Gpr126/Adgrg6 contributes to the terminal Schwann cell response at the NMJ following peripheral nerve injury

Abstract

Gpr126/Adgrg6 is an adhesion G protein-coupled receptor essential for Schwann cell (SC) myelination with important contributions to repair after nerve crush injury. Despite critical functions in myelinating SCs, the role of Gpr126 within non-myelinating terminal Schwann cells (tSCs) at the neuromuscular junction (NMJ), is not known. tSCs have important functions in synaptic maintenance and reinnervation, and after injury tSCs extend cytoplasmic processes to guide regenerating axons to the denervated NMJ. In this study, we show that Gpr126 is expressed in tSCs, and that absence of Gpr126 in SCs (SC-specific Gpr126 knockout, cGpr126) results in a NMJ maintenance defect in the hindlimbs of aged mice, but not in young adult mice. After nerve transection and repair, cGpr126 mice display delayed NMJ reinnervation, altered tSC morphology with decreased S100β expression, and reduced tSC cytoplasmic process extensions. The immune response promoting reinnervation at the NMJ following nerve injury is also altered with decreased macrophage infiltration, Tnfα, and anomalous cytokine expression compared to NMJs of control mice. In addition, Vegfa expression is decreased in muscle, suggesting that cGpr126 non-cell autonomously modulates angiogenesis after nerve injury. In sum, cGpr126 mice demonstrated delayed NMJ reinnervation and decreased muscle mass following nerve transection and repair compared to control littermates. The integral function of Gpr126 in tSCs at the NMJ provides the framework for new therapeutic targets for neuromuscular disease.

Graphical abstract

• Gpr126 is expressed in terminal Schwann cells

• Terminal Schwann cell cytoplasmic processes are reduced in Schwann cell-specific Gpr126 knockout mice after injury

• Gpr126 is integral to terminal Schwann cell response to and NMJ recovery after nerve injury.

INTRODUCTION:

Peripheral nerve injuries are common and debilitating. The peripheral nervous system (PNS) is unique in its ability to regenerate after nerve injury, but motor outcomes decline with time to muscle reinnervation. Clinically, after a critical window of 1-2 years, muscle reinnervation cannot occur. Defining the molecular mechanisms that guide recovery after nerve injury is essential for future therapeutic considerations to improve functional outcome. In the PNS, Schwann cells (SCs) are a major glial cell type. SCs include myelinating SCs, which insulate axons with myelin to speed action potential propagation, and several populations of non-myelinating SCs. One type of non-myelinating SC, terminal SCs (tSCs), maintain and repair the neuromuscular junction (NMJ) in the muscle. Despite advancements in knowledge of regenerative PNS processes, few patients achieve normal function following nerve injury (Noble, Munro, Prasad, & Midha, 1998; Scholz et al., 2009). Investigations of the mechanisms supporting reinnervation are warranted to build upon the wealth of information providing translational benefit, including delineation of tSC contributions to reinnervation after nerve injury.

Perisynaptic tSCs reside at the NMJ and contribute to synaptic development, function, maintenance, reinnervation, and even NMJ disease. After injury, these multifunctional glial cells assume phagocytic roles and also extend long cytoplasmic processes in search of neighboring, innervated NMJs (Kang, Tian, Mikesh, Lichtman, & Thompson, 2014; O'Malley, Waran, & Balice-Gordon, 1999; Reynolds & Woolf, 1992; Son & Thompson, 1995a, 1995b). The tSC processes guide axons to the denervated NMJ and induce nearby axonal sprouting (Son & Thompson, 1995a, 1995b; Son, Trachtenberg, & Thompson, 1996). tSCs thus determine NMJ structural changes, including motor unit configuration. Relatively little, however, is known about tSC signaling and the mechanisms by which tSCs contribute to NMJ reinnervation after injury. Given that all SCs have common precursors and the unique roles of tSCs independent of myelin, we reasoned that investigation of molecular signals known in myelinating SCs might provide insight to tSC biology and overall improve knowledge of motor recovery after nerve injury.

Functional roles and signaling molecules of myelinating SCs are better delineated than those of tSCs. One example is the role of Gpr126/Adgrg6 in myelinating SCs (Hamann et al., 2015; Mogha et al., 2013; Monk, Oshima, Jors, Heller, & Talbot, 2011). Gpr126 is an adhesion G protein-coupled receptor (aGPCR) that is integral for myelinating SC development and function. aGPCRs are characterized by their structure, including a seven-transmembrane domain and a long N-terminus. Like many aGPCRs, Gpr126 is activated via a tethered peptide agonist, termed the Stachel, liberated by an autocleavage event that splits the receptor into an N-terminal fragment and a C-terminal fragment (Arac et al., 2012; Demberg, Rothemund, Schoneberg, & Liebscher, 2015; Liebscher et al., 2014; Stoveken, Hajduczok, Xu, & Tall, 2015). Gpr126 has also been found to be required for remyelination as well as timely macrophage recruitment and axon regeneration following nerve crush injury (Mogha et al., 2016). Gpr126, thus, has important functions in myelinating SC development and response to injury, but its role in the non-myelinating tSCs has not been previously delineated.

Here, we analyze Gpr126 conditional mutant mice to examine the role of Gpr126 in tSCs following nerve transection and immediate repair. We found that Gpr126 is required for tSC process elongation for successful NMJ reinnervation in the muscle. We also show that Gpr126 deficiency in SCs delays the immune response in muscle, necessary for reinnervation, after nerve injury. Gpr126 signaling, therefore, is integral not only to myelinating SCs, but also to nonmyelinating tSC function at the NMJ and contributes to the critical temporal window at the nerve-muscle interface after peripheral nerve injury.

MATERIAL and METHODS:

Experimental Animals

All animals were housed in a central animal facility and were maintained pre- and postoperatively in strict accordance with the National Institutes of Health guidelines and according to protocols approved by the institutional animal research ethics committee (IACUC) at Washington University School of Medicine. Gpr126^−/− constitutive knock out (Gpr126-KO) mice and conditional Dhh^Cre;Gpr126^fl/fl (hereafter called cGpr126 mice) mutant mice have been previously described (Mogha et al., 2013; Mogha et al., 2016; Monk et al., 2011). To generate conditional Dhh^Cre;Gpr126^fl/fl mutants and their sibling controls (Gpr126^fl/fl, hereafter called CTR [control] mice), Dhh^Cre;Gpr126^fl/fl females were crossed to Gpr126^fl/fl males. Dhh^Cre;Gpr126^fl/fl mice expressed body trembling as described before (Mogha et al., 2013). In addition, Dhh^Cre mice (Jaegle et al., 2003) in which the Dhh gene drives Cre expression in Schwann cells were crossed to RosatdTomato (Rosa 26^tdTm/tdTm) reporter mice to generate Dhh^Cre; Rosa26^tdTmtdTm mice. Rosa 26^tdTm/tdTm mice were kindly provided by Dr. Stacey Rentschler (Washington University, St. Louis, MO). Wild-type (WT, a pure C57BL/6 background) and Gpr126^+/− mice were also bred in our laboratory. For all experiments, age-matched (3-4, 12, and 22 months of age) mice of both sexes were analyzed and compared littermate sibling controls.

Denervation of Sternomastoid Muscle

To examine the role of Gpr126 signaling at the NMJ after nerve injury, we chose the spinal accessory nerve (SAN) innervating the sternomastoid (SM) muscle. Young adult mutant and CTR mice (3-4 months of age, both sexes) were anesthetized with ketamine (50 mg/kg BW) and dexmedetomidine (1 mg/kg BW) and placed supine under a surgical microscope. After the appropriate depth of anesthesia, the neck was carefully shaved and cleansed with betadine scrub and alcohol pads. A vertical skin incision was made from the chin to the superior aspect of the sternum, and salivary glands and fat were retracted to allow access to the SAN deep to the SM. The left SAN was cut just lateral to its point of entry into the muscle (~6 mm away from the muscle). Immediately following nerve transection, Seprafilm was placed under the cut nerve ends to prevent retraction, and the nerve was repaired using fibrin glue (Tisseel, Baxter, Deerfield, IL, USA). For all mice, the right SAN was exposed but not injured as a sham uninjured control. Incisions were sutured closed, and anesthesia was reversed with atipamezole hydrochloride (1 mg/kg). Mice were administered pain-reducing medication, and they were recovered on a heating pad and returned to the central animal housing facility within 12 hours of the procedure for close monitoring and postoperative care.

Muscle Harvest

At 1, 3, or 6 weeks following nerve transection and repair, SM from both the injured and sham uninjured sides were harvested under the dissecting microscope. For immunostaining, SM muscles (n=5-6 mice/genotype) were weighed, immediately placed in cold 0.1 M phosphate buffered saline (PBS, pH 7.4), and then fixed in 4% paraformaldehyde as described before (Snyder-Warwick et al., 2018). Muscles were cryoprotected in 15% and 30% sucrose/PBS, frozen, and sectioned. For mRNA expression at the NMJ, SM from both uninjured and injured sites (n=3 animals/genotype) were collected. For each SM the endplate band (centrally located NMJ-dense region) and synapse-free segments (used as control) of SM muscles were dissected under the microscope and placed in RNALater solution (Thermo Fisher Sci.). Samples from individual mice were homogenized in Trizol (Invitrogen) and stored at −80°C until RNA extraction next day (see method below).

To examine NMJ morphology in homeostasis in aged (12 and 22 months of age) cGpr126 mutant mice and control siblings, SM and extensor digitorum longus (EDL) muscles were collected.

Immunofluorescent Staining of Neuromuscular Junctions

NMJs were immunofluorescently stained as previously described in detail (Snyder-Warwick, Satoh, Santosa, Imai, & Jablonka-Shariff, 2018). Briefly, frozen sections (25 μm thick) or whole mounts of the SM or EDL muscles were washed with PBS, incubated for 60 minutes in blocking buffer (containing 5% normal goat serum, 2% Triton-X 100, 5% BSA) and then incubated in primary antibodies overnight at 4° C. The following primary antibodies were used: rabbit anti-NF200 (1:500; Sigma, St. Louis, MO), rabbit anti-S100β (1:1000; Dako North America, Carpinteria, CA), mouse anti-Synaptophysin (1:500, Abcam), rabbit anti-Gpr126^CTF(a kind gift from Xianhua Piao, Boston Children’s Hospital/UCSF; 1:10) (Mogha et al., 2016; Petersen et al., 2015), rat anti-CD68 (1:500, BioRad, Hercules, CA) and anti-Ly6G (1:100, Clone 1A8, BioLegend, San Diego, CA)-. The next day, sections were washed in PBS and incubated with appropriate fluorescently labeled secondary antibodies (1:1000; Invitrogen) for 1 hour at room temperature. In addition, endplates were stained with α-Bungarotoxin-Alexa 594 or -Alexa 647 (1:1000, Life Technologies), which bind specifically to acetylcholine receptors (AChRs) in the postsynaptic membrane, thereby marking synaptic sites. Control sections (either no primary or no secondary antibody) were included in every type of staining to test for non-specific staining and autofluorescence. After the final washing, coverslips were mounted on glass slides using Vectashield with DAPI (Vector Labs, Burlingame, CA). NMJs were imaged with an Axio Imager M2 fluorescent microscope (Zeiss, Thornwood, NY, USA).

Morphometric quantification

Images were analyzed using NIH ImageJ (http://rsb.info.nih.gov/iJ/) (Tse et al., 2014), and figures were prepared using Adobe Photoshop CC 2018 and Adobe Illustrator CC 2018 system (Adobe Systems, San Jose, CA, USA). For quantitative analyses, muscle sections from 3-6 animals from each genotype were processed, and at least 15 random sections and totaling >100 NMJs lying en face were analyzed per muscle. For each time point, muscles from sham and injured sides of mutant and sibling controls were analyzed. All quantitative analyses were confirmed via 2 blinded evaluators. Morphologic changes after the nerve injury were analyzed in major NMJ components: the presynaptic nerve terminals (NF200 antibody labeling), tSCs (S100β antibody labeling), synaptic vesicles (Synaptophysin antibody labeling), postsynaptic AChRs on the motor endplate (α-BTX labeling), and nuclei (DAPI labeling).

To quantify innervation before and after nerve injury, NMJs showing colocalization of NF200 and α-BTX staining were counted. The evaluation was based on the extent of alignment between presynaptic motor nerve terminal and postsynaptic AChR clusters. An NMJ was defined as innervated if nerve and AChR clusters aligned over more than 75% of length. The percentage of innervated NMJs was normalized to the total number of NMJs in the field of view.

To determine the morphological changes in AChRs stained with α-BTX and DAPI, the number of discrete AChR fragments per endplate were counted (Jones et al., 2016). Endplates were considered “normal” if AChRs formed a continuous, “pretzel-like” structure with 7 or fewer fragments (Li & Thompson, 2011). The percentage of normal endplates are shown for uninjured mice and after nerve injury for sham and injured sides of mutant and CTR mice.

To examine tSCs at the NMJ before and after nerve injury, sections were stained with S100β antibody, α-BTX, and DAPI. The fluorescence intensity of S100β expression in tSC bodies was measured as the integrated density on grayscale images using NIH ImageJ program (Tse et al., 2014). The integrated density measures pixel intensity, and this analysis does not depend on the size of cells and nuclei. Images used to generate integrated density values were collected with identical microscopic settings. In addition, the percent of NMJs with tSC processes extending beyond the area of the NMJ were calculated following nerve injury.

Number of macrophages positively stained for CD68 antibody and DAPI were evaluated using NIH ImageJ (n=3 animals/genotype) and presented as number of cells per field of view. The immature neutrophils with ring-shaped nuclei positively stained for Ly6G antibody and DAPI were counted after the nerve injury and presented as percentage of total nuclei per field of view.

RNA Purification and RT-PCR

Total RNA was extracted from homogenized individual muscles using an acid phenol extraction separation with TRIzol (Invitrogen) and then treated with RNase-free DNase (Qiagen, Hilden, Germany) to remove residual DNA. The RNA pellet was then further purified using RNeasy MinElute Cleanup Kit (Qiagen) according to the manufacturer’s protocols and quantified with a NanoDrop ND-1000 spectrophotometer. Only those samples with a 260 nm / 280 nm ratio between 1.6–2.1 were used. Complementary DNA was synthesized from total RNA (100 μg) using High-Capacity cDNA Transverse Transcription Kit (Applied Biosystems, Foster City, CA). Samples without reverse transcriptase (RT) enzyme, to control for genomic DNA contamination, were also included. RNA was extracted from 3 biological replicates for all genotypes and sibling controls (n=3 animals/genotype).

mRNA Expression

Quantitative real-time PCR (qPCR) performed using TaqMan Fast Universal PCR Master mix and specific TaqMan PCR primers-probes combination were purchased from Applied Biosystems (Foster City). The following Taq-Man Gene Expression Assays were used: S100β (Mm00485897_m1), Ccl2 (Mm00441242_m1), Tnfα (Mm00443258_m1), Interleukin-6 (Mm00446190_m1), Interleukin-10 (Mm01288386_m1), Interleukin-12a (Mm00434169_m1), Vegfa (Mm00437306_m1) and Gapdh (Mm99999915_g1). All qPCR studies were performed on a Step One Plus instrument (Applied Biosystems, Foster City), and results were analyzed using Microsoft Excel and ΔΔCt method (Livak & Schmittgen, 2001). Relative expression levels were calculated for each gene by normalizing first to Gapdh level and then to the expression in sham uninjured tissue for each genotype of mice. Relative gene expression in SM from uninjured sides were designated as 1, and results for injured groups (cGpr126 and CTR) are shown as fold change compared to the uninjured site. Each sample was run in triplicate on each plate, and a ‘non-template control’ with water and no RT enzyme were always carried out as negative controls.

Statistical analyses

All analyses were done using Microsoft Excel or GraphPad Prism 7.00 (San Diego California, http://www.graphpad.com). Data are shown as mean ± SD. Significance is shown as *p < 0.05. Quantifications were performed from at least three experimental groups unless otherwise noted. No animals or data points were excluded from analyses. Image analyses were quantified by 2 independent investigators. qRT-PCR reactions were performed in technical and biological triplicate. Count data were assumed to be non-parametric, and appropriate statistical tests were used. Statistical analyses were performed by two-tailed unpaired Student's t tests and Mann-Whitney U test.

RESULTS:

Gpr126 is present in non-myelinating tSCs

Gpr126 expression in tSCs has not previously been reported. To initially investigate the role of Gpr126 in tSC function, we first examined the localization of Gpr126 in non-myelinating tSCs at the NMJ using WT mice. Gpr126 protein was expressed in tSCs at the NMJ in the sternomastoid (SM) muscle (Fig. 1A, A’ B). Absence of Gpr126 staining was confirmed in both Gpr126^−/− (not shown) and cGpr126 mice (Fig. 1C).

Figure 1.

Gpr126 is expressed in tSCs at the NMJ in SM muscle. (A, B) Images reveal Gpr126 (red) localization at NMJs (white arrows) and in myelinating SCs (yellow arrows) of WT mice. (A’) The isolated red channel from panel A, shown in gray, highlights distribution of Gpr126 staining at NMJs, which colocalizes with tSCs (green). (C) No Gpr126 staining is present at NMJ in conditional knockout Gpr126 (cGpr126) mice. Inset shows motor endplate (asterisks) staining with α-bungarotoxin, (BTX, red). S100β = glial cells (S100, green), DAPI = nuclear staining (blue). Scale bar = 20 μm.

Constitutive and conditional Gpr126 transgenic mouse models

We showed previously (Monk et al., 2011) that constitutive Gpr126^−/− mutant mice have early lethality and display profound motor deficits. Although Gpr126^−/− mutants frequently die in utero, we were able to examine Gpr126^−/− pups from P1 (n=3), P2 (n=10) and P12 (n=1) and their Gpr126^+/− littermate controls (n=3-10 animals/age). Gpr126^−/− mice weighed less than controls (Table 1), as previously described (Monk et al., 2011). Immunohistochemical examination of endplates showed more immature acetylcholine receptor (AChR) morphology at the NMJs of Gpr126^−/− mice compared to controls (Fig. 2). At P2, for example, AChRs had an ovoid plaque morphology (Fig. 2B) compared to the perforated-plaque phenotype of controls (Fig. 2A), and control AChR areas were noted to be, on average, 1.8-fold larger. No differences were noted in endplate morphology of the single P12 Gpr126^−/− mouse. These data show delayed NMJ development in the absence of Gpr126, suggesting a contributory, but not obligatory, role for Gpr126 in NMJ maturation.

Table 1.

Body weight of Gpr126^+/− and Gpr126^−/− mice.

Genotype of Mice	Age
	P1	P2	P12
Gpr126+/−	1.62 ± 0.05 (3)	1.92 ± 0.31(10)	8.68 ± 0.11 (3)
Gpr126−/−	1.09 ± 0.02 (3)*	1.39 ± 0.10 (10)*	3.53 (1)

Data presented as mean ± SD. Numbers in parenthesis indicate number of mice.

^*Indicates significance for P1 and P2 (p<0.0001) compared to Gpr126^+/− mice.

Figure 2.

Motor endplates from P2 Gpr126^−/− mice are less mature compared to Gpr126^+/− controls. (A) Representative image from control Gpr126^+/− mice showing perforated-plaque endplate morphology (arrows; yellow arrow highlights higher magnification inset). (B) Image from Gpr126^−/− mice demonstrates more immature aneural/plaque phenotype of endplates (arrowheads, yellow arrowhead highlights higher magnification inset). BTX = α-bungarotoxin (endplates, red), and DAPI = nuclear staining (blue). Scale bar = 20 μm.

Due to the short lifespan of the constitutive Gpr126^−/− mouse, we were unable to assess NMJ response to nerve injury. We therefore investigated a conditional model to assess Gpr126 function. Conditional Dhh^Cre; Gpr126^fl/fl (cGpr126) mutants, in contrast to Gpr126^−/− mutants, are much healthier and exhibit normal lifespans. cGpr126 mice are ambulatory, although they exhibit a body tremor (Mogha et al. 2013). These mice express Cre recombinase under control of the Desert hedgehog (Dhh) promoter, which drives recombination in SC precursors at E12.5 (Jaegle et al., 2003). To test if Dhh drives recombination in tSCs, we crossed Dhh^Cre mice with Rosa26^tdTm/tdTm reporter mice. We evaluated NMJs of young adult Dhh^Cre; Rosa26^tdTm/tdTm mice (n=8 mice from 3 different litters) for the presence of the tdTomato fluorescence at the NMJ, indicating Cre-recombinase driven by Dhh within tSCs (Fig. 3A-C). Rosa26:tdTomato fluorescence was present in 95±8.6% of tSCs from Dhh^Cre; Rosa26^tdTm/tdTm mice and was absent in Rosa26 tdTomato mice (Fig. 3A inset), suggesting efficient Cre-mediated recombination in tSCs.

Figure 3.

Desert hedgehog (Dhh) is expressed in tSCs. (A-C) DhhCre expression in tSCs (white arrows) at NMJs and in myelinating glia along axon (blue arrows) as shown by the red tdTomato co-localization with the glial marker, S100β antibody (green). Inset in panel A shows no tdTomato fluorescence but only some autofluorescence in the muscle fibers of Rosa26tdTomato mice. Asterisk in panel C highlights a tSC with low tdTomato expression level. Images shown are representative of 8 transgenic animals analyzed from 3 independent litters. S100β = glial cells (S100, green), DAPI = nuclear staining (blue). Scale bar = 20 μm.

Young cGpr126 mice have normal NMJ architecture, but decreased tSC expression of S100β

SM muscle synaptic morphology from young adult (3-4 month old) cGpr126 mice showed similar NMJ structure to CTR mice (Fig. 4). Nerve terminals (NF200), synaptic vesicles (synaptophysin), and AChR (αBTX) morphologies did not differ from controls (Fig. 4B-C, E-F). There were no differences in mean tSC number per NMJ (2.3 ± 0.6 tSCs per NMJ in cGpr126 mice versus 2.5 ± 0.5 in CTR). In contrast, S100β staining intensity of tSCs was reduced in cGpr126 mice compared to controls (16.3 ± 5.3 compared to 40.3 ± 10.7, p<0.05 in CTR mice) (Fig. 4A, D, G). This decreased S100β staining intensity was not seen along the nerve in the same mice, but was unique to tSCs at the NMJ.

Figure 4.

tSCs are abnormal in the absence of Gpr126 in SM muscles from 3-month old cGpr126 compared to CTR mice. (A-C) Representative images of NMJs from CTR mice showing distinct tSCs (A, arrows), nerve terminals (B, arrows) and synaptic vesicles (C, arrowhead). (D-F) Images of NMJ components from cGpr126 mice. Note that tSCs (D, arrows) show lower fluorescence intensity (quantified as integrated density/area) after staining with S100β antibody, a glial cell marker, compared to CTR mice. Insets represent the isolated green channel from panels A-B, D-E. (G-J) Quantification of NMJ components in CTR and cGpr126 mice. Endplates were considered “normal” if AChRs formed a continuous, “pretzel-like” structure with 7 or fewer fragments. S100β = glial cells (green), NF200 = anti-neurofilament antibody (axons, green), BTX = α-bungarotoxin (endplates, red), and DAPI = nuclear staining (blue). Scale bar = 20 μm. Data: Mean ± SD; * p < 0.05.

cGpr126 mice have a NMJ maintenance defect in the hindlimb with age

cGpr126 mice displayed gross disturbances in hindlimb function with age. By 12 months, most transgenic mice displayed hindlimb paresis or paralysis, which was not seen in CTR sibling mice. Histological examination of the SM and EDL muscles revealed NMJ structural changes in EDL of cGpr126 mice, including extensive denervation and AChR fragmentation (Figs. 5B, F; 6A). At 12 months of age, only 32.6 ± 10.6% of NMJs were innervated in the EDL muscles of cGpr126 mice compared to 93.8 ± 11.6% of those in CTRs (p<0.001; Fig. 6A). In contrast to EDL, NMJs in the SM muscle in the neck of cGpr126 mice were no different from those of CTR littermates at 12 months of age (Fig. 5A, E and Fig. 6A). As expected, NMJ alterations were noted in the NMJs of the SM in cGpr126 mice at 22 months of age where both EDL and SM showed typical age-related synaptic degeneration (Figs. 5C-D, G-H; 6A). These changes did not differ between cGpr126 mice and CTRs. In addition, loss of S100β immunoreactivity in tSCs at the NMJ was noted in both SM and EDL muscles for aged cGpr126 mice compared to CTR mice (Figs. 6B and 7). For cGpr126 mice, tSC bodies were only clearly observed in SM muscle at 12 months of age, but the relative S100β integrated density was less than half that of controls (Figs. 6B and 7). Moreover, the morphology and the integrated density of S100β in EDL muscles cGpr126 mice at age 12 months was similar to changes observed in 22 month-old CTR mice, consistent with long-term SC disruption observed in sciatic nerve of cGpr126 mutants (Kuffer et al., 2016).

Figure 5.

Absence of Gpr126 in SCs results in an NMJ maintenance defect in EDL muscle of the hindlimb but not in SM muscle of the neck. NMJs from 12 month old CTR mice are fully innervated; nerve terminals (green, arrows) overlap with pretzel-shaped BTX (red) in both SM (A) and EDL (B) muscles, but they show denervation and endplate fragmentation at 22 months (C, D). cGpr126 mice show NMJ denervation and endplate fragmentation (short arrow) in EDL muscle already at 12 months of age (F), but not in SM muscle of the neck (E). NMJs from SM and EDL muscles of 22 month old cGpr126 are denervated (G, H) similar to CTR mice. NF200 = anti-neurofilament antibody (axons, green), BTX = α-bungarotoxin, (endplates, red), and DAPI = nuclear staining (blue). Scale bar = 20 μm.

Figure 6.

Quantification of NMJ changes in SM and EDL muscle (shown in Figures 5 and 7) of aged CTR and cGpr126 mice. (A) Percent of innervated NMJs from 12- and 22-month old CTR and cGpr126 mice. NMJs were defined as innervated if nerve and an AChR cluster with 7 or fewer fragments (designated as normal endplates) aligned over more than 75% of its length. The percentage of NMJs was normalized to the total number of NMJs in the field of view. (B) Bar graph showing the average pixel intensity for S100β protein in tSCs measured as integrated density/area using NIH ImageJ. Data: Mean ± SD; * p < 0.05.

Figure 7.

tSC changes in aged CTR and cGpr126 mice. tSCs (green, arrows) and their processes are present and overlapping with BTX at the NMJ in both the SM (A) and EDL (B) muscles of 12 month old CTR mice. tSCs are fewer in number and have less prominent cell bodies in the SM (C) and EDL (D) muscles of 22-month old CTR mice. In cGpr126 mice (E-H), tSC fluorescent integrated density is less than that of CTR mice at all assessed ages and locations. tSC absence is notable in cGpr126 EDL muscles (F), but not SM muscles (E), at 12 months of age. tSCs are frequently absent at the NMJ (short arrow) in cGpr126 mice by 22 months of age in both the SM (G) and EDL (H). Note, the quantifications are shown in Figure 6B. S100β = glial cells (S100, green), BTX= α-bungarotoxin (endplates, red), and DAPI = nuclear staining (blue). Scale bar = 20 μm.

The impact of Gpr126 deletion at the NMJ was therefore dependent on anatomic location, suggesting the importance of target muscle distance from the neuron cell body, perhaps related to the impact of poor myelin insulation along the nerve and diminished action potential conduction.

NMJ reinnervation is delayed in cGpr126 mice after nerve injury

To further dissect the function of Gpr126 in tSCs, we examined its necessity in NMJ reinnervation following spinal accessory nerve transection and immediate repair. NMJ reinnervation in SM, based on the extent of alignment between presynaptic motor nerve terminals and postsynaptic AChR clusters, was decreased in the cGpr126 mice compared to injured CTR mice at all assessed time points (Fig. 8). Fewer (44.2 ± 14.7%) NMJs were reinnervated in mutant mice 1 week after nerve cut and repair compared to CTR (71.3 ± 10.1%, p<0.006) mice. By 3 weeks, only 59.1 ± 12.3% of NMJs were reinnervated in cGpr126 mice compared to 92 ± 9.8% (p<0.003) in CTR mice. By 6 weeks, mean reinnervation in cGpr126 mice was still low at 65.3 ± 12.4% compared to 96.1 ± 3.5% (p<0.03) in CTRs.

Figure 8.

NMJ reinnervation is decreased after nerve injury in the absence of Gpr126 in SCs. (A-C) Representative images of NMJs from SM muscles of 3-4-month old CTR mice after spinal accessory nerve transection and immediate repair at 1, 3, and 6 weeks post-injury. NMJs show full reinnervation (green, arrows) and overlapping with pretzel-shaped endplates (red) at 3 and 6 weeks-post injury. (D-F) Images from cGpr126 mice reveal significantly fewer innervated NMJs (arrows) with many small denervated endplates (red, asterisks) even at 6 weeks after injury. Insets show NMJs at higher magnification (G, H) Quantification of NMJ innervation and percentage of normal endplates are summarized (n≥3 mice/genotype/per time point). NMJs were defined as innervated if nerve and an AChR cluster with 7 or fewer fragments (designated as “normal” endplates) aligned over more than 75% of its length. The percentage of NMJs was normalized to the total number of NMJs in the field of view. NF200 Ab = anti-neurofilament antibody (axons, green), BTX = α-bungarotoxin (endplates, red), and DAPI (nuclear staining, blue). Scale bar = 20 μm. Data: Mean ± SD; * p < 0.05.

In addition to NMJ reinnervation, we assessed specific AChR morphology. AChR morphology (αBTX) did not differ statistically between cGpr126 and CTR mice 1 and 3 weeks following nerve injury and repair (Fig. 8). At 6 weeks, however, in cGpr126 mice only 67.7 ± 18.5% of AChRs had normal morphology compared to 98.3 ± 2.9% (p<0.04) of injured CTR mice (Fig. 8H).

Gross SM muscle weight was assessed following spinal accessory nerve cut and repair (Table 2). The injured hindlimb of cGpr126 mice had lower SM muscle mass at all time points assessed compared to CTR mice, and these differences were significant at 3 weeks (p<0.0002) and 6 weeks (p<0.0004). In cGpr126 mice, SM mass dipped to 82.1 ± 3.8% of the sham side at 1 week, reached a nadir of 67.4 ± 8.4% at 3 weeks, and weighed 72.2 ± 7.6% of the uninjured side at 6 weeks. The SM muscle mass of CTR mice varied from 88.1 ± 4.7% to 92.6 ± 6.9% of the uninjured side throughout the study time points. Muscle mass is reflective of the innervation differences between the two groups.

Table 2.

Change (%) in SM muscle weight of control (CTR) and cGpr126 mice after the nerve injury compared to uninjured side.

Genotype of Mice	Weeks Post-Injury
	1 week	3 weeks	6 weeks
CTR	91.4 ± 7.2%	88.1 ± 4.7%	92.6 ± 6.9%
cGpr126	82.1 ± 3.8%	67.4 ± 8.4%*	72.2 ± 7.6%*

For each mouse SM muscle weight from injured side was compared to uninjured side (set as 100%). Data presented as Mean ± SD (N=4-8 mice/genotype/time point).

^*Indicates significance for 3 weeks (p<0.0002) and 6 weeks (p<0.0004) after injury compared to CTR mice.

The tSC response to nerve injury is altered in cGpr126 mice

tSCs encapsulate the NMJ and play important roles in repair. Post-denervation/reinnervation loss of S100β immunoreactivity at the NMJ was noted in cGpr126 mice (Fig. 9D-F) compared to CTR mice (Fig. 9A-C) at 1, 3, and 6 weeks. Relative S100βintegrated density in tSC bodies was less than half that of injured controls at all time points (1, 3, and 6 weeks; p<0.0001) following injury (Fig. 9G). Such low S100β immunoreactivity was also observed in tSCs of uninjured cGpr126 mice compared to CTRs (Fig. 4G). In addition, S100β mRNA expression near NMJs corroborated staining data (Fig. 9H). For both cGpr126 and CTR mice, results represent fold change in mRNA expression compared to the uninjured side. In cGpr126 mice, S100β mRNA expression was less than half that of CTRs at 1 week after injury (p<0.02), less than one-fourth of CTRs at 3 weeks (p<0.00001), and less than one-ninth of expression in CTRs at 6 weeks (p<0.0001).

Figure 9.

tSCs have less S100β expression after nerve injury in cGpr126 mice. (A-C) Representative images show S100β fluorescent staining of tSCs (green, arrows) at NMJs from SM muscle of CTR mice 1, 3, and 6 weeks after nerve injury. S100β staining is also shown in myelinating glia along axons. (D-F) Images from cGpr126 mice demonstrate lower intensity of S100β labeling in tSCs (arrows) at NMJs at all experimental time points. Insets show higher magnification of tSCs (asterisks). (G) Bar graph showing the average pixel intensity for S100β protein measured as integrated density/area. This analysis does not depend on cellular or nuclear size. (H) Relative expression of S100β mRNA is significantly reduced at the NMJ in cGpr126 relative to CTR mice (n=3 mice/genotype/time point). For both types of mice, gene expression from the uninjured sides were set as 1, and results for the injured sides are shown as mRNA fold change. S100β = glial cells (S100, green), BTX = α-bungarotoxin (endplates, red), and DAPI = nuclear staining (blue). Data: Mean ± SD; * p < 0.05.

Following nerve injury, tSCs normally extend, in advance of axon sprouts, cytoplasmic processes beyond the denervated NMJ in search of neighboring, innervated NMJs. NMJ reinnervation results in withdrawal of tSC processes (Kang & Lichtman, 2013). In cGpr126 mice, the percentage of tSCs extending processes was less than that of injured CTR mice at all assessed time points, but differed significantly at 3 weeks after nerve transection and repair: 2.0 ± 3.4% cGpr126 vs. 22.3 ± 18.6% CTR at 3 weeks (p<0.05, Fig. 10). Of note, no tSC processes were observed in the absence of injury (Fig. 4A, D). These results suggest Gpr126 in tSCs plays an important role in tSC morphology and their response to nerve injury, likely affecting synapse remodeling post-injury.

Figure 10.

Lack of Gpr126 impairs the extension of tSC cytoplasmic processes after nerve injury to initiate reinnervation. (A, B) NMJs from CTR mice 3 and 6 weeks after nerve injury show tSCs (asterisks) extending long processes beyond the NMJ area (arrows). (C, D) tSCs from cGpr126 stained with S100β lack long processes after nerve injury. Some tSCs form thin cytoplasmic whiskers with low intensity S100β antibody staining (arrows). (E) Bar graph showing the percentage of NMJs with tSC processes beyond the NMJ area after nerve injury. S100β = glial cells (green), BTX = α-bungarotoxin (endplates, red), and DAPI = nuclear staining (blue). Data: Mean ± SD; * p < 0.05.

cGpr126 mice have altered immune response at the NMJ after nerve injury

The immune response has been shown to be integral in sciatic nerve repair after injury (Cattin et al., 2015; Mogha et al., 2016). We assessed whether the immune response was integral at skeletal muscle for NMJ reinnervation. In the current study, we observed an immune response active at the NMJ in the end target muscle after nerve injury in CTR mice, and this immune response is altered in cGpr126 mutant mice (Figs. 11-12). Macrophages are early responders following injury. In SM muscle of CTR mice, the highest number of CD68+ macrophages was noted 1 week following injury, an expected time point given the acute time course of macrophages, and then dramatically decreased by 3 weeks and 6 weeks after injury (Fig. 11A-C, G). In cGpr126 mice, significantly fewer macrophages were present at 1 week (14.6 ± 0.5 per view area) compared to CTR mice (41.4 ± 11.7 per view area, p<0.0001). The number of CD68+ macrophages remained consistent in cGpr126 mice at 3 weeks (9.5 ± 5.8 per view area) and 6 weeks (10.4 ± 3.4 p<0.004) after injury, while CD68+ cells in CTR mice were much lower at these time points (Fig. 11D-G).

Figure 11.

Fewer macrophages and altered chemokine expression are observed in the SM muscle of cGpr126 mice following spinal accessory nerve transection and immediate repair. (A-C) Representative images from CTR mice showing the highest number of CD68+ macrophages at 1 week (arrows) and only few small CD68+ cells (arrows) were observed at 6 weeks following nerve injury. (D-F) Images from cGpr126 mice reveal fewer CD68+ macrophages at 1 week and the same number stayed until 6 weeks (arrows). (G) Quantification indicates significantly fewer macrophages in cGpr126 compared to CTR mice at 1 week post injury. (H, I) Relative CCL2 and TNF chemokine expressions (n=3 mice/genotype/time point) are significantly altered in cGpr126 mice relative to CTR mice at 1 and 3 weeks following injury. For both types of mice, the gene expression from uninjured contralateral hindlimb were set as 1, and results for injured sides are shown as mRNA fold change. CD68+ = macrophages (red) and DAPI = nuclear staining (blue). Data: Mean ± SD; * p < 0.05.

Figure 12.

NMJ areas from cGpr126 mice show infiltration of immature neutrophils with ring-shaped nuclei 3 weeks following nerve injury. These cells accumulate around NMJs from cGpr126 mice (B, arrows) compared with few cells in CTR mice (A, arrow) mostly found away from NMJs. Inset in panel B shows higher magnification of ring-shaped nuclei (gray, arrow). (C) Ly6G antibody, a marker for immature neutrophils, stains cells (purple, arrows) with ring-shaped nuclei. Inset in panel C shows higher magnification of immature neutrophil. (D) Quantitative analysis showing the significant increase in number of immature neutrophils in cGpr126 mice 3 weeks post injury compared with CTR mice. DAPI = nuclear staining (blue in panels A and B or gray in panel C). Data: Mean ± SD; * p < 0.05.

To further evaluate these macrophage differences, we assessed mRNA expression of some pro-inflammatory-associated chemokines (Fig. 11H, I). Ccl2 binds to the macrophage receptor, Ccr2 (Kwon, Yoon, & Kim, 2016). For CTR mice, the highest level of Ccl2 was observed at 1 week and remained low at 3 and 6 weeks after the nerve injury. Thus, the level of Ccl2 corresponded to the number of CD68+ cells. In contrast, Ccl2 mRNA expression was higher in cGpr126 mice than in CTR mice at 1 week following injury, relative to uninjured sides (10.3 and 7.3-fold change, respectively, Fig. 11H). The level remained higher at 3 weeks compared to CTR mice and dropped dramatically by 6 weeks. These data indicate that other immune cells, besides macrophages, contributed to the elevated expression of Ccl2 in cGpr126 mice. TNFα is a first-responder to injury and recruits macrophages. In NMJs of CTR mice, Tnfα mRNA expression was significantly increased by nearly 12-fold compared to uninjured sides at 1 week following nerve injury and repair. Tnfα expression then decreased to ~2.5-fold of baseline at 3 and 6 weeks. Although Tnfα mRNA expression in NMJs of cGpr126 mice increased after injury and repair, it was significantly lower than CTRs at 1 and 3 weeks (p<0.05, Fig. 11I).

In addition to changes in CD68+ macrophage response, different immune modulators were noted in the cGpr126 mice compared to CTR mice during the reinnervation period. At 3 weeks following injury, distinct, CD68− cells with ring-shaped nuclei were seen near NMJs in cGpr126 mice (16.5 ± 7%) and rarely at the NMJs of CTR (1.4 ± 0.9%) mice (p<0.001, Fig. 12). Cells with ring-shaped nuclei were not seen at the other time points. Ring-shaped nuclei are characteristic for immature neutrophils (Bliss, Butcher, & Denkers, 2000; Nathan, 2006). Ly6G staining (Fig. 12C) was used to identify immature neutrophils (Sagiv et al., 2015; Tsiganov et al., 2014). Expression levels of additional related cytokines and immune modulators were therefore assessed with qRT-PCR (Fig. 13). IL-6 mRNA expression was elevated at 1 and 3 weeks after injury in CTRs, but was significantly reduced (p<0.05) in cGpr126 mice at 3 weeks following nerve injury and repair (Fig. 13A). IL-10 expression remained near baseline levels in CTR mice after injury, but was 13-fold above baseline in cGpr126 mice at 1 week after injury (p < 0.05) and remained higher than CTR mice at both 3 and 6 weeks (p<0.05, Fig. 13B). For experimental groups, IL-12α mRNA expression after injury was lower than baseline in uninjured mice, but IL-12α expression in cGpr126 mice was significantly lower than in CTR mice (Fig. 13C). Overall, these results indicate that Gpr126 in tSCs is required for macrophage infiltration and balanced chemokine expression at the NMJ in the muscle during the reinnervation process.

Figure 13.

Cytokines (A-C) and Vascular Endothelial Growth Factor A (VEGFA, D) expressions are altered in synaptic regions of cGpr126 mice compared with CTR mice (n=3 mice/genotype/time point). For both genotypes, the relative gene expression in SM from the uninjured sides were set as 1, and results for injured sides are shown as mRNA fold change. Data: Mean ± SD; * p < 0.05.

Given that the immune response and Gpr126 have been implicated for angiogenesis supporting nerve regeneration (Cui et al., 2014; Shvartsman et al., 2014) and the aberrant changes noted in the immune response to injury in the cGpr126 mice, we also evaluated Vegfa mRNA expression levels. While Vegfa expression was unchanged from baseline in both genotypes at 1 and 3 weeks, cGpr126 Vegfa mRNA levels were less than one-third those of CTRs at 6 weeks after injury (p<0.05), a time point where Vegfa expression increased by 3-fold over baseline in uninjured mice (Fig. 13D).

DISCUSSION:

tSCs are essential to NMJ reinnervation following injury. After injury, tSC cytoplasmic processes induce axonal sprouting and provide trophic support and guidance to regenerating axons (Kang, Tian, & Thompson, 2003; O'Malley et al., 1999; Reynolds & Woolf, 1992; Son & Thompson, 1995a, 1995b; Son et al., 1996). Fast reinnervation helps synapse remodeling, whereas delayed reinnervation leads to structural abnormalities at the synapse. The molecular mechanisms mediating these events have not been elucidated.

The aGPCR Gpr126 is integral to myelinating SC development (Mogha et al., 2013; Monk et al., 2009; Petersen et al., 2015), maintenance (Kuffer et al., 2016), and function, as well as for repair after a crush injury (Mogha et al., 2016). However, to date in the PNS, the only known functions for Gpr126 are restricted to the nerve proper. Here, we show that Gpr126 is also critically important in non-myelinating tSCs at the NMJ within the muscle during long-term maintenance and regeneration. We show for the first time that Gpr126 is present in tSCs, and that loss of Gpr126 in tSCs leads to defects with and without nerve injury. We also found that long term NMJ maintenance is impaired in cGpr126 mice. Moreover, age-associated NMJ degeneration was noted in the hindlimb, but not the neck muscles, despite the fact that both the SM and EDL muscles have >50% Type II fibers (Eddinger, Moss, & Cassens, 1985; Guido, Campos, Neto, Marques, & Minatel, 2010). This result is unexpected considering the Gpr126 deletion is SC-specific. These anatomical differences may suggest that myelin is important to enable efficient action potential propagation to maintain the NMJ, therefore, distance from the cell body is relevant. cGpr126 mutant tSCs also display decreased S100β expression and decreased cytoplasmic process extension after nerve injury. Cytoplasmic process extension is a key morphological feature of repair SCs in peripheral nerves following axon injury (Gomez-Sanchez et al., 2017). It will therefore be interesting in the future to determine precisely how acquisition of repair SC phenotypes are impaired within the nerve of cGpr126 mice.

Furthermore, in the setting of nerve transection and repair, the immune response at the NMJ is altered in cGpr126 mice, with decreased initial macrophage presence, cytokine abnormalities, and subacute decreased VEGFA expression. Ultimately, cGpr126 mice display more NMJ denervation and decreased muscle mass compared to control mice following nerve injury. We conclude that tSCs are integral to NMJ reinnervation, and this reparative role is incomplete in the absence of Gpr126.

The data from the cGpr126 mice suggest a role for Gpr126 specific to tSC function following nerve injury. These cells displayed changes not explained by alterations to myelinating SC function alone. For example, S100β-positive tSC cytoplasmic processes extending away from denervated NMJs were decreased in number compared to those of CTRs. In the absence of a role for Gpr126 within tSCs, decreased process extension would be unexpected. In contrast, diminished reinnervation would impart increased and prolonged tSC process extension. In addition, S100β intensity was decreased in tSCs in the absence of Gpr126 both with and without nerve injury, which may contribute to effects on NMJ remodeling. In both injured and uninjured scenarios, tSC S100β intensity in cGpr126 mice was approximately half that of tSCs in control mice. S100β imparts function as well as form; for example, S100β promotes morphological SC changes after nerve injury (Perrone, Peluso, & Melone, 2008). Intracellularly, S100β is important for calcium and cytoskeletal protein metabolism (Fontaine, Lepape, Piriou, Venet, & Friggeri, 2016; Garbuglia et al., 1999; Gross, Sin, Barraclough, & Rudland, 2014; Zimmer, Cornwall, Reynolds, & Donald, 1998). These functions may relate to cytoplasmic process extension in the tSC; with decreased S100β expression, processes are unable to extend. The extracellular functions of S100β include promotion of injury-induced vascular remodeling (Cao et al., 2017) and activation of the immune response. In macrophages, CD68 binds to specific S100β proteins to regulate immune functions (Okada et al., 2016). S100β may be upregulated after injury to promote monocyte infiltration (Leanderson, Liberg, & Ivars, 2015), although decreased S100β expression has also been observed following injury and prior to reinnervation (Griffin & Thompson, 2008; Magill et al., 2007). In cGpr126 mutant mice, decreased macrophage presence and the altered immune response to nerve injury may be related to the decreased S100β expression within tSCs. It has also been reported that calcium/calmodulin and cAMP (PKA)-dependent protein kinases also affect intracellular levels of S100β (Davey, Murmann, & Heizmann, 2001). Whereas the function of cAMP in SC differentiation has been known for decades (Mirsky, Dubois, Morgan, & Jessen, 1990), the role of this second messenger in tSCs is less clear. Previous reports (Liebscher et al., 2014; Mogha et al., 2013) showed that Gpr126 directly increases cAMP by coupling to heterotrimeric G-protein and that cAMP levels are reduced in peripheral nerve of cGpr126 mice. Therefore, perhaps a lack of cAMP elevation underlies tSC phenotypes and intracellular S100β level observed in cGpr126 mice following nerve injury.

Delayed NMJ reinnervation is a known sequelae of cGpr126 deficiency in Schwann cells and is representative of abnormalities of NMJ components. In addition to the many differences in tSCs in cGpr126 mice compared to controls and the delayed reinnervation, we noted motor endplate (acetylcholine receptor) differences at 6 weeks post injury and repair, when NMJ reinnervation is >60%. While this observation may seem counterintuitive, it is likely related to endplate turnover. Motor endplates turnover consistently at baseline (Salpeter, Cooper, & Levitt-Gilmour, 1986). After denervation, a population of more rapidly turning over endplates is produced. Endplate turnover remains more rapid after nerve injury until the process of reinnervation is complete and until the more rapidly turning over endplate population has dissipated (Shyng & Salpeter, 1990; Vannucci et al., 2019; Xu & Salpeter, 1997). Endplate turnover does not necessarily directly correlate with reinnervation—ie: turnover can increase even as fewer NMJs remain denervated. The finding of endplate fragmentation in the setting of increased NMJ reinnervation is not unique.

In addition to the tSC response and delayed NMJ reinnervation, we observed reduced macrophage recruitment to the NMJ area in cGpr126 mice following nerve injury. In contrast, immature neutrophils, as represented by cells with ring-shaped nuclei staining for Ly6G, were elevated at 3 weeks after injury in cGpr126 mice. While the mechanism resulting in increased numbers of immature neutrophils present at the NMJ is not yet defined, we speculate that these cells may compensate for decreased macrophage presence. Alternatively, immature neutrophils may represent persistent inflammation in the cGpr126 mice, as neutrophils are typically an early mediator of inflammation. Additional hypotheses for the immature neutrophils include a second inflammatory response, an immature myeloid wave, or the result of chemokine imbalance. The role of immature neutrophils in Gpr126 deficient mice should be studied further. It is well documented that immune cell infiltration and response after nerve injury are essential for axon regeneration (Arthur-Farraj et al., 2012; Cattin et al., 2015). It is also reported that Gpr126 is required in myelinating SCs for macrophage recruitment to the peripheral nerve and chemokine expression upregulation following injury (Mogha et al., 2016). In this study, however, we show changes in the immune response downstream from the injury site—at the NMJ, not at the nerve.

The structure and function of NMJs is maintained by a complex micro-environment including the immune system, which creates a localized tissue niche. During nerve terminal degeneration, activated tSCs become phagocytic and together with macrophages remove nerve debris, supporting NMJ reinnervation after injury (Duregotti et al., 2015). In addition, upregulation of several cytokines, chemokines, and angiogenic factors are required for successful reinnervation. Here we show that in cGpr126 mice the mRNA expression patterns of Tnfα, a first-responder to injury (Qin et al., 2008), were low and paralleled reduced macrophage number trends. Concomitantly, Ccl2 mRNA expression, one of the TNF downstream targets (Bose & Cho, 2013; Toews, Barrett, & Morell, 1998), was increased significantly, suggesting a feedback loop in response to the decreased macrophages—i.e., Ccl2 expression may be upregulated in response to decreased macrophages, in an attempt to increase macrophage numbers. The cellular source of these chemokines was not defined in this study and may be macrophages, Schwann cells, or potentially another cell population. These data suggest that Gpr126 is required in tSCs to upregulate chemokine expression during NMJ reinnervation.

The mRNA expression patterns of pro-inflammatory chemokines provide further insight to the altered inflammatory response to nerve injury in the cGpr126 mice. IL-6, IL-10, and IL-12α proteins are all secreted by macrophages, and other cell types, but they do not all share identical expression patterns and timelines. In addition, their secretion patterns are dependent upon TNFα and IFNɣ(Flesch et al., 1995). IL-10 mRNA expression was noted to be significantly increased in cGpr126 mice relative to CTRs at all time points, but most pronounced at 1 week following injury. IL-10 has anti-inflammatory properties and inhibits macrophages when expressed at high levels (Fiorentino, Zlotnik, Mosmann, Howard, & O'Garra, 1991). Increased IL-10 expression, then, may be partially responsible for the diminished macrophage response to injury in the cGpr126 mice. The relationship between IL-10 and Gpr126 warrants further investigation. In contrast to IL-10, both IL-6 and IL-12α expression were noted to be lower in cGpr126 mice compared to CTRs at 3 weeks following injury. IL-6 interacts with IL-10 (Fiorentino et al., 1991; Wang et al., 1994; Wong et al., 2004), and the low IL-6 expression at 3 weeks in cGpr126 mice may result from negative feedback from the substantially elevated IL-10 levels acutely after injury. While the neuroregenerative effects of IL-12α are less known compared to its immunological contributions, IL-12α may affect both neurons and glial cells after nerve injury. IL-12α has been shown to enhance nerve regeneration and motor recovery in a 5 mm murine sciatic conduit model, in part via SC differentiation from neural stem cells (Lee, Wu, Arinzeh, & Bunge, 2017). Sympathetic neuron sprouting has also been increased in vivo with IL-12α treatment (Lin, Hikawa, Takenaka, & Ishikawa, 2000). In addition, IL-12α has anti-angiogenic properties (Bielawska-Pohl et al., 2010). This may, in part, explain the decreased Vegf levels noted in the cGpr126 mice at 6 weeks post-injury. However, the contributions of IL-12α to motor recovery following nerve injury require further investigation.

We also observed decreased Vegfa mRNA levels in cGpr126 mutant mice at 6 weeks following nerve injury. Vegf supports angiogenesis, which impacts nerve regeneration by providing guidance to supporting cells or by increasing the vascularity of the environment and transport of inflammatory and reparative mediators. Interactions of inflammatory mediators and Vegf are well-described (Bielawska-Pohl et al., 2010; Cattin et al., 2015). It has been demonstrated that Gpr126 also plays an important role in angiogenesis (Cui et al., 2014). GPR126 regulates VEGF Receptor 2 expression via PKA-CREB-GATA2/STAT5 signaling in vitro and in a zebrafish model. The interactions of SC-derived Gpr126 with the VEGF signaling pathway remain to be elucidated but may involve activation of protein kinase A (PKA). PKA is an integral downstream mediator of Gpr126 signaling (Glenn & Talbot, 2013). The interaction of these signaling pathways may hold promise for translational therapeutic application.

Our study highlights the importance of Gpr126 to non-myelinating tSCs, suggesting a new arena for scientific investigation as well as translational application. Despite the strengths of our study, the data have limitations. While we show the use of Dhh^Cre to mediate recombination in tSCs, all SCs are affected. We note NMJ anomalies that would not be expected as a result of changes to myelinating SCs alone, however, and observe changes intrinsic to tSCs, supporting the role for tSC-mediated effects. While we report differences in cGpr126 mice at the mRNA transcriptional and protein translational levels, our qPCR data do include other cells and not simply tSCs or macrophages, reflecting the environment at the NMJ. Because tSCs are not able to be extracted in isolation for mRNA extraction, we utilized the area of muscle with the highest density of tSCs. Functional assays are limited due to our SM model. We did observe changes in gross muscle mass as an overall marker of muscle health and physiology. The SM model is useful, however, as it provides the greatest density of NMJs and tSCs for evaluation. The data support the conclusion that Gpr126 is integral for non-myelinating tSC function at the NMJ with both cell autonomous and non-cell autonomous contributions. Without Gpr126 function, motor functional recovery is hindered. Gpr126, as a G-protein receptor, is an attractive therapeutic target in neuromuscular disease and injury for future clinical application.

Conclusions

Gpr126 is an adhesion G protein-coupled receptor with important functions in myelinating SC development, function, and response to nerve injury. Here, we show that Gpr126 is also critical in non-myelinating tSCs at the end target muscle for long-term maintenance and injury response. We demonstrate, for the first time, that Gpr126 is expressed in tSCs. There is no NMJ maintenance defect with loss of Gpr126 in SCs in young adult mice, but an NMJ maintenance defect was present in aging cGpr126 mice. We show important differences in tSCs of cGpr126 mice following nerve injury, with less S100β expression and fewer tSC processes. The immune response at the NMJ following nerve injury is altered in the muscle of cGpr126 mice. Ultimately, cGpr126 mice have decreased NMJ reinnervation and decreased muscle mass up to 6 weeks after nerve injury compared to controls. Gpr126 is integral to the tSC injury response and is a key contributor to recovery after injury. This knowledge is requisite for future therapeutic approaches targeting Gpr126. The implications of this study broaden the boundaries of Gpr126 investigation within the PNS and provide important therapeutic targets for neuromuscular disease.