The p75^NTR neurotrophin receptor is required to organize the mature neuromuscular synapse by regulating synaptic vesicle availability

The C57BL/6 J strain B6.129S4-Ngfrtm1Jae/J p75^NTR−/− mice [48] and their control pairs were purchased from The Jackson Laboratory (Sacramento, California, USA). All in vivo tests were carried out in procedure rooms equipped for that purpose. Mice were fed with pellet (Prolab RMH-3000, LabDiet) and water ad libitum and sacrificed by inhalatory isofluorane anesthesia overdose and posterior cervical dislocation, when indicated. Our procedures have been approved by the Bioethics Committee at the University of Concepcion, Chile, and follow the rules imposed by the Bioethics Committee of the National Commission for Scientific and Technological Research, Chile (CONICYT), and have therefore been performed in accordance with the ethical standards laid down in the Animals (Scientific Procedures) Act 1986, UK.

Diaphragm and Levator auris longus (LAL) muscles were dissected and whole-mount fixed in 0.5% formaldehyde (FA) in 1X Phosphate buffered saline (PBS) at 22 °C for 90 min. Samples were incubated with 0.1 M glycine in 1X PBS, permeabilized with PBST (1X PBS/0.5% TritonX-100) and blocked with 4% Bovine serum albumin (BSA) dissolved in PBST 12-16 h at 4 °C. Muscles were incubated with mouse monoclonal antibodies raised against neurofilament (2H3) (1:300) and synaptic vesicles (SV2) (1:50) (both from the Developmental Studies Hybridoma Bank, DSHB, of the University of Iowa, USA) along with a rabbit anti S100 antibody (1:300) (DAKO, Santa Clara, CA, USA) in 4% BSA-PBST for 30 min at RT and then 12-16 h at 4 °C. The tissues were incubated with the respective secondary antibodies (1:300) (Jackson Immuno Research, West Grove, PA, USA) in 4% BSA-PBST containing Alexa488-conjugated α-bungarotoxin (BTX) (Invitrogen, Carlsbad, CA, USA) (1:500) and DAPI (1:1000) (Thermo Fisher, Waltham, MA, USA) 12-16 h at 4 °C. Samples were post-fixed with 1% FA in 1X PBS for 10 min at 22 °C, flat mounted between two coverslips, and imaged. To analyze NMJ morphological maturation, z-stack images were collected at 1 μm intervals in a Zeiss LSM 780 confocal microscope at the CMA Bio-Bio facility, University of Concepcion, Chile. Maximal intensity projection images were reconstructed in 3D using the ImageJ software and analyzed to determine the proportion of the different postsynaptic NMJ morphologies, which were grouped into those having a single peripheral opening (still maturing) or those having multiple peripheral openings (mature pretzels). The morphology of > 130 NMJs per mice was manually determined and expressed as the percentage of the total. Images were obtained by processing confocal z-stack images using Imaris software. The surface, volume, and area of > 50 AChR densities per mice were determined for each postsynaptic structure using the ImageJ software, as described [40]. The LAL muscle innervation profile was analyzed in low magnification (10X) epifluorescence images (acquired with a Nikon Eclipse 80i microscope) of the entire rostral and caudal portions of the muscle, which were assembled using the ImageJ software. To determine NMJ innervation, confocal images were analyzed as described [40]. Briefly, the free-AChR presynaptic and the total AChR positive areas of > 35 NMJs per mice were calculated. Data are expressed as the fractional apposition between the pre- and postsynaptic domains. To analyze AChR dynamics, isofluorane anesthetized mice from both genotypes were subjected to a subcutaneous injection (in the head/neck region) of a non-saturating concentration (diluted 4 μg/mL in sterile 1X PBS) of Alexa-488 conjugated BTX (BTX-1). Subsequently, the right hemi-LAL muscle was denervated by facial nerve transection, as described [63]. Seven days post-surgery, mice were sacrificed and the LAL muscles were dissected, pinned to a sylgard dish and fixed with 0.5% v/v formaldehyde (Merck-Millipore) for 90 min at room temperature. After fixation, the muscles were washed and labeled with Alexa-555 conjugated BTX for 60 min (BTX-2, diluted 2 μg/mL in 1X PBS). Images (z-stacks) were collected at 1 μm intervals in a Zeiss LSM 700 confocal microscope at the CMA Bio-Bio facility (University of Concepcion, Chile). Following a blind assessment-based quantification, NMJs were categorized in “stable” if BTX-1 and BTX-2 labels were similarly intense, or as “dynamic” if BTX-1 labelling was mostly absent and BTX-2 intensity was comparatively higher. Data are expressed as the percentage of stable and dynamic denervated NMJ postsynaptic apparatuses in both genotypes [29].

NMJ-enriched samples from the diaphragm muscle were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4, incubated with 1% Osmium tetroxide for 2 h and dehydrated with graded ethanol series. After embedding in EPON, 70 nm ultrafine sections were obtained, contrasted with uracyl acetate and lead citrate, and visualized in a transmission electronic microscope PHILIPS TECNAI 12 BIOTWIN [13]. Images were acquired with the following magnifications: 6000X, 16,500X, 26,500X, and 43,000X. Quantifications were performed as described [66]. Briefly, the number of active zones, the number of vesicles per terminal area, the diameter of vesicles, and the number of active zones without docked vesicles (defined as those having their membrane apposing the presynaptic membrane), were quantified as presynaptic parameters. The readily releasable pool (RRP) of vesicles, defined as the vesicles contained within a 480 nm wide strip of the terminal directly across from the synaptic cleft, was also determined. Quantified postsynaptic parameters included the postsynaptic perimeter, the number, depth and width of the secondary folds, as well as the width of the primary fold, defined as the synaptic cleft space. Postsynaptic parameters were divided by the total apposition length of the motor terminal to control for differences in the size of each analyzed junction. All measurements were performed using the ImageJ software, as described [66].

Myofibers from the gastrocnemius muscle were dissociated and incubated for 3–4 days in matrigel-coated (ThermoFisher Scientific) dishes. For enrichment of the myoblast population, adhered cells were trypsinized and pre-plated twice onto uncoated dishes for 1 h at 37 °C and 5% CO₂. Cultures were maintained at less than 50% confluence in Bioamf-2 medium with 1% Pen/Strep at 37 °C and 5% CO₂ [36]. Primary myoblasts were seeded onto cell culture plates coated with poly-ornithine and laminin in L^− 15 medium, as described [45]. Cells were triggered to fuse into myotubes by incubation for 5 days with a differentiation medium containing 1X DMEM supplemented with 1% Glutamax-100, 1% Pen/Strep, 10% Fetal Bovine Serum, and 10% Horse Serum. AChR aggregates were labelled with Alexa488-conjugated BTX (1:500) for 45 min in culture conditions and the cells were subsequently fixed with 2% PFA for 20 min at 4 °C followed by incubation with 100% methanol for 5 min at − 20 °C. Cells were permeabilized with 0.1% TrisPO₄/Triton X-100 and incubated with a mouse anti α-tubulin (1:1000) antibody (Sigma-Aldrich, St. Louis, MO, USA) in 1% TrisPO₄ supplemented with 1% BSA 12-16 h at 4 °C. Cells were then incubated with the respective secondary antibodies (1:300) (Jackson Immuno Research) in 4% BSA-PBST containing Alexa488-conjugated BTX (1:500) and phalloidin (1:200) (Invitrogen) 2 h at 22 °C. Postsynaptic morphologies were categorized into “plaques”, small immature uniformly stained structures, and “complex” shapes, those comprising regions of low AChR density resulting in “O”, “C” and “pretzel-like” structures [45]. The myotube perimeter per field was manually traced on bright field images and their area was calculated using the ImageJ software. The proportions of the different postsynaptic structures in myotubes from both genotypes are expressed as a fraction of the total area of myotubes.

To analyze neuromuscular function, the effects of sciatic nerve stimulation frequency on muscle force-intensity and fatigue were studied after repetitive supramaximal stimulation protocols. Briefly, mice were anesthetized with ketamine (60 mg/kg) and xylazine (6 mg/kg) and an incision was made parallel to the femur bone to expose the sciatic nerve. A longitudinal skin incision was made to expose the Achilles tendon, which was separated from the tibia bone via blunt dissection, and firmly tied to a force transducer (UFI model 1030). Indirect stimulation was accomplished by placing bipolar electrodes in contact with the exposed sciatic nerve using a stimulator (model 611, Phipps and Bird, INC). Nerves were stimulated using 0.1 ms pulses, at supramaximal current intensity (5 mA) delivered at different frequencies. To induce fatigue, repetitive stimulations for 60s at 1, 10, and 100 Hz were performed. The different time phases of single twitch contractions as well as the contractile force and fatigue (force measurements as a percentage of maximal force) were analyzed using a Power Lab 4/35 device and the LabChart 8 software (AD Instruments, Dunedin, New Zealand). To measure skeletal muscle contractile properties, the Tibialis Anterior (TA) muscle was firmly tied from the tendon and the knee, dissected out, placed in a Plexiglas bath filled with oxygenated Krebs–Ringer solution, at 37 °C, and tied to a MLT 1030/D force transducer (ADinstruments, Oxford, UK). To determine isometric muscle force, the optimum muscle length and stimulation voltage were determined from micromanipulations of muscle length to produce the maximum isometric twitch force using a S48 Stimulator (Grass Research Instruments, West Warwick, RI, USA), that was controlled and measured using a PowerLab 4/35 device (AD Instruments, Oxford, UK). Isometric net force was determined with a stimulation frequency in the range 1–200 Hz for 450 ms, with 2 min of rest between stimuli. Specific net force was determined from the relationship between isometric net force (mN/mm2) using total muscle fiber cross-sectional area (mm²) to calculate it, as described [60]. Electromyography (EMG) was performed on isofluorane-anesthetized mice using Powerlab 26 T data acquisition hardware and LabChart 8 software (ADInstruments, Oxford, UK). Electrode distribution setup in mice was performed as described [16]. Briefly, the sciatic nerve was stimulated with two needle electrodes placed over the lumbar vertebral column. EMG was recorded by a needle electrode inserted into the right gastrocnemius muscle. The other recording electrode was subcutaneously placed in the right paw of the hindlimb. Repetitive nerve stimulation (RNS) protocol consisted in a 5 mA stimulation pulse at 50 Hz for 4-s. Finally, a single pulse of 5 mA was performed 7 s after RNS (post RNS). Compound muscle action potential (CMAP) values were calculated by the peak-to-peak amplitude values recorded. All-recorded CMAP values of the RNS stimulation were plotted to show the decreased values after a RNS or the fatigue. The first, the last and the post RNS CMAP obtained values were also separately plotted to determine the initial, fatigue and recovery values, respectively.

Samples of TA muscles, sciatic nerves, and spinal cord were homogenized in Tris-EDTA or RIPA (150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris pH 8.0) buffers containing a cocktail of protease inhibitors along with 2 mM sodium orthovanadate, 100 mM sodium fluoride, and 1 mM phenylmethylsulfonyl fluoride. For immunoblotting, 20–50 μg of total proteins were loaded in each lane and fractionated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred onto polyvinylidene difluoride (PVDF) membranes (EMD Millipore Corp.), and probed with mouse anti myosin heavy chain (MyHC) (1:1000), mouse anti troponin C (both from Developmental Studies Hybridoma Bank of the University of Iowa, USA), rabbit anti fibronectin (1:5000; Sigma-Aldrich), rabbit anti atrogin-1 (1:500), rabbit anti MuRF-1 (1:500; both from ECM Biosciences, Versailles, KY, USA), mouse anti TrkB (1:500; BD Biosciences, Franklin Lakes, NJ, USA), rabbit anti p-TrkB Y816 (1:500; Merck-Millipore, Burlington, MA, USA), rabbit anti BDNF (1:1000; Alomone Labs, Jerusalem, Israel). Mouse anti GAPDH (1:1000; Santa Cruz Biotechnology, TX, USA) or rabbit anti β-actin (1:2000; Abcam, Cambridge, UK) antibodies were used as loading controls. In control experiments of skeletal muscle atrophy, TA muscles were obtained from adult WT mice treated with angiotensin II, as described [60]. All immunoreactions were visualized by enhanced chemiluminescence (Thermo Fisher Scientific, Waltham, MA, USA) and acquired using a Fotodyne FOTO/Analyst Luminary Workstations System (Fotodyne, Inc., Walnut Ridge, WI, USA). Data are expressed as the ratio of the protein of interest v/s loading control band intensity ratios, which were quantified using the ImageJ software [60].

Feshly frozen TA muscles were embedded in optimal cutting temperature (OCT) compound (Sakura Fine technical Co., Torrance, CA), sectioned every 20 μm with a cryostat (Thermo Scientific Microm HM 525) and mounted on Vectabond (Vector Laboratories) coated slides. Muscle fiber morphology was analyzed with conventional haematoxylin/chromotrope staining. Skeletal muscle sarcolemma and nuclei were stained with 1 μg/ml Alexa488-conjugated wheat germ agglutinin (WGA, Invitrogen) and 0.3 μM DAPI (Thermo Fisher), respectively [89]. Cryosections were also stained with a NADH reduced solution (Tris-buffer, pH 7.4, NADH reduced, nitro-blue tetrazolium) (Sigma-Aldrich) for 45 min, and fibers were classified into slow (dark blue), intermediate (blue), and fast twitch (light blue). The identity and cross-sectional area (CSA) of > 100 fibers per type per mouse were determined using the ImageJ software and are expressed as the percentage of the total [89]. For collagen second harmonic generation imaging, cryosections (20 μm) of TA muscles were imaged using an inverted Zeiss LSM 780 multiphoton laser scanning confocal microscope at the CMA Bio-Bio facility (University of Concepcion, Chile). An excitation wavelength of 800 nm was used for 400 nm detection, using a 40X objective. For immunohistochemistry, cryosections (20 μm) of TA muscles were fixed with 0.5% v/v formaldehyde (Merck-Millipore) for 90 min, washed with PBS 1X containing 0.5% Triton X-100, blocked in PBS 1X, 0.5% Triton X-100 and 4% BSA for 16–18 h at 4 °C, and incubated with mouse anti myogenin (1:5), mouse anti embryonic MyHC (1:5; both from the Developmental Studies Hybridoma Bank), rabbit anti fibronectin (1:200; Sigma-Aldrich), rabbit anti p-TrkB Y816 (1:200; Merck-Millipore), and rabbit anti BDNF (1:200; Alomone Labs) antibodies for 16 h at 4 °C. After washing, samples were incubated with a Cy2-conjugated anti mouse antibody along with Alexa555-conjugated BTX, washed, and mounted with fluorescence mounting medium (DAKO). In control experiments of skeletal muscle degeneration/regeneration, the TA muscle of adult WT mice was injected along its whole length with an aqueous 1.2% mass/volume barium chloride solution, as described [7]. Samples were obtained after 10 days of recovery and subjected to immunohistochemistry. In control experiments of skeletal muscle denervation, the sciatic nerve of adult WT mice was exposed and a 3–5 mm section was resected, as described [51]. After 7 days, TA muscles were dissected and processed for immunohistochemistry, as above described.

The kyphotic index (KI) was calculated from X-ray images (Veterinary Clinic, University of Concepcion) based on the distance (D1) from the seventh cervical vertebra (C7) to the sixth lumbar vertebra (L6) and the perpendicular distance (D2) from D1 to the point of maximum vertebra curvature. KI corresponds to the D1/D2 ratio and is inversely proportional to kyphosis [47, 58]. The footprint test measures stride length and the distance between fore and hind limb paw prints left on a filter paper by mice trained to run down a runway with painted feet [6]. Balance and motor coordination was evaluated using the rotarod test. Mice were positioned on horizontal cylinders with increasing acceleration from 4 to 40 rpm for 120 s and the latency time was recorded. Eight measurements were taken per mouse, with 2 min rest, on two consecutive days [6]. In the horizontal triple bar test, mice were positioned in the center of 3 horizontal rods of different diameters (2, 4 and 6 mm). The time taken by the mice to reach the end of each rod was recorded and scored, as described [14]. In the static bars test, mice were positioned at the end of 5 wooden bars of decreasing diameters (31 to 8 mm), which are joined at one end to a platform, and the time the mice take to reach the free end was recorded [14].

In the Kondziela’s inverted screen test, mice were positioned in the center of a metallic mesh which was rotated 180°. The time taken before the mouse detached from the mesh was quantified and a score assigned. Measurements were made 3 times for each mouse separated by 30 min with a maximum test duration of 120 s [15]. In the weights test, mice were challenged to hold different weights (15.5, 23.1, 30.8, 39.4, 46.4, 54.1 g) with their forelimbs while suspended from the tail. A score was assigned according to the time they were able to hold the different weights. The maximum time of maintenance score was 3 s. The final score was normalized to the body weight of each mice [15, 60]. When indicated, p75^NTR−/− mice were subcutaneously injected daily with 3 μmol/Kg of the acetylcholinesterase inhibitor pyridostigmine bromide in commercial physiological saline [50]. Controls were similarly injected with vehicle. One hour after drug administration, mice were challenged in the weights test.

As a first attempt to examine whether p75^NTR is involved in the structural organization of the adult NMJ, cranial LAL muscles from WT and p75^NTR−/− mice were stained to reveal the pre and postsynaptic profiles, as well as Schwann cells. Low magnification analyses indicate no gross differences in NMJ organization, as evidenced by unaltered distribution of the main innervation profiles [61] (Additional file 1: Figure S1). Higher magnification images revealed that nerve terminal branches aligned precisely with their respective postsynaptic specialization in both WT and p75^NTR−/− mice muscles (Fig. 1a, b). Quantitative analyses confirmed that NMJs are fully innervated, as the area of presynaptic staining apposes more than 70% of the postsynaptic domain in both genotypes [40] (Fig. 1c). In addition, myonuclei from TA muscle fibers from both genotypes do not express detectable levels of myogenin (Additional file 2: Figure S2a), a molecular marker of muscle denervation [43]. Similarly, terminal Schwann cells distribute in precise register with presynaptic terminals and postsynaptic AChR staining in p75^NTR−/− and control NMJs (Fig. 1b). Despite these results, we consistently found that postsynaptic domains of p75^NTR−/− mice were smaller and less complex than those from WT mice (Fig. 1b). As postsynaptic maturation is characterized by the presence of AChR-poor regions [54], we quantified postsynaptic structures containing a single (i.e. still maturing) or multiple peripheral openings (i.e. fully mature) (Fig. 1d). NMJs from 2-month-old p75^NTR−/− mice show a significant decrease in the proportion of mature pretzel-like shapes, compared to WT controls (Fig. 1e). In 3D projections of NMJs from both genotypes (Additional file 3: Figure S3), we quantified significantly decreased values of postsynaptic area (Fig. 1f), surface (Fig. 1g) and volume (Fig. 1h), consistent with the relative abundance of less complex maturing structures found in p75^NTR−/− mice. Based on these findings, we next analyzed potential postsynaptic NMJ defects at the ultrastructural level (Fig. 1i). Whereas the secondary clefts of control NMJs are long, thin, and closely spaced, these structures are reduced in number (Fig. 1j) and give rise to decreased postsynaptic perimeter in p75^NTR−/− mice (Fig. 1k). Our quantitative analyses revealed that other parameters, such as the depth and width of the secondary folds, as well as the width of the primary fold, are not significantly affected in p75^NTR−/− mice (Fig. 1l–n). Together, our findings reveal that the absence of p75^NTR negatively affects the organization of the NMJ postsynaptic domain.

We next aimed to correlate our morphological observations with functional NMJ parameters by analyzing the effect of sciatic nerve stimulation on hind limb muscle force generation and fatigue. After single twitch stimulation (Fig. 2a) we found no differences in the latency period (Fig. 2b) and a slight but significant decrease in the contraction time in p75^NTR−/− mice (Fig. 2c). After repetitive nerve stimulation protocols (Fig. 2d), we found that p75^NTR−/− mice muscles displayed a significantly increased fatigue after 30s of tetanic 100 Hz nerve stimulation (Fig. 2e). We complemented these studies with electromyography recording after presynaptic stimulation (Fig. 2f). We found that p75^NTR−/− mice display decreased CMAP, a feature that was enhanced after repetitive nerve stimulation (Fig. 2g). After 7 s of repetitive nerve stimulation, single CMAP values were similar in control and p75^NTR−/− mice muscles (Fig. 2g). Together, these results reveal that p75^NTR−/− mice display altered nerve-induced muscle responses.

To analyze if sustained defective synaptic activity resulted in skeletal muscle defects in p75^NTR−/− mice, we first quantified the total levels of the sarcomeric proteins myosin heavy chain (MyHC) and troponin C (TnC) by Western blot (Fig. 3a). Compared to the loading control GAPDH, we found no differences in the levels of both proteins in the different genotypes (Fig. 3b). Next, we measured the force elicited after direct stimulation of the TA muscle at different frequencies. Our findings show a significant decrease in the net isometric force in the range of 30 to 100 Hz stimulation in p75^NTR−/− TA muscles (Fig. 3c). These findings correlate with a significant reduction in the percentage of maximum isometric contraction force developed by p75^NTR−/− mice muscles after 60 Hz stimulation (Fig. 3d). As these findings reveal defective muscle contraction in p75^NTR−/− mice, we next studied muscle structure by histological analyses of transversal TA muscle cryosections (Fig. 3e). Our results show no gross differences in muscle fiber distribution or mononuclear infiltration (H&C staining). We also observed no differences in spontaneous muscle fiber degeneration/regeneration cycles, as cryosections from both genotypes show similar WGA/DAPI staining (Fig. 3e), a similarly low proportion of myofibers displaying central nuclei (Fig. 3f), and the absence of fibers expressing an embryonic form of MyHC (Additional file 2: Figure S2b), a molecular marker of muscle regeneration [32]. To evaluate potential differences in inter-fiber extracellular matrix deposition, we first performed immunohistochemical staining (Fig. 3g) and Western blot analyses (Fig. 3h, i) to study the levels of fibronectin. We found no differences in fibronectin levels or distribution in TA muscle samples from both genotypes. To complement these studies, we standardized the second harmonic generation technique, a two-photon laser confocal analysis that allows label-free imaging of collagen fibers at high resolution on native tissue [23]. Following this approach, we consistently found that the levels of collagen deposition, as well as the normally linear organization of collagen fibers, are not affected in skeletal muscles of the p75^NTR−/− mice, compared to controls (Fig. 3g). We next evaluated muscle fiber plasticity by histochemical staining to reveal NADH-thioreductase (NADH-TR) activity (Fig. 3j). Quantification shows no differences in the relative composition of p75^NTR−/− and control TA muscles regarding fast (light blue), intermediate (blue), and slow-twitch (dark blue) fibers (Fig. 3k). However, we found a significant reduction in the cross-sectional area (CSA) of muscle fibers of p75^NTR−/− mice, evidenced by a strong increase in the proportion of low caliber fibers and a corresponding decrease in higher caliber fibers of both, slow and fast-twitch muscle fibers (Fig. 3l–m). To analyze if the decrease in muscle fiber size observed in p75^NTR−/− mice was due to atrophy, we conducted Western blot experiments to analyze the levels of MuRF-1 and atrogin-1, two E3 ubiquitin ligases that act as key regulators of ubiquitin-mediated protein degradation in skeletal muscle [4]. We detected similar low levels of both proteins in TA muscle samples from both, control and p75^NTR−/− mice (Additional file 2: Figure S2c). Together, these findings evidence that the absence of p75^NTR results in functional NMJ defects, which correlate with altered structure and contractile capabilities of skeletal muscles.

To study if the alterations in muscle contraction, ultrastructure, and function of the neuromuscular synapse found in p75^NTR−/− mice resulted in defective motor behavior, several motor coordination and balance tests were performed. We first found that p75^NTR−/− mice display altered walking gait, evidenced by an increase in the step distance (Fig. 4a) and in the traces between fore and hind limbs (Fig. 4b). Also, p75^NTR−/− mice displayed a shorter latency time to fall in the rotarod test (Fig. 4c), and required more time to reach the end, or fell out, either in the triple horizontal bars (Fig. 4d) or in the static bar tests (Fig. 4e). As motor alterations are due to anatomical defects in mice models displaying NMJ phenotypes [58], we also measured the kyphotic index (i.e. abnormal spine curvature); however, no differences were found between p75^NTR−/− and WT mice (Fig. 4f–g). Based on our previous findings, we next evaluated whether the absence of p75^NTR results in force impairment in the context of the entire animal. Our results show that p75^NTR−/− mice are able to hold their body weights in the inverted grid test, as WT mice do (Fig. 4h). However, when challenged to hold increasing weights with their fore limbs, p75^NTR−/− mice show a significantly impaired performance (Fig. 4i). Thus, the absence of p75^NTR results in defective locomotor behavior.

Mature NMJs express low levels of p75^NTR in motor axon terminals, terminal Schwann cells, and muscle fibers, in close vicinity to AChR aggregates [26, 27]. In order to gain insights on the cellular source of the p75^NTR fraction potentially involved in postsynaptic NMJ organization and ultrastructural assembly, we first analyzed the aneural formation of complex AChR structures on cultured myotubes obtained from satellite cells of WT and p75^NTR−/− mice (Fig. 5a). Our analyses revealed similar myotube formation and size in both genotypes (Fig. 5b). When seeded onto polyornithine/laminin matrices, myotubes assemble complex postsynaptic structures resembling those observed in vivo [45]. We categorized AChR aggregates into “immature plaques”, having homogenous AChR distribution, and “complex maturing forms”, corresponding to those having AChR-poor subregions forming “O-”, “C-” or “pretzel-like” shapes (Fig. 5c). Our morphological classification indicates no differences in the complexity of AChR aggregates amongst myotubes derived from p75^NTR−/− and WT mice (Fig. 5d). Consistently, no differences were detected in the size of postsynaptic structures (Fig. 5e) or in the ability of myotubes from both genotypes to assemble complex postsynaptic structures, quantified as the number of AChR complex structures per myotube area (Fig. 5f). These findings suggest that the absence of the muscle-derived p75^NTR receptor does not alter the assembly of mature postsynaptic structures in the p75^NTR−/− mice in vivo.

As a way to investigate the possibility that muscle-derived p75^NTR affects postsynaptic maintenance, the dynamics of AChRs after NMJ denervation were analyzed following an in vivo two-color BTX method [29]. With this aim, NMJ postsynaptic domains of LAL muscles were labeled with a non-saturating dose of a fluorescently tagged BTX (BTX-1) in live mice from both phenotypes and the right hemi-LAL muscle was subsequently subjected to denervation through facial nerve resection. After a 7-day recovery period, the pre-labeled muscles were dissected and labeled with a different fluorescently tagged variant of BTX (BTX-2). Using confocal microscopy, AChR aggregates were categorized in “stable” if BTX-1 and BTX-2 labels were similarly intense, or as “dynamic” if BTX-1 labelling was mostly absent and BTX-2 intensity was comparatively higher (Fig. 5g). As expected, denervation resulted in a significant increase in the proportion of dynamic AChR aggregates in LAL muscles from both, control and p75^NTR−/− mice; however, no differences were detected in the proportion of dynamic AChR aggregates amongst denervated LAL muscles from p75^NTR−/− and WT mice (Fig. 5h). These findings suggest that the absence of the muscle-derived p75^NTR does not differentially affect the stability of mature postsynaptic structures after denervation.

Even though we did not find differences in the profile of muscle fiber innervation (Fig. 1a–c), our previous findings showing functional NMJ defects in p75^NTR−/− mice prompted us to examine the presynaptic ultrastructure (Fig. 6a). Remarkably, quantitative analyses showed a drastic decrease in the total vesicle density in p75^NTR−/− motor terminals (Fig. 6b). p75^NTR−/− mice also displayed a significant reduction in the density of the readily releasable pool of vesicles (RRP) (Fig. 6c), as well as a strong decrease in the number of active zones per unit length of terminal apposition (Fig. 6d). No differences were found in the diameter of individual synaptic vesicles (Fig. 6e).

Considering that p75^NTR−/− mice display decreased number of vesicles in the motor axon terminal as well as postsynaptic morphological defects, we hypothesized that impaired neurotransmission could contribute to deficient NMJ function, subsequent muscle contraction impairment, and defective locomotor performance. To experimentally address this idea, pharmacological interventions were conducted in which the enzyme acetylcholinesterase was inhibited by subcutaneous administration of pyridostigmine bromide. After daily treatment with the drug, p75^NTR−/− mice were challenged in the weights test to evaluate muscle strength. Our results show that p75^NTR−/− mice had partially rescued strength at 2 and 3 days after daily drug administration compared to p75^NTR−/− mice treated with physiological saline (Fig. 6f). Acetylcholinesterase inhibition also improved the electromyographic recording of p75^NTR−/− NMJs, reflected in increased CMAP activity of p75^NTR−/− pyridostigmine treated animals from 1 to 3 days (Fig. 6g). Our results also showed that the positive effects of acetylcholinesterase inhibition are not sustained in time, as CMAP values did not increase after 5 days of treatment (Fig. 6g). Altogether, these findings reinforce the idea that impaired neurotransmission contributes to deficient NMJ function in p75^NTR−/− mice.

It has been demonstrated that the levels of NTs and their Trk receptors are altered in conditions affecting the neuromuscular synapse. For instance, hind limb muscles of amyotrophic lateral sclerosis mouse models display a significant decline in the number of NMJs expressing BDNF, NT-4, GDNF, p75^NTR, TrkB, and TrkC [31]. Therefore, as a first hint to analyze the potential involvement of NT-dependent signaling on the morphological and functional alterations observed in p75^NTR−/− mice, we analyzed the levels of expression of BDNF and its TrkB receptor at the NMJ, as BDNF/TrkB signaling plays important roles on NMJ organization, including synaptic vesicle availability [34, 75]. With that aim, cryosections of TA muscles from control and p75^NTR−/− mice were subjected to immunohistochemistry to detect BDNF and an active form of TrkB (p-TrkB Y816) (Fig. 7). Our findings show that NMJs from control and p75^NTR−/− mice display positive staining for BDNF and p-TrkB Y816 (Fig. 7a). Immunofluorescence quantification shows that the vast majority of NMJs express both proteins in TA muscle cryosections of both genotypes (Fig. 7b, c). Following Western blot experiments, we analyzed the levels of TrkB and BDNF. We also included the antibody that specifically detects active TrkB when phosphorylated at residue Y816 of both, the glycosylated (145 kDa) and the unprocessed (110 kDa) forms of TrkB [38]. Our results show that the total levels of BDNF and TrkB are not significantly altered in p75^NTR−/− mice muscles (Fig. 7d). Consistently, immunoblot quantifications show that the levels of pro-BDNF (the precursor form of BDNF) or the active p-TrkB Y816 form of TrkB were unchanged by the absence of p75^NTR at the NMJ (Fig. 7e, f). Similarly, the levels and distribution of BDNF and TrkB in samples of sciatic nerve and the spinal cord of p75^NTR−/− mice were similar to control mice (Additional file 4: Figure S4). Together, these findings suggest that BDNF/TrkB signaling is not significantly altered at the neuromuscular synapse of p75^NTR−/− null mice.