wls deficiency (Figure 7). In agreement, Zebrafish Wls mutants are surprisingly normal in early development, in contrast to Wnt5B and Wnt11 mutants (Kuan et al., 2015). It is possible that some Wnts may be regulated in a Wls-independent manner. For example, Wls-antisense morpholino was shown to severely impair membrane localization of Wnt5B, but not Wnt11 in zebrafish (Wu et al., 2015a). WntD in Drosophila does not require lipidation by Porcupine and thus may be secreted independent of Wls (Chen et al., 2012; Ching et al., 2008; Richards et al., 2014). Wnt3A seems not responsive to Wls overexpression (Galli et al., 2016). It is also possible that there is another not-yet-reported Wls-like homology gene in the rodents. It would be interesting to determine whether NMJ development is regulated by Wls-insensitive and/or Wls-independent Wnts.

Materials and methods

Materials

Wlsf/f mice, in which the exon 3 of Wls gene was flanked by loxp sites, were described previously (Fu et al., 2011). The following mouse lines were described in the literature: HB9::Cre (Wu et al., 2012b), HSA::Cre (Wu et al., 2012b), and Wnt1::Cre (Fu et al., 2011). Rosa-tdTomato (Jax, Cat# 007909) and Rosa-Laz (Jax, Cat# 003309) were purchased from Jackson Laboratory (Bar Harbor, ME). Unless otherwise indicated, control mice were either relevant floxed or Cre littermates. Mice were housed in a room with a 12 hr light/dark cycle with ad libitum access to water and rodent chow diet. Experiments with animals were approved by Institutional Animal Care and Use Committees of Augusta University, Case Western Reserve University, and Zhejiang University. HEK293T cells were purchased from ATCC (RRID:CVCL_0063) and were certified authentic and mycoplasma free.

Grip strength measurement

Muscle strength was determined using an SR-1 hanging scale (American Weigh Scales), as described previously (Shen et al., 2013). Briefly, mice were held by the tail, allowed to grip a grid connected to the scale. They were gently pulled horizontally until the grip was released. Top five values will be scored.

Light microscopy analysis

Whole-mount staining of muscles was performed as described previously (Shen et al., 2013; Wu et al., 2012b). Briefly, diaphragm or gastrocnemius were fixed in 4% paraformaldehyde (PFA) in PBS at 4°C overnight. Muscles were rinsed with PBS and incubated with 0.1 M glycine in PBS for 30 min. After being permeabilized with 0.5% Triton X-100 in PBS for 1 hr, muscles were blocked in the blocking buffer (3% BSA, 5% goat serum, and 0.5% Triton X-100 in PBS) for 2 hr, followed by incubation with primary antibodies against neurofilament (1:1000, Cell Signaling, Cat# 2837S), synaptophysin (1:1000, Dako, Cat# M7315), or SV2 (1:1000, Developmental Studies Hybridoma Bank, Cat # SV2) in the blocking buffer at 4°C overnight. Samples were washed 3 times with 0.5% Triton X-100 in PBS and incubated with a mixture of Alexa Fluor 488-conjugated secondary antibodies (1:750; Thermo Fisher Scientific, Cat# A11001 and A11008) and Alexa Fluor 594-conjugated β-bungarotoxin (BTX) (1:2000, Sigma-Aldrich, Cat# 0006) for 2 hr. After washing 3 times with 0.5% Triton X-100 in PBS, samples were flat-mounted in Vectashield mounting medium including DAPI (Vector Laboratories). Other primary antibodies for immunostaining were anti-Wls (rabbit, 1:200, home-made) (Yu et al., 2010) and anti-ChAT (goat, 1:200, Millipore, Cat# AB144P). Z-serial images were collected with a Zeiss confocal laser scanning microscope (LSM 510 META 3.2) and collapsed into a single image.

Electron microscopy (EM) analysis

EM analysis was performed as described previously (Barik et al., 2014a). Diaphragms were fixed in 2% glutaraldehyde and 2% PFA in PBS for 1 hr and then fixed in 1% osmium tetroxide in sodium cacodylate buffer (pH 7.3) for 1 hr at 25°C. After washing 3 times with PBS, tissues were dehydrated through a series of ethanol (30%, 50%, 70%, 80%, 90%, and 100%). After 3 rinses with 100% propylene oxide, samples were embedded in plastic resin (EM-bed 812; EMSciences). Serial thick sections (1–2 μm) of tissue blocks were stained with 1% toluidine blue, cut into ultrathin sections, mounted on 200-mesh unsupported copper grids, and stained with uranyl acetate (3% in 50% methanol) and lead citrate (2.6% lead nitrate and 3.5% sodium citrate, pH 12.0). Electron micrographs were taken using a JEOL 100CXII operated at 80 KeV. Synaptic vesicles in randomly-selected areas within 500 nm from presynaptic membrane were quantified. Junctional folds and synaptic cleft width were quantified in randomly-selected areas where more than 5 consecutive junctional folds could be visibly defined.

Electromyography

Mice were anesthetized with ketamine and xylazine mixture (100 and 10 mg/kg body weight, respectively). The stimulation needle electrode (092-DMF25-S; TECA) was inserted near the sciatic nerve in the thigh. The reference needle electrode was inserted near the Achilles tendon, and the recording needle electrode was inserted into the middle of the gastrocnemius or TA muscles. The reference and recording electrodes were connected to an Axopatch 200B amplifier (Molecular Devices). Supramaximal stimulation was applied to the sciatic nerve with trains of 10 stimuli at 1, 5, 10, 20, 30, or 40 Hz (with a 30 s pause between trains). Compound muscle action potentials (CMAPs) were collected with a Digidata 1322A (Molecular Devices). Peak-to-peak amplitudes were analyzed in Clampfit9.2 (Molecular Devices).

Electrophysiological recording

Electrophysiological recording of neuromuscular transmission was performed as described previously (Li et al., 2016; Zhao et al., 2017). Left hemidiaphragms with ribs and phrenic nerve distal endings were quickly dissected from anesthetized mice and pinned on Sylgard gel in oxygenated (95% O2, 5% CO2), 26–28°C Ringer’s solution (136.8 mM NaCl, 5 mM KCl, 12 mM NaHCO3, 1 mM NaH2PO4, 1 mM MgCl2, 2 mM CaCl2, 11 mM d-glucose, pH 7.3). Microelectrodes (20–40 MΩ with 3 M KCl) were pierced into the center of muscle fibers. Resting membrane potentials remained stable throughout the experiment at approximately −65 to −75 mV. From each hemidiaphragm, more than 5 muscle fibers were recorded for a period longer than 3 min. Data were collected with an Axopatch 200B amplifier, digitized (10 kHz low-pass filtered) with Digidata 1322A, and analyzed in Clampfit 9.2.

Western blot

Western blot was performed as previously described (Shen et al., 2008), with primary antibodies: anti-Wls (rabbit, 1:500, home-made) (Yu et al., 2010), Synapsin 1 (rabbit, 1:1000, CST, 5297), anti-GAPDH (mouse, 1:5000, Proteintech, Cat# 60004), and anti-α-tubulin (mouse, 1:3000, Santa Cruz, Cat# 23948). After washing, membranes were incubated with secondary antibodies HRP-conjugated goat anti-mouse or rabbit IgG (1:5000; Thermo Fisher Scientific, Cat# 31430 and 31460). Immunoreactive bands were visualized using enhanced chemiluminescence (Thermo Fisher Scientific). Autoradiographic films were scanned with an Epson 1680 scanner, and captured images were analyzed with Image J.

Motoneuron sorting

Spinal cords were freshly isolated from HB9:Cre;tdTomato mice (P0) with cold Ca2+/Mg2+ free PBS. Tissues were minced to small pieces with scissors and digested by 0.25% Trypsin-EDTA at 37°C for 15–30 min. Trypsin activity was stopped by adding the same amount of 10% FBS/DMEM. Cells were dissociated from each other by repeated pipetting. After filtering through 40 µm mesh (BD, Becton,Dickinson and Company), cells were spun down by centrifugation at 1000 rpm for 3 min at 4°C and suspended in the sorting buffer (Ca2+/Mg2+ free PBS with 1 mM EDTA; 25 mM HEPES, pH 7.0, 1% FBS). GFP-labeled cells were isolated by fluorescent-activated cell sorting (FACS).

Real-time PCR

RNA was isolated with Trizol as previously described (Tao et al., 2013). Two micrograms of RNA were reverse-transcribed with iScript cDNA synthesis kit (Bio-Rad). Real-time PCR was performed with SYBR Green qPCR Master Mix (Thermo Fisher Scientific). PCR primers of Wnt ligands were previously described (Zhang et al., 2012). RNA levels were normalized with internal controls (GADPH) that were assayed simultaneously on same reaction plates.

Injection of Wnt proteins

Indicated Wnt proteins (10 ng/µl, 20 µl) (BD Company) were injected into right tibialis anterior (TA) with Hamilton syringe, once every three days. As control, TA muscles of left hindlimb were injected with vehicle. Thirty days later, mice were subjected for CMAP analysis, or TA muscles were collected for NMJ image analysis.

Statistics

Data were analyzed by two-tailed unpaired Student’s t test and one or two-way ANOVA. Unless otherwise indicated, data were expressed as mean ±SEM. p value less than 0.05 were considered significant.

Acknowledgments

We are grateful to histology core in the Augusta University and Zhejiang University for EM analysis; the members from Mei-Xiong and Shen laboratories for suggestions. This work was supported in part by grants from the National Key Research and Development Program of China (2017YFA0104903 to SC); the Zhejiang Provincial Natural Science Foundation of China (LR17H090001 to SC); the National Natural Science Foundation of China (31671040 to SC, 31701036 to KZ); and the National Institutes of Health (DE15654 and DE26936 to WH; NS082007, NS090083 and AG051510 to LM; and AG045781 to WCX).

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

Chengyong Shen, Email: cshen@zju.edu.cn.

Lin Mei, Email: lin.mei@case.edu.

Jeremy Nathans, Johns Hopkins University School of Medicine, United States.

Marianne E Bronner, California Institute of Technology, United States.

Funding Information

This paper was supported by the following grants:

  • National Key Research and Development Program of China 2017YFA0104903 to Chengyong Shen.

  • Natural Science Foundation of Zhejiang Province LR17H090001 to Chengyong Shen.

  • National Natural Science Foundation of China 31671040 to Chengyong Shen.

  • National Natural Science Foundation of China 31701036 to Kejing Zhang.

  • National Institutes of Health DE15654 and DE26936 to Wei Hsu.

  • National Institutes of Health AG045781 to Wen-Cheng Xiong.

  • National Institutes of Health NS082007, NS090083, AG051510 to Lin Mei.

Additional information

Competing interests

No competing interests declared.

Author contributions

Conceptualization, Formal analysis, Funding acquisition, Investigation, Writing—original draft, Writing—review and editing.

Conceptualization, Formal analysis, Writing—original draft, Writing—review and editing.

Conceptualization, Formal analysis, Investigation, Writing—original draft, Writing—review and editing.

Conceptualization, Formal analysis, Investigation, Writing—review and editing.

Formal analysis, Investigation, Writing—review and editing.

Investigation, Writing—review and editing.

Investigation, Writing—review and editing.

Investigation, Writing—review and editing.

Formal analysis, Investigation, Writing—review and editing.

Formal analysis.

Formal analysis.

Resources, Writing—review and editing.

Resources, Writing—review and editing.

Conceptualization, Supervision, Funding acquisition, Writing—review and editing.

Conceptualization, Supervision, Funding acquisition, Writing—original draft, Writing—review and editing.

Ethics

Animal experimentation: Experiments with animals were approved by Institutional Animal Care and Use Committees of Augusta University (2011-0393), Case Western Reserve University (2017-0115), and Zhejiang University (10262).

Additional files

Transparent reporting form
elife-34625-transrepform.docx (244.7KB, docx)
DOI: 10.7554/eLife.34625.015

Data availability

All data generated or analysed during this study are included in the manuscript and supporting files.

References

  1. Arber S, Han B, Mendelsohn M, Smith M, Jessell TM, Sockanathan S. Requirement for the homeobox gene Hb9 in the consolidation of motor neuron identity. Neuron. 1999;23:659–674. doi: 10.1016/S0896-6273(01)80026-X. [DOI] [PubMed] [Google Scholar]
  2. Barik A, Lu Y, Sathyamurthy A, Bowman A, Shen C, Li L, Xiong WC, Mei L. LRP4 is critical for neuromuscular junction maintenance. Journal of Neuroscience. 2014a;34:13892–13905. doi: 10.1523/JNEUROSCI.1733-14.2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Barik A, Zhang B, Sohal GS, Xiong WC, Mei L. Crosstalk between agrin and wnt signaling pathways in development of vertebrate neuromuscular junction. Developmental Neurobiology. 2014b;74:828–838. doi: 10.1002/dneu.22190. [DOI] [PubMed] [Google Scholar]
  4. Bartscherer K, Pelte N, Ingelfinger D, Boutros M. Secretion of Wnt ligands requires Evi, a conserved transmembrane protein. Cell. 2006;125:523–533. doi: 10.1016/j.cell.2006.04.009. [DOI] [PubMed] [Google Scholar]
  5. Bänziger C, Soldini D, Schütt C, Zipperlen P, Hausmann G, Basler K. Wntless, a conserved membrane protein dedicated to the secretion of wnt proteins from signaling cells. Cell. 2006;125:509–522. doi: 10.1016/j.cell.2006.02.049. [DOI] [PubMed] [Google Scholar]
  6. Bezakova G, Ruegg MA. New insights into the roles of agrin. Nature Reviews Molecular Cell Biology. 2003;4:295–309. doi: 10.1038/nrm1074. [DOI] [PubMed] [Google Scholar]
  7. Bhattacharya MR, Geisler S, Pittman SK, Doan RA, Weihl CC, Milbrandt J, DiAntonio A. TMEM184b promotes axon degeneration and neuromuscular junction maintenance. Journal of Neuroscience. 2016;36:4681–4689. doi: 10.1523/JNEUROSCI.2893-15.2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Byrd DT, Kawasaki M, Walcoff M, Hisamoto N, Matsumoto K, Jin Y. UNC-16, a JNK-signaling scaffold protein, regulates vesicle transport in C. elegans. Neuron. 2001;32:787–800. doi: 10.1016/S0896-6273(01)00532-3. [DOI] [PubMed] [Google Scholar]
  9. Chen Q, Takada R, Takada S. Loss of porcupine impairs convergent extension during gastrulation in zebrafish. Journal of Cell Science. 2012;125:2224–2234. doi: 10.1242/jcs.098368. [DOI] [PubMed] [Google Scholar]
  10. Ching W, Hang HC, Nusse R. Lipid-independent secretion of a Drosophila Wnt protein. The Journal of Biological Chemistry. 2008;283:17092–17098. doi: 10.1074/jbc.M802059200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Darabid H, Perez-Gonzalez AP, Robitaille R. Neuromuscular synaptogenesis: coordinating partners with multiple functions. Nature Reviews Neuroscience. 2014;15:703–718. doi: 10.1038/nrn3821. [DOI] [PubMed] [Google Scholar]
  12. DeChiara TM, Bowen DC, Valenzuela DM, Simmons MV, Poueymirou WT, Thomas S, Kinetz E, Compton DL, Rojas E, Park JS, Smith C, DiStefano PS, Glass DJ, Burden SJ, Yancopoulos GD. The receptor tyrosine kinase MuSK is required for neuromuscular junction formation in vivo. Cell. 1996;85:501–512. doi: 10.1016/S0092-8674(00)81251-9. [DOI] [PubMed] [Google Scholar]
  13. Escher P, Lacazette E, Courtet M, Blindenbacher A, Landmann L, Bezakova G, Lloyd KC, Mueller U, Brenner HR. Synapses form in skeletal muscles lacking neuregulin receptors. Science. 2005;308:1920–1923. doi: 10.1126/science.1108258. [DOI] [PubMed] [Google Scholar]
  14. Fu J, Ivy Yu HM, Maruyama T, Mirando AJ, Hsu W. Gpr177/mouse Wntless is essential for Wnt-mediated craniofacial and brain development. Developmental Dynamics. 2011;240:365–371. doi: 10.1002/dvdy.22541. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Fu J, Jiang M, Mirando AJ, Yu HM, Hsu W. Reciprocal regulation of Wnt and Gpr177/mouse Wntless is required for embryonic axis formation. PNAS. 2009;106:18598–18603. doi: 10.1073/pnas.0904894106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Galli LM, Zebarjadi N, Li L, Lingappa VR, Burrus LW. Divergent effects of porcupine and Wntless on WNT1 trafficking, secretion, and signaling. Experimental Cell Research. 2016;347:171–183. doi: 10.1016/j.yexcr.2016.07.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Galli S, Lopes DM, Ammari R, Kopra J, Millar SE, Gibb A, Salinas PC. Deficient Wnt signalling triggers striatal synaptic degeneration and impaired motor behaviour in adult mice. Nature Communications. 2014;5:4992. doi: 10.1038/ncomms5992. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gautam M, Noakes PG, Mudd J, Nichol M, Chu GC, Sanes JR, Merlie JP. Failure of postsynaptic specialization to develop at neuromuscular junctions of rapsyn-deficient mice. Nature. 1995;377:232–236. doi: 10.1038/377232a0. [DOI] [PubMed] [Google Scholar]
  19. Goodman RM, Thombre S, Firtina Z, Gray D, Betts D, Roebuck J, Spana EP, Selva EM. Sprinter: a novel transmembrane protein required for Wg secretion and signaling. Development. 2006;133:4901–4911. doi: 10.1242/dev.02674. [DOI] [PubMed] [Google Scholar]
  20. Gross JC, Chaudhary V, Bartscherer K, Boutros M. Active Wnt proteins are secreted on exosomes. Nature Cell Biology. 2012;14:1036–1045. doi: 10.1038/ncb2574. [DOI] [PubMed] [Google Scholar]
  21. Hall AC, Lucas FR, Salinas PC. Axonal remodeling and synaptic differentiation in the cerebellum is regulated by WNT-7a signaling. Cell. 2000;100:525–535. doi: 10.1016/S0092-8674(00)80689-3. [DOI] [PubMed] [Google Scholar]
  22. Hausmann G, Bänziger C, Basler K. Helping wingless take flight: how WNT proteins are secreted. Nature Reviews Molecular Cell Biology. 2007;8:331–336. doi: 10.1038/nrm2141. [DOI] [PubMed] [Google Scholar]
  23. Heasley LE, Winn RA. Analysis of Wnt7a-stimulated JNK activity and cJun phosphorylation in non-small cell lung cancer cells. Methods in Molecular Biology. 2008;468:187–196. doi: 10.1007/978-1-59745-249-6_14. [DOI] [PubMed] [Google Scholar]
  24. Hubbard SR, Gnanasambandan K. Structure and activation of MuSK, a receptor tyrosine kinase central to neuromuscular junction formation. Biochimica et Biophysica Acta (BBA) - Proteins and Proteomics. 2013;1834:2166–2169. doi: 10.1016/j.bbapap.2013.02.034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Jing L, Lefebvre JL, Gordon LR, Granato M. Wnt signals organize synaptic prepattern and axon guidance through the zebrafish unplugged/MuSK receptor. Neuron. 2009;61:721–733. doi: 10.1016/j.neuron.2008.12.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Katz B, Miledi R. The statistical nature of the acetycholine potential and its molecular components. The Journal of Physiology. 1972;224:665–699. doi: 10.1113/jphysiol.1972.sp009918. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kim N, Stiegler AL, Cameron TO, Hallock PT, Gomez AM, Huang JH, Hubbard SR, Dustin ML, Burden SJ. Lrp4 is a receptor for agrin and forms a complex with MuSK. Cell. 2008;135:334–342. doi: 10.1016/j.cell.2008.10.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kuan YS, Roberson S, Akitake CM, Fortuno L, Gamse J, Moens C, Halpern ME. Distinct requirements for Wntless in habenular development. Developmental Biology. 2015;406:117–128. doi: 10.1016/j.ydbio.2015.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Li L, Cao Y, Wu H, Ye X, Zhu Z, Xing G, Shen C, Barik A, Zhang B, Xie X, Zhi W, Gan L, Su H, Xiong WC, Mei L. Enzymatic activity of the scaffold protein rapsyn for synapse formation. Neuron. 2016;92:1007–1019. doi: 10.1016/j.neuron.2016.10.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Li L, Xiong WC, Mei L. Neuromuscular junction formation, aging, and disorders. Annual Review of Physiology. 2018;80:159–188. doi: 10.1146/annurev-physiol-022516-034255. [DOI] [PubMed] [Google Scholar]
  31. Li XM, Dong XP, Luo SW, Zhang B, Lee DH, Ting AK, Neiswender H, Kim CH, Carpenter-Hyland E, Gao TM, Xiong WC, Mei L. Retrograde regulation of motoneuron differentiation by muscle beta-catenin. Nature Neuroscience. 2008;11:262–268. doi: 10.1038/nn2053. [DOI] [PubMed] [Google Scholar]
  32. Liu Y, Sugiura Y, Wu F, Mi W, Taketo MM, Cannon S, Carroll T, Lin W. β-Catenin stabilization in skeletal muscles, but not in motor neurons, leads to aberrant motor innervation of the muscle during neuromuscular development in mice. Developmental Biology. 2012;366:255–267. doi: 10.1016/j.ydbio.2012.04.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Luo ZG, Wang Q, Zhou JZ, Wang J, Luo Z, Liu M, He X, Wynshaw-Boris A, Xiong WC, Lu B, Mei L. Regulation of AChR clustering by dishevelled interacting with MuSK and PAK1. Neuron. 2002;35:489–505. doi: 10.1016/S0896-6273(02)00783-3. [DOI] [PubMed] [Google Scholar]
  34. Lupa MT, Hall ZW. Progressive restriction of synaptic vesicle protein to the nerve terminal during development of the neuromuscular junction. The Journal of Neuroscience. 1989;9:3937–3945. doi: 10.1523/JNEUROSCI.09-11-03937.1989. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Lüningschrör P, Binotti B, Dombert B, Heimann P, Perez-Lara A, Slotta C, Thau-Habermann N, R von Collenberg C, Karl F, Damme M, Horowitz A, Maystadt I, Füchtbauer A, Füchtbauer EM, Jablonka S, Blum R, Üçeyler N, Petri S, Kaltschmidt B, Jahn R, Kaltschmidt C, Sendtner M. Plekhg5-regulated autophagy of synaptic vesicles reveals a pathogenic mechanism in motoneuron disease. Nature Communications. 2017;8:678. doi: 10.1038/s41467-017-00689-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. McMahan UJ. The agrin hypothesis. Cold Spring Harbor Symposia on Quantitative Biology. 1990;55:407–418. doi: 10.1101/SQB.1990.055.01.041. [DOI] [PubMed] [Google Scholar]
  37. Messéant J, Dobbertin A, Girard E, Delers P, Manuel M, Mangione F, Schmitt A, Le Denmat D, Molgó J, Zytnicki D, Schaeffer L, Legay C, Strochlic L. MuSK frizzled-like domain is critical for mammalian neuromuscular junction formation and maintenance. Journal of Neuroscience. 2015;35:4926–4941. doi: 10.1523/JNEUROSCI.3381-14.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Miyamoto MD, Volle RL. Enhancement by carbachol of transmitter release from motor nerve terminals. PNAS. 1974;71:1489–1492. doi: 10.1073/pnas.71.4.1489. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Nichols P, Croxen R, Vincent A, Rutter R, Hutchinson M, Newsom-Davis J, Beeson D. Mutation of the acetylcholine receptor epsilon-subunit promoter in congenital myasthenic syndrome. Annals of Neurology. 1999;45:439–443. doi: 10.1002/1531-8249(199904)45:4&#x0003c;439::AID-ANA4&#x0003e;3.0.CO;2-W. [DOI] [PubMed] [Google Scholar]
  40. Posokhova E, Shukla A, Seaman S, Volate S, Hilton MB, Wu B, Morris H, Swing DA, Zhou M, Zudaire E, Rubin JS, St Croix B. GPR124 functions as a WNT7-specific coactivator of canonical β-catenin signaling. Cell Reports. 2015;10:123–130. doi: 10.1016/j.celrep.2014.12.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Remédio L, Gribble KD, Lee JK, Kim N, Hallock PT, Delestrée N, Mentis GZ, Froemke RC, Granato M, Burden SJ. Diverging roles for Lrp4 and Wnt signaling in neuromuscular synapse development during evolution. Genes & Development. 2016;30:1058–1069. doi: 10.1101/gad.279745.116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Richards MH, Seaton MS, Wallace J, Al-Harthi L. Porcupine is not required for the production of the majority of Wnts from primary human astrocytes and CD8+ T cells. PLoS One. 2014;9:e92159. doi: 10.1371/journal.pone.0092159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Rogozhin AA, Pang KK, Bukharaeva E, Young C, Slater CR. Recovery of mouse neuromuscular junctions from single and repeated injections of botulinum neurotoxin A. The Journal of Physiology. 2008;586:3163–3182. doi: 10.1113/jphysiol.2008.153569. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Sanes JR, Lichtman JW. Induction, assembly, maturation and maintenance of a postsynaptic apparatus. Nature Reviews Neuroscience. 2001;2:791–805. doi: 10.1038/35097557. [DOI] [PubMed] [Google Scholar]
  45. Shen C, Chen Y, Liu H, Zhang K, Zhang T, Lin A, Jing N. Hydrogen peroxide promotes abeta production through JNK-dependent activation of gamma-secretase. Journal of Biological Chemistry. 2008;283:17721–17730. doi: 10.1074/jbc.M800013200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Shen C, Lu Y, Zhang B, Figueiredo D, Bean J, Jung J, Wu H, Barik A, Yin DM, Xiong WC, Mei L. Antibodies against low-density lipoprotein receptor-related protein 4 induce myasthenia gravis. Journal of Clinical Investigation. 2013;123:5190–5202. doi: 10.1172/JCI66039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Shen C, Xiong WC, Mei L. LRP4 in neuromuscular junction and bone development and diseases. Bone. 2015;80:101–108. doi: 10.1016/j.bone.2015.05.012. [DOI] [PubMed] [Google Scholar]
  48. Strochlic L, Falk J, Goillot E, Sigoillot S, Bourgeois F, Delers P, Rouvière J, Swain A, Castellani V, Schaeffer L, Legay C. Wnt4 participates in the formation of vertebrate neuromuscular junction. PLoS One. 2012;7:e29976. doi: 10.1371/journal.pone.0029976. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Tao Y, Chen YJ, Shen C, Luo Z, Bates CR, Lee D, Marchetto S, Gao TM, Borg JP, Xiong WC, Mei L. Erbin interacts with TARP γ-2 for surface expression of AMPA receptors in cortical interneurons. Nature Neuroscience. 2013;16:290–299. doi: 10.1038/nn.3320. [DOI] [PubMed] [Google Scholar]
  50. Tapia JC, Wylie JD, Kasthuri N, Hayworth KJ, Schalek R, Berger DR, Guatimosim C, Seung HS, Lichtman JW. Pervasive synaptic branch removal in the mammalian neuromuscular system at birth. Neuron. 2012;74:816–829. doi: 10.1016/j.neuron.2012.04.017. [DOI] [PubMed] [Google Scholar]
  51. Thaler J, Harrison K, Sharma K, Lettieri K, Kehrl J, Pfaff SL. Active suppression of interneuron programs within developing motor neurons revealed by analysis of homeodomain factor HB9. Neuron. 1999;23:675–687. doi: 10.1016/S0896-6273(01)80027-1. [DOI] [PubMed] [Google Scholar]
  52. Valdez G, Tapia JC, Kang H, Clemenson GD, Gage FH, Lichtman JW, Sanes JR. Attenuation of age-related changes in mouse neuromuscular synapses by caloric restriction and exercise. PNAS. 2010;107:14863–14868. doi: 10.1073/pnas.1002220107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. von Maltzahn J, Bentzinger CF, Rudnicki MA. Wnt7a-Fzd7 signalling directly activates the Akt/mTOR anabolic growth pathway in skeletal muscle. Nature Cell Biology. 2011;14:186–191. doi: 10.1038/ncb2404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Winn RA, Marek L, Han SY, Rodriguez K, Rodriguez N, Hammond M, Van Scoyk M, Acosta H, Mirus J, Barry N, Bren-Mattison Y, Van Raay TJ, Nemenoff RA, Heasley LE. Restoration of Wnt-7a expression reverses non-small cell lung cancer cellular transformation through frizzled-9-mediated growth inhibition and promotion of cell differentiation. Journal of Biological Chemistry. 2005;280:19625–19634. doi: 10.1074/jbc.M409392200. [DOI] [PubMed] [Google Scholar]
  55. Wu BT, Wen SH, Hwang SP, Huang CJ, Kuan YS. Control of Wnt5b secretion by Wntless modulates chondrogenic cell proliferation through fine-tuning fgf3 expression. Journal of Cell Science. 2015a;128:2328–2339. doi: 10.1242/jcs.167403. [DOI] [PubMed] [Google Scholar]
  56. Wu H, Barik A, Lu Y, Shen C, Bowman A, Li L, Sathyamurthy A, Lin TW, Xiong WC, Mei L. Slit2 as a β-catenin/Ctnnb1-dependent retrograde signal for presynaptic differentiation. eLife. 2015b;4:e07266. doi: 10.7554/eLife.07266. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Wu H, Lu Y, Barik A, Joseph A, Taketo MM, Xiong WC, Mei L. β-Catenin gain of function in muscles impairs neuromuscular junction formation. Development. 2012a;139:2392–2404. doi: 10.1242/dev.080705. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Wu H, Lu Y, Shen C, Patel N, Gan L, Xiong WC, Mei L. Distinct roles of muscle and motoneuron LRP4 in neuromuscular junction formation. Neuron. 2012b;75:94–107. doi: 10.1016/j.neuron.2012.04.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Wu H, Xiong WC, Mei L. To build a synapse: signaling pathways in neuromuscular junction assembly. Development. 2010;137:1017–1033. doi: 10.1242/dev.038711. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Yang G, Liu Y, Yang K, Liu R, Zhu S, Coquinco A, Wen W, Kojic L, Jia W, Cynader M. Isoform-specific palmitoylation of JNK regulates axonal development. Cell Death & Differentiation. 2012;19:553–561. doi: 10.1038/cdd.2011.124. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Yang PT, Lorenowicz MJ, Silhankova M, Coudreuse DY, Betist MC, Korswagen HC. Wnt signaling requires retromer-dependent recycling of MIG-14/Wntless in Wnt-producing cells. Developmental Cell. 2008;14:140–147. doi: 10.1016/j.devcel.2007.12.004. [DOI] [PubMed] [Google Scholar]
  62. Yu HM, Jin Y, Fu J, Hsu W. Expression of Gpr177, a Wnt trafficking regulator, in mouse embryogenesis. Developmental Dynamics. 2010;239:2102–2109. doi: 10.1002/dvdy.22336. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Zhang B, Liang C, Bates R, Yin Y, Xiong WC, Mei L. Wnt proteins regulate acetylcholine receptor clustering in muscle cells. Molecular Brain. 2012;5:7. doi: 10.1186/1756-6606-5-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Zhang B, Luo S, Wang Q, Suzuki T, Xiong WC, Mei L. LRP4 serves as a coreceptor of agrin. Neuron. 2008;60:285–297. doi: 10.1016/j.neuron.2008.10.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Zhang X, Zhu J, Yang GY, Wang QJ, Qian L, Chen YM, Chen F, Tao Y, Hu HS, Wang T, Luo ZG. Dishevelled promotes axon differentiation by regulating atypical protein kinase C. Nature Cell Biology. 2007;9:743–754. doi: 10.1038/ncb1603. [DOI] [PubMed] [Google Scholar]
  66. Zhao K, Shen C, Lu Y, Huang Z, Li L, Rand CD, Pan J, Sun XD, Tan Z, Wang H, Xing G, Cao Y, Hu G, Zhou J, Xiong WC, Mei L. Muscle yap is a regulator of neuromuscular junction formation and regeneration. The Journal of Neuroscience. 2017;37:3465–3477. doi: 10.1523/JNEUROSCI.2934-16.2017. [DOI] [PMC free article] [PubMed] [Google Scholar]

Decision letter

Editor: Jeremy Nathans1

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

[Editors’ note: this article was originally rejected after discussions between the reviewers, but the authors were invited to resubmit after an appeal against the decision.]

Thank you for submitting your manuscript "A novel autonomous regulation by Wnt for presynaptic development in the neuromuscular junction to eLife. Three experts carefully reviewed the manuscript. While the reviewers found the work interesting, the number of substantive questions raised was such that we feel we must reject it. We hope that the reviewers' comments will be useful to you. We apologize for not being able to deliver better news, and we hope that you will continue to consider eLife for future submissions.

Reviewer #1:

Synapse formation between neurons and their follower cells is an example of extreme subcellular differentiation at the site of their contact. It is driven by reciprocal interactions of pre- and postsynaptic cells. However, the underlying molecular and cellular mechanisms are poorly understood, even at the best investigated synapse, the neuromuscular junction (NMJ). The crucial factor from the motor neuron driving postsynaptic muscle differentiation is neural agrin acting through its receptor/effector LRP4 and RTK MuSK. In addition to driving postsynaptic muscle membranes in response to MuSK activation muscle fibers are thought to act back on motor neurites to drive nerve terminal differentiation through mechanisms that may involve Wnt signalling.

In recent years, evidence, some of it conflicting, for an involvement of Wnts, in various ways, in modulating rodent NMJ formation has been accumulating. Investigating their involvement in NMJ formation has been complicated by the fact, that some 18 Wnts have been identified in mammals. In the present paper, authors demonstrate by genetic approaches that Wnts secreted from motor neurons is involved in presynaptic terminal differentiation and function. Specifically, genetic deletion of Wnt ligand secretion mediator, Wls, in motor neurons, but not in Schwann cells or in muscle fibers, impair functional differentiation and maintenance of rodent NMJs. As Wls is required for sorting and transport to the cell surface and, thus, for secretion of various Wnts, these data indicate that Wnts secreted from motor neurons are involved in NMJ maintenance. Specifically, evidence is presented that the Wnt required for NMJ maintenance is Wnt 7A. Reinnervation experiments further show that Wls mutant motor axons show a deficit in their capacity to reinnervate old endplate sites upon severing the nerve. This paper is thus an important contribution towards full understanding of NMJ formation and implicates dysregulation of Wnt signalling as a potential mechanism of certain neuromuscular disorders.

Although overall, the data support the conclusion, I feel that some of the mechanisms proposed appear overinterpreted. Attention should be given to phrasing that accurately reflects what the figures actually show. The points below can probably be dealt with readily.

Figure 1D: Authors conclude from grip strength reduction in Wls mutants that muscle function is impaired. For a meaningful comparison of muscle function grip strengths should be compared not only between adults of same age, but also between animals of same weight! Part of the grip strength reduction in mutants could arise from lower muscle mass associated with lower body weight of the latter.

Results section: NMJ formation appears………: It is stated that NMJs are characterized in mouse embryos. Figure 2 suggests that images were taken from older animals: fetuses or newborn?

The resolution of Figure 2A is not sufficient to show that morphologies of nerve terminals were comparable. – Note: typo: "defasciculation".

Age-dependent……: "The mild phenotypes…….". Are the phenotypes mild or absent? P- Values in the para above suggests they NMJ parameters are indistinguishable. Either improve statistics by increasing sample size in Figure 2C or accept that phenotypes are indistinguishable.

In Figure legend 2, the age of the animals should be given.

Figure 3: In panel B, calibration bar is missing. Varicosities similar to those seen primarily in Wls mutants are also observed in NMJs recovering from Botulinus toxin poisoning. Thus, they may be indicative of reinnervation in response to functional denervation, as is suggested by neuromuscular transmission failure in the mutants (Figure 5). This should be mentioned in the Discussion section.

It is striking that at two months, SV2 staining in w.t. muscle is still strong, as in Figure 3A. Lupa and Hall, (1989) have described that unlike during fetal development, SV2 is no longer visible in motor axons of adult animals (>2 weeks postnatal). Have the authors made out a difference between SV2 stainings in w.t. vs Wls mutant axons, as is suggested by the Western blot in Figure 6D? This could be important for the interpretation of varicosity formation potentially reflecting axon regeneration in Wls mutants. Discuss!

In Figure legend 3, the age of the animals should be given.

Reduced synaptic vesicles……..: It is not clear, how from EM sections such as those in Figure 4A centre, with their inhomogenous distribution of synaptic vesicles in the nerve terminals, synaptic vesicle densities can be derived quantitatively (as claimed in Figure 4B). Describe!. Were there also mutant terminals with normal vesicle densities, as would be expected from high SV2 aggregates opposed to high density AChR clusters (see Figure 6B)? Furthermore, is the lack of a basal lamina in the synaptic cleft (as shown in Figure 4A bottom), commonly observed in the mutants? It could be related to loss of postsynaptic structures.

Subsection “Impaired neuromuscular transmission”: it is stated that "postsynaptic deficits were likely caused by reduced neuromuscular transmission". This interpretation is unlikely to be correct. For example, in AChR deficiency myasthenic syndrome (through a mutation in the AChRe promoter (specifically its N-box) with severely impaired transmission, postsynaptic membrane appears normal.

Clustered synaptic vesicles: based on differences in SV2 stainings, in their localized distribution in nerve terminals and accumulations in varicosities, and the higher SV2 immunoreactivity in sciatic nerves (Figure 6B, C, D) between w.t. and Wls mutant mice, it is suggested that synaptic vesicle transport may be impaired in Wls ablated axons, which would explain the lower vesicle density in the mutant nerve terminals. Although SV2 is considered a vesicle-specific protein, I find it risky to conclude from local aggregations in SV2 staining intensity the local aggregation of synaptic vesicles. For confirmation, EM analysis of vesicle densities in axonal varicosities would be required. Furthermore, in Figure 6B, some of the strong presynaptic SV2 accumulations in Wls mutants are opposed to strong AChR clusters, which is not consistent with the EM data in Figure 4. Explain!

Statistical tests: For every experiment analysed by statistical tests, authors should give sample sizes: e.g. how many mepp amplitudes were sampled for calculation of a mean {plus minus} S.E. When were paired or unpaired t-tests used?

Reviewer #2:

This manuscript reported results to support the conclusion that Wnt signaling pathways act autonomously in presynaptic development at the neuromuscular junction (NMJ). The key approach is genetic deletion of the Wnt ligand secretion mediator (Wls) gene in motor neurons, muscles or Schwann cells. The analysis and the results in the current version of manuscript have not convincingly supported the conclusion.

1) Even though no detectable NMJ defects were found in the HSA::Cre and Wnt1::Cre Wls mutants, it would be crucial to show Wls is indeed completely absent in muscles and Schwann cells and there is no possibility that other Wls-like molecules can compensate for the loss. In particular, muscle fibers contain multiple nuclei and the reporter system (td-Tomato) cannot prove a complete deletion of Wls in all nuclei because the reporter molecule will diffuse and label the entire muscle fiber. It is a lingering concern for the HAS::Cre line when used in studying the NMJ at early developmental stages.

2) As Wnts are presumably secreted and bind their cognate receptors to function, a central question is what Wnts are expressed and secreted from muscles and Schwann cells. For example, does either cell type express Wnt 7A?

3) The HB9::Cre Wls mutants display grip strength deficits. But the analysis of the NMJ has only been done in the diaphragm muscle. It would be important to analyze other muscles that are directly relevant to grip strength phenotype.

4) If exogenous ACh agonists are applied, is the endplate potential normal?

5) The defect in synaptic vesicle aggregation is interesting. The authors tried to show increased SV2 protein levels in the sciatic nerve to support the finding, but the Figure 6C is not convincing. Since the sciatic nerve is mixed with many non-motor axons, it would have been more convincing to use pure motor nerves, e.g. phrenic nerve.

6) The nerve transplantation experiment is interesting, but it is less relevant to the conclusion. They can be completely removed.

7) The partial rescue of NMJ defects via injecting individual Wnt proteins is interesting but the results do not directly support the conclusion on cell autonomous Wnt signaling. It is a little bit hard to interpret the results without knowing what relevant Wnt receptors are expressed in cells involved in NMJ formation or maintenance. For example, even Wnt7A expression/secretion from motor neurons is reduced, it cannot rule out that motor neuron-derived Wnt7A activates muscle (or Schwann cell) pathways to elicit retrograde signals to promote presynaptic development unless the receptor(s) for Wnt7A (or other relevant Wnt molecules) is selectively deleted from motor neurons, not other cell types.

Reviewer #3:

Here, Shen et al. report a very interesting study on a novel role for Wnt in presynaptic assembly in the neuromuscular junction. The authors used the main Wnt secretion mediator Wls to specifically abolish Wnt secretion in each of the three NMJ compartments (skeletal muscle, motor neurons and Schwann cells). They showed that only Wls deletion in motor neurons leads to age-dependent NMJ deficits and neurotransmission impairment in a cell-autonomous manner. Using a combination of morphological analysis, electron microscopy and electrophysiology, they characterized the NMJ deficits and nicely demonstrate that motoneuronal specific deletion of Wls results in progressive NMJ degeneration, muscle weakness, impaired neurotransmission, motor axonal swelling and synaptic vesicle reduction. They also showed that motoneuronal Wls is required for NMJ regeneration. From there, they screened for the Wnt ligands that are expressed in motoneurons and identified those whose secretion is regulated by Wls in heterologous HEK cells. Wnt5A and 7A were the most sensitive to Wls mutation. Injection of recombinant Wnt7A, but not Wnt5A in TA muscle can partially rescue the morphological presynaptic deficits of HB9-Wl-/-. Overall this is a very interesting study, the analysis of data was carefully performed and most of the main conclusions are supported by the provided results. However, although the finding that Wnt7A may play a role in presynaptic assembly is novel, the current study does not provide sufficient evidence to support this conclusion and it remains unclear how Wnts secretion from motoneurons contributes to presynaptic functions resulting in age-dependent NMJ dysfunction when altered.

The authors state that Wls deletion in motoneurons, but not skeletal muscle or Schwann cells shows NMJ defects. Given that several Wnts identified by the authors in motoneurons are known to regulate postsynaptic differentiation, some of them being low or not sensitive to Wls, there is a possibility that Wnts secreted by the motoneurons may have compensated the muscle and/or Schwann cells NMJ phenotypes. A more detailed analysis of the NMJ phenotypes during NMJ formation, including combined HSA and/or HB9 and/or Wnt1-Wls-/- should be performed to support their conclusions. I realized that it is a demanding task, but this is not unfeasible.

The authors identified several Wnts molecules expressed in motoneurons and showed that apart from Wnt7A and Wnt5A, several other (Wnt7B and Wnt9B) were sensitive to Wls. It is not clear why these Wnts were not tested in the rescue experiment. Also, how Wnt overexpression in the muscle can rescue the presynaptic NMJ deficits? Neurotransmission analysis could be performed following Wnt7A injection.

The authors previously reported that Wnt7A modulates agrin-induced AChR clustering in muscle cell (Barik et al., 2014) suggesting that Wnt7A regulates postsynaptic differentiation. This should be included and commented in the manuscript.

Developmental, post-natal stages and muscle types used should be mentioned in the figure legends.

In several figures the title is misleading. The authors should replace Wl-/- by HB9-Wl-/-.

Figure 5 and Figure 3—figure supplement 1Figure: Please clarify the title.

Figure 3—figure supplement 1Figure: Please clarify how synapse elimination was evaluated.

Figure 8D: The graph legend is missing.

Figure 1C: Co-staining of Wls with a motoneuron marker should be performed to support the conclusion.

eLife. 2018 Aug 16;7:e34625. doi: 10.7554/eLife.34625.018

Author response

[Editors’ note: the author responses to the first round of peer review follow.]

Reviewer #1:

Synapse formation between neurons and their follower cells is an example of extreme subcellular differentiation at the site of their contact. It is driven by reciprocal interactions of pre- and postsynaptic cells. However, the underlying molecular and cellular mechanisms are poorly understood, even at the best investigated synapse, the neuromuscular junction (NMJ). The crucial factor from the motor neuron driving postsynaptic muscle differentiation is neural agrin acting through its receptor/effector LRP4 and RTK MuSK. In addition to driving postsynaptic muscle membranes in response to MuSK activation muscle fibers are thought to act back on motor neurites to drive nerve terminal differentiation through mechanisms that may involve Wnt signalling.

In recent years, evidence, some of it conflicting, for an involvement of Wnts, in various ways, in modulating rodent NMJ formation has been accumulating. Investigating their involvement in NMJ formation has been complicated by the fact, that some 18 Wnts have been identified in mammals. In the present paper, authors demonstrate by genetic approaches that Wnts secreted from motor neurons is involved in presynaptic terminal differentiation and function. Specifically, genetic deletion of Wnt ligand secretion mediator, Wls, in motor neurons, but not in Schwann cells or in muscle fibers, impair functional differentiation and maintenance of rodent NMJs. As Wls is required for sorting and transport to the cell surface and, thus, for secretion of various Wnts, these data indicate that Wnts secreted from motor neurons are involved in NMJ maintenance. Specifically, evidence is presented that the Wnt required for NMJ maintenance is Wnt 7A. Reinnervation experiments further show that Wls mutant motor axons show a deficit in their capacity to reinnervate old endplate sites upon severing the nerve. This paper is thus an important contribution towards full understanding of NMJ formation and implicates dysregulation of Wnt signalling as a potential mechanism of certain neuromuscular disorders.

Although overall, the data support the conclusion, I feel that some of the mechanisms proposed appear overinterpreted. Attention should be given to phrasing that accurately reflects what the figures actually show. The points below can probably be dealt with readily.

We thank reviewer 1 for the remarks that “This paper is thus an important contribution towards full understanding of NMJ formation”; “overall, the data support the conclusion” and his/her constructive critiques and suggestions that have significantly improved the manuscript.

Figure 1D: Authors conclude from grip strength reduction in Wls mutants that muscle function is impaired. For a meaningful comparison of muscle function grip strengths should be compared not only between adults of same age, but also between animals of same weight! Part of the grip strength reduction in mutants could arise from lower muscle mass associated with lower body weight of the latter.

This is a good suggestion. The revision also compared grip strength between two groups of mice with similar body weights (29.3 + 1.1 g vs 27.7 + 1.6 g for control and mutants, respectively, p > 0.05). As shown in revised Figure 2G, HB9-Wls-/- mice had reduced grip strength (129 + 9.7 g in control mice; 71.8 + 7.9 g in mutant mice; p < 0.01) (–subsection “Reduced body weight and muscle strength in Wls motoneuron knockout mice”).

Results section: NMJ formation appears………: It is stated that NMJs are characterized in mouse embryos. Figure 2 suggests that images were taken from older animals: fetuses or newborn?

Sorry for the confusion. It shows NMJs of E18.5 mice. More representative, high-resolution images are provided in revised Figure 2—figure supplement 2A.

The resolution of Figure 2A is not sufficient to show that morphologies of nerve terminals were comparable. – Note: typo: "defasciculation".

Corrected.

Age-dependent……: "The mild phenotypes…….". Are the phenotypes mild or absent? P- Values in the para above suggests they NMJ parameters are indistinguishable. Either improve statistics by increasing sample size in Figure 2C or accept that phenotypes are indistinguishable.

Good point. The reviewer is correct that there were no detectable deficits of NMJs in mutant embryonic and P0 mice (in terms of gross morphology, endplate width, AChR size, and NMJ density). “mild” was a word of bad choice. This sentence has now been revised as “Normal NMJs at P0 pups were unable to explain…” (subsection “Age-dependent nerve terminal swellings and poor innervation”).

In Figure legend 2, the age of the animals should be given.

Good point. Information on ages of animals has now provided to legend to revised Figure 2—figure supplement 2 and all other figures in revised manuscript.

Figure 3: In panel B, calibration bar is missing.

Calibration bar was added in revised Figure 3B.

Varicosities similar to those seen primarily in Wls mutants are also observed in NMJs recovering from Botulinus toxin poisoning. Thus, they may be indicative of reinnervation in response to functional denervation, as is suggested by neuromuscular transmission failure in the mutants (Figure 5). This should be mentioned in the Discussion section.

Good suggestion. This is now discussed as suggested (Discussion section).

It is striking that at two months, SV2 staining in w.t. muscle is still strong, as in Figure 3A. Lupa and Hall, (1989) have described that unlike during fetal development, SV2 is no longer visible in motor axons of adult animals (>2 weeks postnatal). Have the authors made out a difference between SV2 stainings in w.t. vs Wls mutant axons, as is suggested by the Western blot in Figure 6D? This could be important for the interpretation of varicosity formation potentially reflecting axon regeneration in Wls mutants. Discuss!

The reviewer is correct that SV2 staining is weak in axons at ages of 2 weeks and old although remains positive at terminals. This is in agreement with our data in Figure 6B where SV2 signal was weak in axons at two months of mice (note axons could be readily labeled by anti-neurofilament antibody). Please note that NMJ images in Figure 3A were obtained with antibodies also against neurofilament and thus showed strong axon staining.

We would like to point out that SV2 staining was detectable in axons of wild type mice at ages of two months, but the staining was sparse (Figure 6B, empty arrowhead). However, SV2 was increased in axons of adult mutant mice by staining and Western blot analysis (Figure 6B and Figure 6C). This is similar to a previous report that enhancement of synaptic vesicle proteins in sciatic nerves when autophagy is impaired (Luningschror et al., 2017).

In addition to increased SV2 levels, SV2-positive aggregates were detected in Wls mutants (Figure 6B). Interestingly, such varicosities were also observed in NMJs recovering from Botulinum toxin poisoning (Rogozhin et al., 2008). Together, these observations are indicative of reinnervation in response to functional denervation. As suggested, the revised manuscript added a brand-new paragraph to discuss implication of these results (Discussion section). We thank the reviewer for this suggestion.

In Figure legend 3, the age of the animals should be given.

Good point. Information on ages of animals has now provided to legend to Figure 3 and all other figures in revised manuscript.

Reduced synaptic vesicles……..: It is not clear, how from EM sections such as those in Figure 4A centre, with their inhomogenous distribution of synaptic vesicles in the nerve terminals, synaptic vesicle densities can be derived quantitatively (as claimed in Figure 4B). Describe!.

Sorry for the oversight. Synaptic vesicles in randomly-selected areas within 500 nm from presynaptic membrane were quantified. Junctional folds and synaptic cleft width were quantified in randomly-selected areas where more than 5 consecutive junctional folds could be visibly defined. This information is now provided in the revised manuscript (subsection “Electron microscopy (EM) analysis”).

Were there also mutant terminals with normal vesicle densities, as would be expected from high SV2 aggregates opposed to high density AChR clusters (see Figure 6B)?

Yes, some nerve terminals in mutant mice appeared to be normal (~47%) (Figure 3A, 3E and 3F).

Furthermore, is the lack of a basal lamina in the synaptic cleft (as shown in Figure 4A bottom), commonly observed in the mutants? It could be related to loss of postsynaptic structures.

Good point. As shown in Figure 4A, in control NMJs, basal lamina was visible in synaptic cleft and between junctional folds, with highest electron density in the middle (empty arrowheads). However, basal lamina density was reduced and poorly defined in mutant mice. This could be a mechanism of loss of postsynaptic structures. The revised manuscript now describes findings on basal lamina in Results section and discussed their implications in Discussion section.

Subsection “Impaired neuromuscular transmission”: it is stated that "postsynaptic deficits were likely caused by reduced neuromuscular transmission". This interpretation is unlikely to be correct. For example, in AChR deficiency myasthenic syndrome (through a mutation in the AChRe promoter (specifically its N-box) with severely impaired transmission, postsynaptic membrane appears normal.

We agree. The revised Discussion section has discussed alternative interpretations. For example, it may be caused by reduced and poorly organized synaptic basal lamina.

Clustered synaptic vesicles: based on differences in SV2 stainings, in their localized distribution in nerve terminals and accumulations in varicosities, and the higher SV2 immunoreactivity in sciatic nerves (Figure 6B, C, D) between w.t. and Wls mutant mice, it is suggested that synaptic vesicle transport may be impaired in Wls ablated axons, which would explain the lower vesicle density in the mutant nerve terminals. Although SV2 is considered a vesicle-specific protein, I find it risky to conclude from local aggregations in SV2 staining intensity the local aggregation of synaptic vesicles. For confirmation, EM analysis of vesicle densities in axonal varicosities would be required. Furthermore, in Figure 6B, some of the strong presynaptic SV2 accumulations in Wls mutants are opposed to strong AChR clusters, which is not consistent with the EM data in Figure 4. Explain!

We agree with the reviewer’s comments. First, we have performed additional EM analysis; however, we have encountered a technical difficulty. NMJ occupies 0.01-0.1% of muscle fiber surface. Presynaptic vesicles/membrane are identified by their proximity to characteristic postsynaptic junctional folds. In Wls mutant mice, NMJs were disorganized and junctional folds were reduced. In addition, as shown in Figure 6B, the area of SV2+ aggregates was about 5 µm and less than 2% of a NMJ. We tried but were unable to convincingly identify vesicle clusters in axons by EM. Second, the nature of SV2+ aggregates is unknown. They may be inclusion bodies containing SV2 proteins for degradation. Third, we wish to point out that SV2 staining was low in wild type axons in adult mice, but it was increased in axons of mutant mice (Figure 6B). To test this hypothesis, we performed additional Western blot analysis of Synapsin 1, another marker of synaptic vesicles. As shown in revised Figure 6C, Synapsin 1 level was increased in Wls mutant axons, compared with control axons. This is in agreement with increased levels of SV2 in axons (Figure 6C). Together, the results suggest possible accumulation of vesicle-specific proteins in mutant axons. Although it is difficult to attribute these proteins to aggregated vesicles (because technical difficulty described above), their increased levels in axons may lead to a reduction at presynaptic terminals.

Regarding “strong presynaptic SV2 accumulations in Wls mutants”, we would like to point out that SV2 signaling is discontinuous and spotty, faint or missing in many areas where AChR was present. As described above, they may result from inclusion bodies containing SV2 proteins for degradation.

We have discussed these points in revised Discussion section.

Statistical tests: For every experiment analysed by statistical tests, authors should give sample sizes: e.g. how many mepp amplitudes were sampled for calculation of a mean {plus minus} S.E. When were paired or unpaired t-tests used?

Statistic information has now been provided in legends to all figures in revised manuscript.

Reviewer #2:

This manuscript reported results to support the conclusion that Wnt signaling pathways act autonomously in presynaptic development at the neuromuscular junction (NMJ). The key approach is genetic deletion of the Wnt ligand secretion mediator (Wls) gene in motor neurons, muscles or Schwann cells. The analysis and the results in the current version of manuscript have not convincingly supported the conclusion.

We thank for reviewer 2 for the above remark that “manuscript reported results to support the conclusion that Wnt signaling pathways act autonomously in presynaptic development at the neuromuscular junction (NMJ)” and the following remarks in individual comments “The defect in synaptic vesicle aggregation is interesting…”; “The nerve transplantation experiment is interesting…” and “The partial rescue of NMJ defects via injecting individual Wnt proteins is interesting…” and his or her constructive critiques and suggestions that have significantly improved the manuscript.

1) Even though no detectable NMJ defects were found in the HSA::Cre and Wnt1::Cre Wls mutants, it would be crucial to show Wls is indeed completely absent in muscles and Schwann cells and there is no possibility that other Wls-like molecules can compensate for the loss. In particular, muscle fibers contain multiple nuclei and the reporter system (td-Tomato) cannot prove a complete deletion of Wls in all nuclei because the reporter molecule will diffuse and label the entire muscle fiber. It is a lingering concern for the HAS::Cre line when used in studying the NMJ at early developmental stages.

Good suggestions. We have performed additional experiments. As shown in revised suppl Figure 1B and 1F, levels of Wls were reduced in muscles of HSA-Wls-/- mice and sciatic nerves of Wnt1-Wls-/- mice, respectively, compared with control mice. The reduction in mutant muscles and sciatic nerves was not 100% because these tissues contain cells where Cre was not expressed (such as blood vessels, nerve terminals, and Schwann cells in suppl Figure 1B and axons in suppl Figure 1F). Although there was no observable defects of NMJ formation in Wnt1-Wls-/- mice, we found that Wnt1-Wls-/- mice died soon after birth with severe craniofacial skeleton defects, in accord with the notion that Wnt signaling in neural crests is critical for craniofacial skeleton development (Fu et al., 2011).

It is a good question whether Wls was completely knocked out by HSA::Cre. Cre in HSA::Cre mice is expressed in myotomal somites as early as E9.5, a time prior to NMJ formation (around E13.5). This line has been widely used by various laboratories to study NMJ development (Li et al., 2008; Escher et al., 2005; Jaworski et al., 2006; Wu et al., 2012; Seaberg et al. 2015). In newly provided data, Wls in HAS-Wls-/- muscles was reduced to ~5% of that of control mice (revised Figure 2—figure supplement 1B). As described above, the residual Wls protein could be from other cell types in muscle tissues such as blood vessel cells and even axon terminals. In agreement with these considerations, a recent study found no apparent NMJ deficits when Wls was knocked out by Pax3::Cre (Remedio et al. 2016).

Whether there is a Wls-like molecule is an outstanding question. However, no such a molecule has been identified so far. Wls knockout is embryonic lethal (Fu et al. 2009). Its conditional knockout has been shown to cause deficits in various tissues including gut (Valenta et al., 2016), spinal cord (Onishi et al. 2018), and craniofacial and brain development (Fu et al., 2011). This paper focuses on whether and which Wnts from motoneurons contribute to NMJ development. We would hope that the reviewer would agree whether there is Wls-like molecules could be addressed in a future study. This point has been discussed in the revised manuscript (Discussion section).

2) As Wnts are presumably secreted and bind their cognate receptors to function, a central question is what Wnts are expressed and secreted from muscles and Schwann cells. For example, does either cell type express Wnt 7A?

We performed additional experiments to analyze expression pattern of all 19 Wnts in Schwann cells and muscle cells. Dominant Wnts appeared to be Wnt5A and Wnt9A in Schwann cells and Wnt4, Wnt6 and Wnt9A in muscle cells (revised Figure 1). Wnt7A is not highly expressed in muscle cells [in agreement with an earlier report (Zhang et al., 2012)] or Schwann cells.

3) The HB9::Cre Wls mutants display grip strength deficits. But the analysis of the NMJ has only been done in the diaphragm muscle. It would be important to analyze other muscles that are directly relevant to grip strength phenotype.

Sorry for not being clear. In addition to diaphragm, NMJ analysis has been done in gastrocnemius (revised Figure 3A-3E, Figure 5A-5C, Figure 6A and 6B, and Figure 3—figure supplement 1) and tibialis anterior (Figure 7C-7E). These muscles are now clearly indicated in respective figure legends.

4) If exogenous ACh agonists are applied, is the endplate potential normal?

Good question. In newly performed experiments, we treated diaphragms with CCh (10 µM), an ACh agonist, for 2 min and observed its effects on mEPPs. Consistent with previous reports (Katz et al., 1972; Miyamoto et al. 1974), CCh could enhance mEPP frequency and amplitude in control diaphragms (revised Figure 5D-5I). Interestingly, CCh also elevated mEPP frequency and amplitude in Wls mutant diaphragms to similar or higher levels of controls in the absence of CCh (Figure 5D-5I). Nevertheless, mEPP frequency and amplitude in the presence of CCh were lower in mutant, compared with control at age of P30. At P1, mEPP amplitudes in the absence of CCh were similar between HB9-Wls-/- and control muscles (Figure 5H). These results indicate that transmission impairment in mutant mice could be attenuated by ACh agonist.

5) The defect in synaptic vesicle aggregation is interesting. The authors tried to show increased SV2 protein levels in the sciatic nerve to support the finding, but the Figure 6C is not convincing. Since the sciatic nerve is mixed with many non-motor axons, it would have been more convincing to use pure motor nerves, e.g. phrenic nerve.

Good suggestion. We now performed additional Western blot with phrenic nerves for SV2 and included another vesicle marker Synapsin 1. As shown in revised Figure 6C, both SV2 and Synapsin 1 were increased in phrenic nerves of mutant mice.

6) The nerve transplantation experiment is interesting, but it is less relevant to the conclusion. They can be completely removed.

We agree that this result is less relevant and has been completely removed, as suggested.

7) The partial rescue of NMJ defects via injecting individual Wnt proteins is interesting but the results do not directly support the conclusion on cell autonomous Wnt signaling. It is a little bit hard to interpret the results without knowing what relevant Wnt receptors are expressed in cells involved in NMJ formation or maintenance. For example, even Wnt7A expression/secretion from motor neurons is reduced, it cannot rule out that motor neuron-derived Wnt7A activates muscle (or Schwann cell) pathways to elicit retrograde signals to promote presynaptic development unless the receptor(s) for Wnt7A (or other relevant Wnt molecules) is selectively deleted from motor neurons, not other cell types.

How motoneuron Wnts regulate NMJ formation is a great question. They could act via activating Wnt receptors on motoneurons, on muscles and on Schwann cells. To address this question, we performed additional experiments to determine the expression patterns of Wnt receptors including 10 Frizzled proteins. As shown in revised Figure 1—figure supplement 1, many were detectable in the ventral horns of the spinal cord, Schwann cells, and muscle cells. Figuring out which Frizzled (plus Wnt co-receptors LPR5/LPR6) is involved in what cells could be a daunting task.

This paper provides first genetic evidence that Wnts from motoneurons are critical for NMJ development. Effects could be mediated by Wnt receptors on motoneurons, Schwann cells and/or muscle fibers, as the reviewer pointed out. Considering that our paper has 7 figures and 6 supplemental figures, each with multiple panels, we would hope that the reviewer would agree that how motoneuron Wnts act could be addressed in a future study.

We deleted Wls in the motoneuron and found the defects in the motoneuron, which we believe it is the cell-autonomous effect of Wls. But we agree that we don’t know whether it is through a cell-autonomous molecular mechanism without knowing the receptors. To get rid of the confusion, we have avoided using the term – cell autonomous – in the revised manuscript and also changed the paper title. We thank the reviewer for this question.

Reviewer #3:

Here, Shen et al. report a very interesting study on a novel role for Wnt in presynaptic assembly in the neuromuscular junction. The authors used the main Wnt secretion mediator Wls to specifically abolish Wnt secretion in each of the three NMJ compartments (skeletal muscle, motor neurons and Schwann cells). They showed that only Wls deletion in motor neurons leads to age-dependent NMJ deficits and neurotransmission impairment in a cell-autonomous manner. Using a combination of morphological analysis, electron microscopy and electrophysiology, they characterized the NMJ deficits and nicely demonstrate that motoneuronal specific deletion of Wls results in progressive NMJ degeneration, muscle weakness, impaired neurotransmission, motor axonal swelling and synaptic vesicle reduction. They also showed that motoneuronal Wls is required for NMJ regeneration. From there, they screened for the Wnt ligands that are expressed in motoneurons and identified those whose secretion is regulated by Wls in heterologous HEK cells. Wnt5A and 7A were the most sensitive to Wls mutation. Injection of recombinant Wnt7A, but not Wnt5A in TA muscle can partially rescue the morphological presynaptic deficits of HB9-Wl-/-. Overall this is a very interesting study, the analysis of data was carefully performed and most of the main conclusions are supported by the provided results. However, although the finding that Wnt7A may play a role in presynaptic assembly is novel, the current study does not provide sufficient evidence to support this conclusion and it remains unclear how Wnts secretion from motoneurons contributes to presynaptic functions resulting in age-dependent NMJ dysfunction when altered.

We thank the reviewer for the remarks that “Shen et al., report a very interesting study on a novel role for Wnt in presynaptic assembly in the neuromuscular junction” and that “Overall this is a very interesting study, the analysis of data was carefully performed and most of the main conclusions are supported by the provided results” and his or her constructive critiques and suggestions that have significantly improved the manuscript.

The authors state that Wls deletion in motoneurons, but not skeletal muscle or Schwann cells shows NMJ defects. Given that several Wnts identified by the authors in motoneurons are known to regulate postsynaptic differentiation, some of them being low or not sensitive to Wls, there is a possibility that Wnts secreted by the motoneurons may have compensated the muscle and/or Schwann cells NMJ phenotypes. A more detailed analysis of the NMJ phenotypes during NMJ formation, including combined HSA and/or HB9 and/or Wnt1-Wls-/- should be performed to support their conclusions. I realized that it is a demanding task, but this is not unfeasible.

The reviewer wanted us to determine whether motoneuron Wnts could compensate Wls deletion in muscles or Schwann cells. To test whether motoneuron Wnts could compensate Wls deletion in muscles, we generated mice where Wls was knocked out in both motoneurons and muscles (HB9/HSA-Wls-/-). As shown in revised Figure 2—figure supplement 3A and 3B, NMJ morphology in these mice was similar to that of HSA-Wls-/- mice, suggesting that motoneuron Wnts had little compensating effects on Wls deletion in muscles.

Along the same line, to determine whether motoneuron Wnts compensate Wls deletion in Schwann cells, we characterized HB9/Wnt1-Wls-/- mice where Wls was deleted in both motoneurons and Schwann cells. As shown in revised Figure 2—figure supplement 3C and 3D, NMJ morphology in these mice was also similar to that of Wnt1-Wls-/- mice, suggesting that motoneuron Wls could not compensate Wls deletion in Schwann cell.

We have tried hard to obtain triple-cell knockout mice (i.e., Wls deletion in motoneurons, muscles and Schwann cells). However, because Wnt1-Wls-/- mice die at birth, it has been very difficult to generate sufficient numbers of HB9/HSA/Wnt1-Wls-/- mice.

In newly performed experiments, we checked expression of 19 Wnts in motoneurons, Schwann cells, and muscle cells. Noticeably, Wnt7A appeared to be the most abundant, followed by Wnt3, Wnt4, and Wnt7B in motoneurons. Expression of other Wnts appeared to be low in motoneurons. Dominant Wnts in Schwann cells appeared to be Wnt5A and Wnt9A whereas Wnt4, Wnt6 and Wnt9A were most abundant in muscle cells (revised Figure 1). By knocking out Wls in motoneurons, we provide first genetic evidence that Wnts from motoneurons are critical for NMJ. In revision, we avoided implications that Wnts from muscles or Schwann cells are not important. The revision has discussed these points in revised Discussion section.

The authors identified several Wnts molecules expressed in motoneurons and showed that apart from Wnt7A and Wnt5A, several other (Wnt7B and Wnt9B) were sensitive to Wls. It is not clear why these Wnts were not tested in the rescue experiment. Also, how Wnt overexpression in the muscle can rescue the presynaptic NMJ deficits? Neurotransmission analysis could be performed following Wnt7A injection. The authors previously reported that Wnt7A modulates agrin-induced AChR clustering in muscle cell (Barik et al., 2014) suggesting that Wnt7A regulates postsynaptic differentiation. This should be included and commented in the manuscript.

These are good questions/suggestions.

1) We performed new experiments and demonstrated that Wnt7B, like Wnt7A, can also rescue; but Wnt9B, like Wnt5A, cannot rescue the defects in HB9-Wls-/- mice (revised Figure 7C and 7D).

2) “How Wnt overexpression in the muscle can rescue the presynaptic NMJ deficits?”

- Because Wls was mutated in motoneurons in HB9-Wls-/- mice, motoneurons are unable to process Wnt proteins for secretion. Therefore, we provided Wnts in muscles.

Exactly how motoneuron Wnts regulate NMJ development is a great question. They could act via activating Wnt receptors on motoneurons, on muscles and on Schwann cells. We performed additional experiments to determine the expression patterns of Wnt receptors including 10 Frizzled proteins. As shown in revised Figure 1—figure supplement 1, many were detectable in the ventral horns of the spinal cord, Schwann cells, and muscle cells. Figuring out which Frizzled (plus Wnt co-receptors LPR5/LPR6) is involved in what cells could be a daunting task.

This paper provides first genetic evidence that Wnts from motoneurons are critical for NMJ maintenance. Without Wnts, synaptic vesicles were deposited ectopically and reduced at nerve terminals, and these dysfunctional synapses were gradually degenerated with a feature of axonal swelling. The effects could be mediated by Wnt receptors on motoneurons, Schwann cells and/or muscle fibers. Considering that the paper has packed 7 figures and 6 supplemental figures, each with multiple panels, we would hope that the reviewer would agree that how motoneuron Wnts act could be addressed in a future study. We have included this in the revised Discussion section.

3) “Neurotransmission analysis could be performed following Wnt7A injection.”

- We performed new experiments and found that impaired neurotransmission in the HB9-Wls-/- was rescued by Wnt7A injection (revised Figure 7E).

4) The authors previously reported that Wnt7A modulates agrin-induced AChR clustering in muscle cell (Barik et al., 2014) suggesting that Wnt7A regulates postsynaptic differentiation. This should be included and commented in the manuscript.

- Yes, a previous in vitro study of ours showed that Wnt7A could reduce the number of agrin-induced AChR clusters in C2C12 myotubes. This result could suggest that Wnt7A may act by regulating postsynaptic differentiation. However, we did not observe a change in AChR cluster size or intensity in mice after Wnt7A injection (data not shown) although this reduced the number of axonal swellings. This suggests that Wnt7A could act directly on nerve terminals in the HB9-Wls-/- mutant. The revised manuscript discussed these possibilities (Discussion section).

Developmental, post-natal stages and muscle types used should be mentioned in the figure legends.

Sorry for the oversight. Such information has now been provided in revision in figure legends.

In several figures the title is misleading. The authors should replace Wl-/- by HB9-Wl-/-. Figure 5 and Figure 3—figure supplement 1: Please clarify the title.

We have gone through figures and text to ensure HB9-Wls-/- is used.

Figure 3—figure supplement 1: Please clarify how synapse elimination was evaluated.

To visualize AChR clusters on single muscle fibers, gastrocnemius was fixed in 4% PFA and stained with R-BTX and DAPI to show AChR clusters and myonuclei. Muscles were washed in PBS and teased into single fibers and mounted with Vectashield mounting medium. We counted the number of AChR cluster in each single muscle fiber and defined that synapse elimination was impaired when more than one AChR cluster was found after P15. Detailed description of synapse elimination was provided in the revision (Figure 3—figure supplement 1).

Figure 8D: The graph legend is missing.

The legend was provided in the revised Figure 7C.

Figure 1C: Co-staining of Wls with a motoneuron marker should be performed to support the conclusion.

Thanks for the suggestion. We performed new experiments to co-stain Wls with a motoneuron marker ChAT. Wls signal was diminished in the ventral horn of HB9-Wls-/- mice, compared with that in control mice. The new data (revised Figure 2C) were added in the revised Results section.

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    Transparent reporting form
    elife-34625-transrepform.docx (244.7KB, docx)
    DOI: 10.7554/eLife.34625.015

    Data Availability Statement

    All data generated or analysed during this study are included in the manuscript and supporting files.

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES