Acute blockade of signaling through the tyrosine kinase receptor B (TrkB) attenuates neuromuscular transmission and fragments postsynaptic acetylcholine receptors (AChRs) in adult mice, suggesting that TrkB signaling is a key regulator of neuromuscular function. Using immunohistochemical, histological, and in vitro muscle contractile techniques, we tested the hypothesis that constitutively reduced TrkB expression would disrupt neuromuscular pre- and postsynaptic structure, neurotransmission, muscle fiber size, and muscle function in the soleus muscle of 6- to 8-mo-old TrkB+/− mice compared with age-matched littermates. Age-like expansion of postsynaptic AChR area, AChR fragmentation, and denervation was observed in TrkB+/− mice similar to that found in 24-mo-old wild-type mice. Neurotransmission failure was increased in TrkB+/− mice, suggesting that these morphologic changes were sufficient to alter synaptic function. Reduced TrkB expression resulted in decreased muscle strength and fiber cross-sectional area. Immunohistochemical and muscle retrograde labeling experiments show that motor neuron number and size are unaffected in TrkB+/− mice. These results suggest that TrkB- signaling at the neuromuscular junction plays a role in synaptic stabilization, neurotransmission, and muscle function and may impact the aging process of sarcopenia.
- neuromuscular junction
- muscle weakness
- tyrosine kinase receptor B
multiple lines of evidence indicate that presynaptic nerve terminals and muscle fibers communicate with each other to establish a precisely aligned neuromuscular junction during postnatal synapse elimination (38, 44). During the aging process, however, this precise alignment of synaptic structure is lost. The postsynaptic acetylcholine receptor (AChR) area expands as extrajunctional receptors are inserted into the end-plate region, the neuromuscular junction becomes disorganized as presynaptic terminals are alternately lost and added, excitation-contraction coupling is weakened, neuromuscular safety factor is reduced, and ultimately muscle fibers become progressively denervated (1, 3, 7, 8, 12, 47, 50).
Neurotrophins participate in both stabilization of synaptic structure and activity-induced modulation of synaptic transmission in multiple central nervous system areas, including the hippocampus (9, 43, 53), the visual system (26, 48), and the cerebellum (5). A number of neurotrophins could potentially affect neuromuscular structure (28), but brain-derived neurotrophic factor (BDNF) and neurotrophin-4/5 (NT-4/5) are prime candidates because they are produced by motor neurons and muscle fibers (15, 24, 27), and their high-affinity tyrosine kinase B (TrkB) receptor is coexpressed with AChR along the postsynaptic membrane of the neuromuscular junction. Furthermore, expression of NT-4/5 and TrkB is innervation dependent (15, 41).
TrkB-mediated signaling affects synaptic stabilization at the neuromuscular junction. Adenoviruses have been used to overexpress a truncated TrkB isoform (TrkB.t1), which lacks the intracellular tyrosine kinase domain, to reduce full-length TrkB signaling in muscle fibers of young mice (17). This dominant-negative manipulation results in the acute fragmentation of postsynaptic AChRs into multiple discrete regions. Manipulating NT-4/5 and BDNF levels also impacts neuromuscular stabilization. Intramuscular administration of NT-4/5 results in sprouting of intact adult motor nerves (15), and BDNF and NT-4/5 alter neuregulin and agrin expression, which may influence AChR synthesis and clustering (35, 52). Constitutive NT-4/5 knockout mice demonstrate AChR disassembly (similar to that found after TrkB.t1 overexpression), reduced AChR binding, reduced electromyographic responses, and increased muscle fatigue in slow-twitch muscle (4). Thus TrkB-mediated signaling, via NT-4/5 or BDNF, may act as an activity-dependent neurotrophic signal for stabilization of the adult neuromuscular junction, whereas reduced TrkB-mediated signaling may underline age-associated changes in the neuromuscular junction during senescence.
Here, we test the hypothesis that TrkB+/− mice, which have a constitutive ∼50% reduction in full-length and marginally reduced TrkB.t1 expression (17, 23), will demonstrate precocious age-like alterations in pre- and postsynaptic neuromuscular structure, neurotransmission failure, and muscle weakness. Fast-twitch type IIA and IIB fibers demonstrate more profound changes in neuromuscular junction structure and function during aging than slow-twitch type I fibers (2, 7, 14, 22, 50, 51). This study, however, focuses on the predominately slow-twitch soleus muscle because 1) NT-4 expression is maximized in adult slow-twitch fibers (15), 2) NT-4 knockout results in greater physiological effects in slow-twitch muscle (4), and 3) NT-4 mRNA levels in adult muscle decrease after blockade of neuromuscular junction transmission and increase after electrical stimulation, suggesting a role for NT-4/TrkB-mediated signaling in activity-dependent neuromuscular junction remodeling (15).
MATERIALS AND METHODS
Studies were performed in heterozygous B6.129S2Ntrk2tmlBbd/J (TrkB+/−) mice or wild-type (WT) littermates (TrkB+/+) obtained from Jackson Laboratories (Bar Harbor, ME). Animals were backcrossed for eight generations, removing ∼99.6% of founder strain polymorphisms (personal communication with Dr. Steven Welle). B6.129S2-Ntrk2tmlBbd/J mice were generated through targeting the protein-tyrosine kinase domain, which is necessary for intracellular signaling (23). All experimental groups were age and sex matched. The average ages of TrkB and WT mice were 6.9 ± 0.3 and 6.4 ± 0.3 mo, respectively (P = 0.212, Student t-test; range of 6 to 8 mo). A total 12 TrkB+/− and 12 WT mice were used for these experiments. Experimental groups consisted of eight male and four female mice for each genotype. A subset of experiments were performed in 24-mo-old C57BL/6 mice (n = 5) to verify age-like changes. All procedures were approved by the University at Buffalo Animal Care and Use Committee.
Animals were anesthetized to a surgical plane of anesthesia by intraperitoneal injection of Nembutal at 100 mg/kg, and both hindlimbs were removed at the proximal femur followed by death from exsanguination. The limbs were transferred into cooled Krebs solution (137 mM NaCl, 4 mM KCl, 1 mM MgCl2, 1 mM NaH2PO4, 12 mM NaHCO3, 2 mM CaCl2, and 6 mM d-glucose) constantly aerated with 95% O2-5% CO2. The soleus was removed with both proximal and distal attachments while maintaining a segment of the tibial nerve. The soleus from the second hindlimb was removed and cleared of connective tissue to allow maximal antibody penetration for immunohistochemistry. Tails were cut and stored at −80°C for genotyping by quantitative PCR.
Muscle contractile properties and neurotransmission failure.
The soleus was transferred to a water jacket bath filled with constantly aerated Krebs solution at 27°C. The muscle prep was then fixed at the calcaneal tendon via alligator clip, while the proximal tendon was fixed to the arm of an isometric force transducer via suture (300B Series Lever System; Aurora Scientific, Ontario, Canada). The muscle was directly stimulated by two chloride silver plate electrodes using rectangular pulses of anodal current with 2.0-ms duration at supramaximal voltage (Grass S88 stimulator with custom current amplifier; Astro-Med). Before maximal isometric muscle force was determined, optimal muscle fiber length for force production (Lo) was obtained by determining the muscle length that produced the greatest twitch force. Isometric force/frequency data were acquired by measuring force output during administration of 900-ms pulse trains with intratrain pulse rates of 10, 20, 35, 50, 65, 80, 100, 125, 150, and 200 Hz. The tibial nerve was then pulled into a suction electrode so that the muscle could be activated via indirect nerve stimulation using rectangular pulses of anodal current with a 2.0-ms duration (Master 8; AMPI, Jerusalem, Israel). Force measurements were recorded and analyzed using Spike2 V6.07 (Cambridge Electronic Design, Cambridge, England).
A 100-Hz stimulation generally produces the greatest isometric force in the soleus muscle. Thus the neurotransmission failure (NF) protocol was carried out at 100 Hz. NF was measured using methods previously described (40). The soleus muscle was stimulated via the tibial nerve with 330-ms trains at 1 Hz with an intratrain pulse rate of 100 Hz. A single train of direct muscle stimulation with similar train characteristics was superimposed on the nerve stimulation every 15 s. The alternating stimulation paradigm between nerve and direct muscle stimulation continued for 5 min. NF was determined every 15 s using the equation NF = (F − MF)/(1 − MF) where F is percent force loss during nerve stimulation and MF is percent force loss due to contractile fatigue (25). This method of determining NF assumes the forces generated by nerve and direct muscle stimulation are initially similar; thus we determined force produced by nerve stimulation and direct muscle stimulation before initiating the NF protocol. Force output between the two stimulation methods needed to be within 90% of each other to be included in the data set. Following measurement of NF, the soleus was trimmed of excess connective tissue, measured for wet weight, stretched to Lo on cork, “snap” frozen in liquid nitrogen cooled isopentane, and stored at −80° C for muscle histology.
Neuromuscular junction immunohistochemistry, image collection, and analysis.
The second hindlimb from each animal was fixed in 4% paraformaldehyde (pH 7.4) for 15 min with the knee and ankle joints pinned at 90°. The soleus muscle was subsequently removed, and postsynaptic AChRs were labeled with rhodamine-conjugated α-bugarotoxin (RαBTX; 1:10 concentration; Molecular Probes, Carlsbad, CA). The tissue was incubated in −20°C MeOH for 5 min and then blocked in 0.2% Triton-X, 2% BSA, and 0.1% sodium azide for 1 h (36). Samples were bathed overnight with primary mouse monoclonal antibodies, anti-2H3 (1:10), and anti-SV2 (1:20), to label neurofilament and presynaptic vesicles, respectively (Dev. Studies Hybridoma Bank, University of Iowa). AlexaFluor 488 goat anti-mouse (1:100; Molecular Probes, Invitrogen) was used to detect the presence of primary antibodies. Muscle samples were bathed in secondary antibodies for 4 h at room temperature. Secondary-antibody only controls were used to verify primary antibody specificity.
Soleus muscles were filleted, mounted surface side up in Vectashield mounting media (Vector Laboratories, Burlingame, CA), and imaged using a Zeiss Axioimager Z1 mounted on a Zeiss LSM-510 Meta NLO confocal [objective: Plan-Apochromat 63×/1.4 oil differential interference contrast (DIC)]. Neuromuscular junctions were selected for imaging based on their enface orientation and clear nerve labeling. AChR and nerve image stacks were collapsed into a single plane maximum-intensity projection using ImageJ (NIH) before analysis (http://rsbweb.nih.gov/ij/). The extent of AChR-nerve overlap (colocalization) for each synapse was determined using a colocalization plug-in. Colocalization plug-in settings were standardized across animal genotype. AChR area, perimeter, inner radius, outer radius, mean radius, average gray value, optical density variance, and AChR-TrkB colocalization area were measured using MetaMorph 5.0 (Universal Imaging, Downingtown, PA). The total number of discrete AChR clusters, defined by clear visible separation from neighboring regions, for each neuromuscular junction was counted manually. The percentage of AChR-nerve overlap was determined by dividing the area of AChR-nerve colocalization area by total AChR area (39).
Tetrodotoxin-resistant sodium channel (Nav1.5) immunoreactivity may be used to identify denervated skeletal muscle (50). The heart was used as a positive control as cardiac muscle normally expresses Nav1.5. Frozen transverse muscle sections (20 μm) were labeled with affinity-purified rabbit polyclonal anti-Nav1.5 (1:200; Sigma, Atlanta, GA) and visualized by AlexaFluor 488. Soleus muscle sections from TrkB+/− and WT mice along with heart tissue sections were placed on the same slide to control for differences in immunolabel intensity across preparations. Sections were imaged using a Zeiss Axioimager Z1 mounted on a Zeiss LSM-510 Meta NLO confocal (objective: Plan-Apochromat 20×/0.8 DIC). Exposure time and thresholding were set based on the labeling intensity of the heart tissue sections for each slide. The ratio of muscle fibers with >50% of their membrane immunolabeled for Nav1.5 vs. fibers with <50% label was determined from five muscle sections per animal. A total of 1,408 and 1,442 muscle fibers were assessed from WT and TrkB+/− mice, respectively. The average percentage of muscle fibers with >50% labeling was calculated per animal. Muscle fibers with >50% Nav1.5 membrane labeling were considered denervated.
Muscle fiber and motor neuron: number and cross-sectional area.
Soleus muscles were cryosectioned mid-belly (20 μm) and placed on electrically charged slides (Mercedes Medical, Sarasota, FL). Muscle sections were fixed in 4% formaldehyde for 15 min and stained with hematoxylin and eosin as previously described (37). The entire muscle cross section was visualized via AxioVision LE V4.8 (Carl Zeiss MircoImaging, Thornwood, NY) so that the total number of muscle fibers within the soleus muscle could be counted. Fiber cross-sectional area (CSA) was measured in 50 fibers within the center of the muscle using MetaMorph 5.0 (Universal Imaging).
To assess whether reduced TrkB expression altered motor neuron number or morphology, lumbar spinal cords were fixed in situ via transcardial perfusion of 4% paraformaldehyde, cryoprotected in 20% sucrose solution, mounted in molds containing Tissue-Tek OCT compound (Sakura Finetek, Torrance, CA), and frozen in dry-ice cooled acetone. Frozen tissue sections (20 μm) were picked up on charged glass slides, rinsed with PBS, blocked with 1% BSA and 0.1% TritonX-100 in PBS, incubated with anti-choline acetyltransferase (ChAT) (1:200; Chemicon International, Temecula, CA), and labeled by AlexaFluor 488 goat anti-mouse (1:100; Molecular Probes Invitrogen,). Sections were visualized by confocal microscopy (objective: Plan-Apochromat 20×/0.8 and 63×/1.4 oil DIC). Motor neurons were identified on the basis of their large soma, anterior horn location and ChAT labeling and were only evaluated if a nucleus was visible. The number of motor neurons was hand counted in 16 cross sections per animal and averaged to determine the average motor neuron number per hemisection. Motor neuron soma CSA was assessed in these same neurons using MetaMorph 5.0 (Universal Imaging).
Retrograde labeling of motor neurons.
To determine the number of motor neurons innervating the soleus muscle, mice were anesthetized as described above and the soleus muscle was exposed without damaging surrounding musculature. Motor neurons that innervate the soleus were retrogradely labeled with Fluoro-Gold (Fluorochrome, Denver, CO) as described by Rott-Percelay et al. (42). Briefly, one injection of 400 μl 4% Fluoro-Gold in saline was made into each muscle. The wounds were sutured closed and 4 days later, and the mice were euthanized with an overdose of a solution of 17.4 mg/ml ketamine and 2.6 mg/ml xylazine (Phoenix Pharmaceuticals, St. Joseph, MO). Their spinal cords were dissected and processed as described above. Serial horizontal sections were made at 35 μm and mounted on charged slides in Vectashield. Fluoro-Gold-labeled motor neurons were hand counted using a fluorescence microscope with an ultraviolet filter set (Olympus BX51; Olympus, Center Valley, PA) and imaged with a cooled color CCD camera (CoolSNAP; Roper Scientific, Ottobrunn, Germany), and images were acquired via MetaMorph. Only cells in which a nucleus was seen were counted to prevent double counting.
Isolation of DNA for genotyping of animals was performed using MasterPure DNA purification kit (Epicentre Biotechnologies, Madison, WI). Tails were incubated at 65° C in 300 μl of tissue and cell lysis aolution containing 1 μl of 50 μg/μl proteinase K for 15 min, mixed on a vortex every 5 min. Samples were then cooled to 37° C and left to soak overnight. DNA was precipitated by the addition of 175 μl of MPC protein precipitation reagent, which was then vortex mixed and centrifuged at 10,000 g for 10 min. The pellet was discarded and the DNA was precipitated with 500 μl of isopropanol and a second microcentrifugation. The DNA was rinsed twice with 75% ethanol and then resuspended in 35 μl of sterile water. Genotyping was carried out via quantitative PCR according to the generic neomycin cassette (NEOTD) protocol provided by Jackson Laboratories.
A two-way ANOVA was used to compare time and experimental group in the experiments investigating neurotransmission failure. In the event of a significant ANOVA, a Student-Newan-Keul's post hoc analysis was used for pair-wise comparisons. For all other experiments, a one-way ANOVA was used to determine differences between experimental groups. Dunn's post hoc test was used for all pair-wise comparisons. Single-tailed Student t-tests were used to compare each dependent variable between TrkB+/− and WT littermates for contractile and Nav1.5 immunohistochemistry experiments. A single-tailed Student t-test was used because we predicted a direction for the hypothesized differences between TrkB+/− and WT littermates. All analyses were done using SigmaStat (Systat Software, San Jose, CA) and Excel (Microsoft, Seattle, WA). A P value ≤ 0.05 was considered to statistically significant.
Reduced TrkB expression results in fragmentation of AChRs and end-plate expansion.
Adult neuromuscular junctions are composed of numerous AChRs grouped into one or a few contiguous pretzel-shaped clusters. AChR end-plate regions have been shown to expand and become fragmented during the normal aging process (1, 11, 12), while expression of TrkB appears to decrease at the neuromuscular junction during senscence (32). Whether reduced TrkB expression causes or is an effect of these age-associated alterations in end-plate structure is unknown. Figure 1 shows confocal microscopy images of AChRs labeled with RαBTX from 6- to 8-mo-old TrkB+/− and WT mice, as well as 24-mo-old WT mice. The AChR end plates of 6- to 8-mo-old WT mice are made of only a few contiguous clusters of receptors with clearly delineated borders (Fig. 1, A and B). For example, Fig. 1A shows an end-plate region with one contiguous AChR cluster. However, the end-plate region in age-matched TrkB+/− littermates contains multiple clusters of receptors (Fig. 1, D–F), similar to that found in 24-mo-old C57BL/6 mice (Fig. 1C). Figure 1D shows an end-plate region with 14 AChR clusters.
The end-plate region in TrkB+/− junctions also expands in size and become less clearly delineated with variable AChR labeling intensity, giving them an overall blotchy appearance. To verify the architectural changes in the neuromuscular junction described above, we counted the number of discrete AChR regions (measurement of declustering or fragmentation) and measured end-plate perimeter and total AChR area. TrkB+/− and 24-mo-old mice demonstrated a similar increase in the number of AChR clusters compared with 6- to 8-mo-old WT mice (Fig. 1G). The 6- to 8-mo-old WT mice averaged 2.8 ± 0.2 clusters, while TrkB+/− littermates and 24-mo-old mice averaged 5.0 ± 0.4 and 5.3 ± 0.5 clusters, respectively (P < 0.001, one-way ANOVA; n = 87, 98, and 61 junctions, respectively).
End-plate size is also increased in 6- to 8-mo-old TrkB+/− and 24-mo-old WT mice compared with 6- to 8-mo-old WT littermates (Fig. 1H). AChR area is enlarged from 469 ± 17 μm2 in 6- to 8-mo-old WT mice to 552 ± 22 and 590 ± 25 μm2 in TrkB+/− mice and 24-mo-old mice, respectively (P = 0.002, One-way ANOVA with Dunn's post hoc test). The 6- to 8-mo-old TrkB+/− and 24-mo-old WT mice demonstrate increased AChR perimeter compared with 6- to 8-mo-old WT littermates, but increases are greater in senescent WT mice. AChR perimeter is increased from 248 ± 10 μm in 6- to 8-mo-old WT mice to 303 ± 12 and 432 ± 31 μm in TrkB+/− littermates and 24-mo-old mice, respectively (data not shown; P < 0.001, one-way ANOVA with Dunn's post hoc test).
In adult mice, there is a precise overlap between presynaptic nerve terminals and postsynaptic AChR clusters at the neuromuscular junction, but during aging, the precision of this overlap is disrupted as muscle fibers become morphologically denervated (2, 3, 12). The extent of denervation, however, is reduced in the slow-twitch soleus muscle compared with muscles with mixed fiber types (14, 49). We used two methods to measure morphological denervation across experimental groups (confocal microscopy of pre- and postsynaptic overlap and Nav1.5 expression). Figure 2 shows sample confocal images of presynaptic nerve terminals and postsynaptic AChR clusters in 6- to 8- and 24-mo-old WT and 6- to 8-mo-old TrkB+/− mice. Precise overlap between pre- and postsynaptic partners was seen in both TrkB+/− and 24-mo-old WT mice, despite postsynaptic fragmentation. Nerve terminals shown in green (Fig. 2, A–D) precisely overlap with their red postsynaptic targets (Fig. 2, E–H). The area of nerve/receptor overlap was pseudocolored white for easier visualization (Fig. 2, I-L). No significant difference in nerve/AChR colocalization was found between any experimental group, although overlap tended to be slightly reduced in 24-mo-old mice. (Fig. 2M; P = 0.707, one-way ANOVA; 53.4 ± 2.1 and 48.3 ± 10 vs. 53.1 ± 1.4% for 6- to 8- and 24-mo-old WT and 6- to 8-mo-old TrkB+/− mice, respectively).
Nav1.5 immunolabeling, however, was consistently increased along muscle fibers from 6- to 8-mo-old TrkB+/− compared with WT littermates (Fig. 3). A 5.5 ± 0.5% of muscle fibers had >50% of their membrane immunolabeled for Nav1.5 in 6- to 8-mo-old WT mice (n = 6 mice) compared with 9.6 ± 1.4% in age-matched TrkB+/− mice (P = 0.028, one-way Student t-test; n = 6 mice). The percentage of muscle fibers from 6- to 8-mo-old TrkB+/− mice with >50% of their membrane immunolabeled for Nav1. was similar to the percentage of denervated fibers previously reported in the soleus of 24-mo-old mice (14). Thus quantification verified that 6- to 8-mo-old TrkB+/− mice demonstrate age-like changes in synaptic structure and muscle fiber denervation similar to those seen in 24-mo-old mice.
Reduced TrkB expression increases neurotransmission failure at the neuromuscular junction:.
Aging is associated with impaired neuromuscular function (8, 47). We used an in vitro differential nerve stimulation vs. direct muscle stimulation protocol to determine whether physiological changes in synaptic function are seen in TrkB+/− mice compared with WT littermates. Figure 4A shows a sample tracing of force output where direct muscle stimulation at 100 Hz was intermittently superimposed every 15 s on 100-Hz trains of nerve stimulation. The difference in force production between stimulation forms represents the contribution of neuromuscular failure to muscle fatigue. The first three waveforms (nerve-muscle-nerve stimulation) serve as a control to demonstrate that force output between the two stimulation methods was initially similar. Neurotransmission failure increased over time for all experimental groups, but neurotransmission failure was significantly increased in TrkB+/− and 24-mo-old WT animals compared with 6- to 8-mo-old WT (P < 0.001, two-way ANOVA; n = 7, 3, and 8 respectively). Figure 4B shows the maximum neurotransmission failure measured during the 330-s stimulation protocol. Maximum neurotransmission failure was significantly increased in TrkB+/− and 24-mo-old WT mice compared with 6- to 8-mo-old WT mice (77.8 ± 2.5 and 78.9 ± 1.8% failure compared with 68.7 ± 2.0% failure, respectively; P = 0.012, one-way ANOVA). Thus reduced TrkB expression potentiates neurotransmission failure to a similar degree to that seen in 24-mo-old mice.
Reduced TrkB expression alters muscle contractile properties and reduces muscle fiber CSA.
To determine whether reduced TrkB expression affects muscle force production, we measured isometric force output at increasing frequencies of direct muscle stimulation (Table 1). Contraction kinetics (time to peak and half relaxation time) and twitch force were similar between all experimental groups, suggesting that TrkB expression does not alter the distribution of muscle fiber phenotypes. Maximal muscle force production (Po), however, trended downward from 207 ± 16 mN in 6- to 8-mo-old WT to 180 ± 26 mN in 24-mo-old to 172 ± 10 mN in TrkB+/− mice (P = 0.196, one-way ANOVA; n = 8, 3, and 10 mice, respectively). A significant difference in Po was seen when 6- to 8-mo-old TrkB+/− mice were compared with WT littermates (P = 0.037, one-tailed Student's t-test). Specific force (sPo), which incorporates the calculated muscle belly CSA, was decreased from 20.6 ± 1.5 N/cm2 in WT to 15.4 ± 1.2 N/cm2 in TrkB+/− littermates (P = 0.040, one-way ANOVA with Dunn's post hoc test). The sPo produced by 24-mo-old mice (18.9 ± 2.8 N/cm2) was similar to that previously reported (6). No differences in muscle length (lo) or calculated muscle belly CSA was seen between genotypes. Soleus muscle wet weight, however, was increased in TrkB+/− compared with young and old WT mice. This trend was unexpected considering the reduced force-generating capability of TrkB+/− mice but may be explained by the significantly increased body weight of TrkB+/− mice compared with WT mice (P = 0.020, one-way ANOVA with Dunn's post hoc test; 44.9 ± 3.6, 33.0 ± 2.2, and 34.4 ± 0.6 g for TrkB+/−, 6- to 8-mo-old WT, and 24-mo-old WT mice, respectively).
To further investigate the influence of reduced TrkB expression on muscle, the soleus muscle was cryosectioned and hematoxylin and eosin labeled (Fig. 5, A and B). Reduced TrkB expression did not affect muscle fiber number (P = 0.403, Student's t-test; 841 ± 30 vs. 777 ± 65 muscle fibers for TrkB+/− and WT mice, respectively; n = 4 for each genotype). Muscle fiber CSA was significantly reduced in TrkB+/− mice compared with WT littermates (Fig. 5C; P ≤ 0.001, Student's t-test; 1,124 ± 18 vs. 1,290 ± 31 μm2 for TrkB+/+ and WT mice, respectively; n = 4 mice of each genotype). Muscle fiber perimeter was similarly decreased (P = 0.009, Student's t-test; 139 ± 1.2 vs. 144 ± 1.7 μm2 for TrkB+/− and WT mice, respectively). Thus reduced TrkB expression attenuates muscle force production and decreases muscle fiber CSA.
Reduced TrkB expression does not affect motor neuron size or number.
To verify that 6–8 mo of constitutively reduced TrkB expression does not alter neuron number or morphology, lumbar spinal cord sections were immunolabeled for ChAT. The number of motor neurons per section and motor neuron cell body CSA was determined (Fig. 6, A–D). Only motor neurons with a visible nucleus were included in the CSA measurement. Motor neuron CSA was similar between genotypes (Fig. 6E; P = 0.495, t-test; 521 ± 39 vs. 490 ± 26 μm2, WT vs. TrkB+/−, respectively). Motor neuron number per section was also similar between WT and TrkB+/− littermates. (P = 0.885, Student's t-test; 11.0 ± 1.3 vs. 11.3 ± 1.0, respectively; n = 3 mice per genotype). Motor neurons were backlabeled using Fluorogold to determine the number of motor neurons specifically innervating the soleus muscle. The soleus muscle was innervated by a similar number of motor neurons in both genotypes (Fig. 7; P = 0.826, Student's t-test, 32.0 ± 4.2 vs. 30.7 ± 3.8; WT vs. TrkB+/−, respectively; n = 3 per genotype). Thus, constitutively reduced TrkB expression does not influence motor neuron cell body size or number in adult mice (30).
Constitutive reduction of TrkB expression affects both the structure and function of the neuromuscular junction. Decreased TrkB expression alters synaptic morphology by 1) dissociating AChRs from one or a few contiguous regions (clusters) into multiple less clearly delineated regions, 2) expanding overall AChR end-plate area and perimeter, and 3) increasing the number of denervated fibers, although presynaptic nerve-terminal/AChR overlay was similar across genotypes. Neuromuscular transmission failure was increased in TrkB+/− mice, indicating that reduced TrkB expression also alters synaptic physiology. Importantly, these changes in synaptic morphology and physiology occur in the absence of attenuated motor neuron number or size.
Constitutive reduction of TrkB expression moreover alters structure and function of the soleus muscle. The soleus muscle of TrkB+/− mice demonstrated reduced muscle fiber CSA but no change in muscle wet weight or fiber number. Physiological changes included reduced maximum tetanic force (Po) and specific Po (N/cm2), but no differences in contraction kinetics were found between genotypes. Thus reduced TrkB expression leads to smaller, weaker muscle fibers but does not appear to influence muscle fiber type or number.
Reduced TrkB expression recapitulates changes seen at the neuromuscular junction during aging.
Relatively early in the aging process, AChR area expands, fragments, the bugarotoxin-labeling intensity becomes variable, and the area of pre- and postsynaptic overlap becomes progressively smaller (2, 3, 11, 12), as the number of presynaptic vesicles decreases (19, 20). These structural changes influence synaptic function, since both the safety factor for neurotransmission and excitation-contraction coupling is reduced in aged rodents (8, 10, 22, 33, 34, 47, 50).
In the current study, reduced TrkB expression resulted in postsynaptic fragmentation, expanded AChR area, and variable AChR labeling intensity similar to that found in 24-mo-old muscle. Reduced TrkB expression resulted in a 1.8-fold increase in AChR regions, which is smaller than the threefold increase reported by Gonzalez et al. (17) in the sternomastiod muscle of TrkB+/− mice. This greater fold-increase is mostly likely due to the larger AChR area and cluster number seen in muscles with mixed fiber types compared with the slow-twitch soleus muscle used in the current study (46). Curiously, Gonzalez et al. (17) also reported that AChR area was reduced in the sternomastiod muscle of TrkB+/− mice. Our preliminary studies in the EDL muscle of adult TrkB+/− mice show expansion similar to that found for the soleus muscle. One potential reason for this discrepancy in AChR area is the young age of the mice used by Gonzalez et al. (17).
The slow-twitch soleus muscle undergoes limited denervation during aging (14, 49). Muscle fiber Nav 1.5 expression, a marker of denervation, was increased in 6- to 8-mo-old TrkB+/− mice to a comparable level to that previously reported in the soleus of 24-mo-old mice (14). However, the area of pre- and postsynaptic overlap was similar across genotypes. We most likely missed indentifying changes in pre- and postsynaptic overlap for two reasons. First, only ∼10% of soleus muscle fibers are denervated in TrkB+/− mice compared with 50% fiber denervation in the flexor digitorum brevis (50) or 25–50% fiber denervation in the gastrocnemius of senescent mice (14). Thus the chance of indentifying a denervated neuromuscular junction in the soleus was limited. Second, confocal images were obtained only when presynaptic structures were clearly immunolabeled and threshold intensities were set to visualize as many presynaptic terminals possible. Reduced TrkB expression resulted in muscle fiber denervation similar to that seen in 24-mo-old mice, but changes in changes in pre- and postsynaptic overlap were not found in part because we were biased towards imaging only well-innervated junctions.
Reduced TrkB expression recapitulates changes seen in slow-twitch muscle during aging.
The hallmarks of sarcopenia are a decline in muscle bulk and performance. Reduced muscle bulk is thought to be caused through a combination of muscle fiber atrophy secondary to reduced protein synthesis/reduced satellite cell activation and the gradual loss of fiber number secondary to denervation (7, 51). Impaired muscle performance is caused by a mixture of factors, including denervation, reduced reinnervation, the replacement of large force-producing type IIB muscle fibers by weaker type I muscle fibers, and slower fractional protein turnover resulting in longer retention of damaged contractile elements (51).
In the current study, reduced TrkB expression resulted in decreased tetanic force production and muscle fiber CSA but did not affect muscle fiber number, weight, or contractile kinetics. The preservation of muscle fiber number in TrkB+/− mice is unsurprising, since limited denervation is seen in the soleus muscle during aging (14, 49). Similarly, no change in contractile speed was expected, since time to peak tension and half relaxation time is uneffected in the soleus during aging (6). The 17% reduction in maximal tetanic force found in TrkB+/− compared with age-matched WT mice is similar to the 14–22% reduction in soleus force output previously reported in rodents during aging (6, 13), and the 13% reduction we found in 24-mo-old WT mice. The reduction in sPo, which incorporates the calculated muscle belly CSA, was greatest in 6- to 8-mo-old TrkB+/− mice due to their increased soleus muscle mass. The sPo produced by 24-mo-old WT mice was similar to previous reports (6, 13). Soleus muscle wet weight in C57BL/6 mice has been reported to be reduced from 10.3 to 8.27 mg in 9- to 26-mo-old mice (6). Our 6-mo-old mice soleus wet weight (10.1 mg) was similar to that found in adult mice by Brooks and Faulkner (6), but the soleus weight tended to be heavier in TrkB+/− mice (11.7 mg). This result may be explained, however, by the significant increase in body weight found in TrkB+/− compared with 6- to 8-mo-old WT mice (33.0 to 44.9 g, respectively). The body weight to soleus wet muscle weight ratio (g/mg) normally increases during aging from 2 to 9 to 26 mo (3.09 to 3.45 to 3.59 in C57BL/6 mice; Ref. 6). We found a ratio of 3.74 in 24-mo-old C57BL/6 mice along with an increase in this ratio from 3.27 to 3.84 when comparing 6-mo-old WT and TrkB+/− mice. Thus, when body weight is controlled, TrkB+/− mice demonstrate a greater relative loss of muscle mass than that seen in 24- to 26-mo-old mice, suggesting that muscle bulk may be relatively reduced in TrkB+/− mice compared with their age-matched WT littermates.
Model for TrkB-mediated signaling at the neuromuscular junction throughout the lifespan.
The findings presented here, in conjunction with previous studies, suggest a model regarding the role of TrkB-mediated signaling at the neuromuscular junction. During embryonic development, innervating motor neurons and muscle progenitors express high levels of BDNF, p75NTR, and TrkB, facilitating of neuromuscular synaptogenesis (15, 16, 54), while the subsequent downregulation of BDNF may promote myogenic differentiation and myotube formation (31). During postnatal development, NT-4/5 expression in muscle is increased and BDNF expression remains low, suggesting NT-4/5 autocrine signaling through TrkB may impact the postnatal upregulation of postsynaptic AChR expression (52).
During adulthood, we and others (17, 29, 45) have shown that TrkB-mediated signaling is necessary to maintain normal neuromuscular morphology and physiology and possibly muscle satellite cell population. However, the importance of TrkB-mediated signaling appears to be more significant in slow-twitch then fast-twitch muscle fibers, because 1) activity-dependent expression of NT-4/5 is more profound in the soleus than EDL muscle (15), 2) muscle force production and muscle endurance are impaired to a greater extent in the soleus muscle in NT-4 knockout mice (4), and 3) hindlimb suspension studies (which decrease muscle activity) show that function of slow-twitch muscle fibers is more affected than fast-twitch fibers even when levels of physical atrophy are similar between muscles (18).
It is likely that reduced TrkB-mediated signaling may underlie several of the age-associated changes that occur in the neuromuscular system, since reduced TrkB expression recapitulates many of the changes seen in the soleus muscle in senescence. Clinical treatment of sarcopenia is often aimed at maintaining greater force-producing type IIB muscle fibers, but the loss of type I fibers may have greater functional significance, since these slow-twitch fibers reinnervate denervated type IIB and IIB fibers and are used for most activities of daily living. Together, this suggests that exercise therapy may slow the onset of sarcopenia by maintaining adequate TrkB signaling for stabilization of the neuromuscular system in slow-twitch muscle fibers (21). In conclusion, TrkB appears to play a role in synaptic stabilization and synaptic potentiation of the neuromuscular junction. Reduction of this receptor elicits significant impairments in synaptic stabilization, neurotransmission, and muscle contractile properties. Thus, reduced TrkB expression results in precocious neuromuscular aging of the predominately slow-twitch soleus muscle, while methods to maintain TrkB-mediated signaling may be a potential treatment for sarcopenia.
We are grateful to the National Skeletal Muscle Research Center and the University at Buffalo (to K. E. Personius) for financial support.
No conflicts of interest, financial or otherwise, are declared by the author(s).
We thank Dr. Wade Sigurdson of the Confocal Microscope and Flow Cytometry Facility in the School of Medicine and Biomedical Sciences, University at Buffalo, for technical assistance and Drs. Burkard, Stachowiak, and Slaughter for reading earlier versions of this manuscript.
- Copyright © 2011 the American Physiological Society