Neuromuscular Junction Defects in Mice with Mutation of dynein heavy chain 1

Disruptions in axonal transport have been implicated in a wide range of neurodegenerative diseases. Cramping 1 (Cra1/+) and Legs at odd angles (Loa/+) mice, with hypomorphic mutations in the dynein heavy chain 1 gene, which encodes the ATPase of the retrograde motor protein dynein, were originally reported to exhibit late onset motor neuron disease. Subsequent, conflicting reports suggested that sensory neuron disease without motor neuron loss underlies the phenotypes of Cra1/+ and Loa/+ mice. Here, we present behavioral and anatomical analyses of Cra1/+ mice. We demonstrate that Cra1/+ mice exhibit early onset, stable behavioral deficits, including abnormal hindlimb posturing and decreased grip strength. These deficits do not progress through 24 months of age. No significant loss of primary motor neurons or dorsal root ganglia sensory neurons was observed at ages where the mice exhibited clear symptomatology. Instead, there is a decrease in complexity of neuromuscular junctions. These results indicate that disruption of dynein function in Cra1/+ mice results in abnormal morphology of neuromuscular junctions. The time course of behavioral deficits, as well as the nature of the morphological defects in neuromuscular junctions, suggests that disruption of dynein function in Cra1/+ mice causes a developmental defect in synapse assembly or stabilization.


Introduction
Rapid axonal transport is essential for neurons with long axons, both for development and mature functioning. Without transport, neurotransmitter vesicles cannot move to synapses, damaged organelles in the axon cannot be replaced by healthy organelles, and target-derived survival signals cannot reach the cell body. Because of the importance of axonal transport, it is not surprising that disruptions in this process have been implicated in a wide range of neurodegenerative diseases, including motor and sensory neuropathies, Alzheimer's disease, Parkinson's disease, and Huntington's disease [1][2][3][4]. Perhaps because of the long distance between cell bodies and distal axons in these neurons (over 1 meter in an adult human), sensory and motor neurons appear to be particularly sensitive to disruptions in axonal transport.
Disruptions in retrograde transport have been implicated in neurodegenerative disease in mice as well as in humans. Disrupting the function of the retrograde motor dynein in motor neurons by overexpression of dynamitin causes late-onset motor neuron degeneration in mice [5,6]. A mutation in the p150 Glued subunit of the dynactin complex has been discovered in a family with a progressive autosomal dominant form of lower motor neuron disease, and mice with mutations in p150 Glued develop motor neuron disease [7,8,6].
In 2003, Fisher and colleagues reported two lines of mice with late onset, progressive motor neuron disease, called Legs at odd angles (Loa/+) and Cramping 1 (Cra1/+) [9]. Both Loa/+ and Cra1/+ mice have a missense mutation in the dynein heavy chain 1 gene [9]. Heterozygous mice from both lines were reported to show an agerelated, progressive motor neuron disease, with cramping of hindlimbs, gait abnormalities, and eventual paralysis. While the number of motor neurons was normal in Cra1/+ mice at three months of age, a severe reduction in motor neuron number was reported at 16 months of age in Cra1/+ mice and 18 months of age in Loa/+ mice, consistent with the progressive symptoms. In vitro, neurons from Loa/Loa embryos exhibited a defect in the fast component of retrograde transport without other alterations in the ubiquitous functions of dynein, such as nuclear motility and Golgi complex morphology and positioning. Furthermore, purified dynein from Loa/+ and Loa/Loa mice exhibited reduced processivity, premature run termination and altered coordination of motor domains. These defects together lead to impaired retrograde axonal transport [10]. Thus, it was concluded that neuron degeneration in these mice arises from specific disruptions in rapid retrograde axonal transport [9,10].
In contrast, two recent studies provide evidence that the Loa/+ and Cra1/+ mice exhibit a sensory neuropathy without motor neuron involvement. Electromyography recordings, motor neurons counts, and neuromuscular junction analysis in aged (18-22 months) Loa/+ and Cra1/+ mice all demonstrate normal motor neuron numbers and innervation of muscle [11,12]. Both lines of mice were reported to exhibit an early-onset sensory neuropathy, with reduced numbers of large caliber proprioceptive axons in the dorsal root and reduced innervation of muscles spindles. These results are similar to the phenotype reported for Sprawling (Swl/+) mice, another hypomorphic mutation in dynein heavy chain 1 that exhibits proprioceptive deficits and loss of proprioceptive neurons [13].
In order to clarify the link between disruption in axonal transport and neuronal functioning we undertook a detailed examination of the behavioral and anatomic phenotype in Cra1/+ mice. Behavioral analysis revealed an early onset, stable phenotype in Cra1/+ mice, with hindlimb cramping and impaired grip strength. We find no evidence of motor or sensory neuron loss in the Cra1/+ mice at symptomatic ages. Strikingly however, neuromuscular junctions in Cra1/+ mice are abnormal. While the number and size of these synaptic zones are unchanged, the complexity of the pretzel-like NMJ is significantly reduced in mutant animals. Furthermore, analysis of muscle fibers provides evidence of continuous muscle degeneration and regeneration. These results suggest that the primary pathology in Cra1/+ animals may be an early onset, non-progressive synaptic dysfunction that affects the neuromuscular junction.

Cra1/+ Mice Show an Early Onset, Stable Phenotype of Neuronal Dysfunction
Cra1/+ mice were originally reported to exhibit late onset, progressive degeneration of motor neurons [9]. In order to characterize disease progression in these mice, we undertook a long-term analysis of Cra1/+ mice. We find that Cra1/+ mice show reduced survival compared to wild-type (+/+) mice ( Figure 1A). The cause of death was not evident in the majority of cases. We did not observe hindlimb paralysis, increased seizure activity, or other signs of nervous system or motor neuron dysfunction in the majority of animals that died prematurely. The original identification of the Cra1/+ line of mice was a result of their unusual ''cramping'' posture of the hindlimbs when suspended by the tail. We documented the emergence of this cramping posture in Cra1/ + mice between the ages of 15 days old and 12 months old by scoring animals as either 1 (no hindlimb abnormalities), 2 (cramping of hindlimbs), or 3 (paralysis of one or both hindlimbs). Surprisingly, we find that Cra1/+ mice demonstrate pronounced cramping posture as early as 15 days of age, the earliest time that could be tested ( Figure 1B).
In order to assess motor neuron function in a quantitative fashion in Cra1/+ mice, we used a digital grip strength meter to record the power used to resist an opposing force with either foreor hindlimbs. This behavioral test requires the mice to grip a wire grid attached to a force gauge and maintain their grip while they are pulled away from the grid, necessitating both proper gripping with the fore-and hindfeet as well as pulling with the limb muscles. Measurement of both fore-and hindlimb grip strength shows an early onset, stable reduction in grip strength compared to +/+ animals ( Figure 1C). Interestingly, Cra1/+ mice show stable grip strength measurements over time, indicating that the dysfunction in grip strength does not progress with age.

Cra1/+ Mice Do Not Show Evidence of Motor or Sensory Neuron Loss at Symptomatic Ages
To investigate motor neuron degeneration in Cra1/+ mice, we examined motor neuron cell bodies in the spinal cord for evidence of pathology. Choline acetyltransferase (ChAT) labeling of motor neuron cell bodies in the ventral horn of the lumbar spinal cord in 6 and 24 month old Cra1/+ and +/+ mice demonstrates no observable difference in motor neuron morphology or ChAT expression in Cra1/+ mice (Figure 2A). Quantification of ChAT labeled motor neurons reveals no difference in neuronal number between Cra1/+ and +/+ mice ( Figure 2B). The overall appearance and number of anterior horn motor neurons is also normal in 6 and 24 month old Cra1/+ mice as observed by standard histochemistry with toluidine blue ( Figure 2C, 2D). As 6 months is long after the emergence of abnormal hindlimb posture (15 days of age) and impairments in hindlimb grip strength (3 months of age), this result indicates that the observed neuronal dysfunction is not a result of motor neuron cell body loss.
Recent studies suggested that neurologic deficits in Cra1/+ and Loa/+ mice are instead due to loss of proprioceptive sensory neurons [11,12], a defect that would also result in the observed phenotypes of hindlimb cramping and reduced grip strength [14]. Using several markers for DRG proprioceptive neurons, including parvalbumin, ER81, and TrkC [15,16], we observe no difference in proprioceptive neuron morphology or number between Cra1/+ and +/+ mice at 6 months of age ( Figure 3A, 3B, 3C, respectively). Additionally, we examined the connection between proprioceptive afferents and primary motor neurons within the spinal cord. Labeling of these fibers in the spinal cord with parvalbumin revealed that proprioceptive fibers innervate motor neurons in the spinal cord to a similar extent in both Cra1/+ and +/+ mice ( Figure 3D). Thus, proprioceptive neuron cell bodies, as well as their central projections, are intact in symptomatic Cra1/+ mice.

Cra1/+ Mice Exhibit Reduced Complexity of Neuromuscular Junctions
Although motor neuron cell bodies appear to be spared in symptomatic Cra1/+ animals, synaptic dysfunction at the neuromuscular junction (NMJ) could potentially underlie the observed behavioral abnormalities. This possibility would be consistent with genetic data from Caenorhabditis elegans and Drosophila melanogaster, where mutations in dynein components cause defects in synapse assembly [17][18][19][20]. Therefore, we examined neuromuscular junction synapses in 2 month and 6 month old Cra1/+ and +/+ mice using abungarotoxin labeling of the tibialis anterior muscle ( Figure 4A). In Cra1/+ mice, there is no reduction in the number of NMJs ( Figure 4B). We examined the ''total area'' of each NMJ, defined by the area contained within a boundary encompassing the outermost edges of the NMJ, as revealed by a-bungarotoxin label ( Figure 4Ai). There is no difference in NMJ total area between Cra1/+ and +/+ at 2 months and 6 months of age ( Figure 4C). However, because of the complex, pretzel-like structure of the NMJ, total area contained within the outside boundary does not accurately reflect the ''synaptic area'' of pre-and post-synaptic alignment where intercellular communication occurs. Therefore, we also assessed the synaptic area by quantifying the area within each NMJ that is labeled by a-bungarotoxin (Figure 4Aii). The synaptic area within each NMJ is significantly reduced in Cra1/+ mice at 2 and 6 months of age ( Figure 4D) suggesting a reduction in complexity of the postsynaptic component. To assess the presynaptic zone of the NMJ synapse, we imaged motor neuron axons as they innervate the NMJ, using mice expressing GFP under the control of the HB9, motor neuron-specific promoter [21]. When we costained the NMJs with a-bungarotoxin and anti-GFP ( Figure 5A), we found that there is decreased innervation of NMJs in Cra1/+ mice ( Figure 5B). Similar results were seen when we co-stained with a-bungarotoxin and neurofilament to visualize presynaptic motor neurons. Notably, quantitative analysis of presynaptic innervation of NMJs revealed that in Cra1/+ animals, there are fewer NMJs that exhibit complete overlap of the presynaptic axon and the post-synaptic acetylcholine (ACh) receptors detected by a-bungarotoxin stain. Instead, the proportion of NMJs that exhibit partial overlap between the presynaptic axon and the post-synaptic ACh receptors is increased in mutant animals. Such partial overlap has also been observed in mouse models of spinal muscular atrophy [22], and suggests that there is instability of the presynaptic components of the NMJ. Together these data provide evidence of defects in both the pre and post-synaptic components of the mutant NMJs.
During adult life, muscle fibers can degenerate and be replaced by new muscle fibers; higher rates of muscle fiber turnover are seen in response to nerve degeneration or to NMJ dysfunction. As centrally localized nuclei mark the presence of regenerating muscle fibers [23,21] we quantified centrally localized nuclei in the tibialis anterior muscles of 6 month old Cra1/+ and +/+ mice. Figure 5C demonstrates that there are more muscle fibers with central nuclei in Cra1/+ animals, indicating increased muscle turnover. Such changes in muscle fiber could be a consequence of NMJ dysfunction, or might be a very early indicator of motor neuron axon degeneration and regeneration.
The results presented here suggest a novel mechanism to explain the behavioral abnormalities found in Cra1/+ mice. We find evidence of an early-onset, stable phenotype of hindlimb cramping and reduced grip strength. At symptomatic ages, we find reduced morphological complexity in the neuromuscular junction. This pathology occurs in the absence of either motor or sensory neuron loss, suggesting that a developmental synaptic dysfunction rather than an adult onset neurodegenerative syndrome underlies the behavioral deficits in Cra1/+ mice.

Discussion
Previous reports on mice with hypomorphic mutations in dynein, the Cra1/+, Loa/+ or Swl/+ mice, described a late-onset degeneration of motor neurons, with progressive decline in motor neuron function and number [9], or an early-onset degeneration of proprioceptive sensory neurons. Our recurrent behavioral assessment of Cra1/+ mice shows an early onset, stable phenotype of hindlimb cramping and impaired grip strength, with no further progression in adulthood over the 24 months of testing. We examined the anatomical underpinnings of this behavior. We find no loss of either motor or sensory neurons at symptomatic ages; instead, our analysis identifies decreased complexity of neuromuscular junction synapses, with associated muscle fiber abnormalities.
Dynein and its associated dynactin complex have been shown to be important for synapse formation and stabilization in both Drosophila and C. elegans [17,11,19,20]. In C. elegans, mutations in genes encoding components of the dynein motor, including dynein heavy chain, results in abnormal assembly of synapses. Additionally, these mutants exhibit intracellular protein aggregations and behavioral abnormalities including locomotor impairment, and shorter lifespan [19]. Drosophila with mutations in genes encoding components of the dynein/dynactin complex, as well as those with mutations in dynein light chain genes, exhibit defects in synapses of sensory and motor neurons [17][18][19][20]. These studies indicate that dynein functions in the presynaptic terminal to promote synapse stabilization and prevent synapse elimination [18].
Our findings of abnormal NMJ complexity in Cra1/+ mice suggests that dynein plays a role in development and maintenance of mammalian synapses as well. The morphological changes in NMJ synapses and the associated alterations in muscle fibers, as well as the early onset of behavioral deficits in Cra1/+ mice, suggest that synapse formation, stability and refinement are altered by defects in dynein function. Thus, the Cra1/+ mice provide a useful model to examine the role of dynein in the dynamic processes of NMJ synapse formation and stabilization.
It is currently unclear why there are such large discrepancies in the phenotypes of the Cra1/+ and Loa/+ mice as described by different groups. Clearly these mutations in the dynein heavy chain gene affect the lower motor neuron circuitry comprising the primary motor neurons, the innervated muscle and the proprioceptive neurons that innervate the muscles. It is apparent that deterioration in a particular portion of this circuitry can ultimately result in dysfunction of the entire circuit. However, it has not been clear where or how dynein mutations initiate abnormalities in this circuit.
Our behavioral analysis reveals that the mice exhibit deficits from an early age, while anatomical analysis indicates that disruption of NMJ synaptic structure is present and precedes any motor or sensory neuron loss. We did not observe any evidence of late degeneration; however, it is possible that early disruption in synaptic function at the NMJ could prime motor neuron or sensory neuron cell bodies for later degeneration under some conditions [11,9,12]. Thus, it may be important to consider the effect of environment on the manifestation of phenotypes in these mice. Recent evidence suggests that exercise and calorie intake can dramatically affect age-related changes in neuromuscular junctions in mice [24]. Since we find that one of the primary sites of dysfunction in Cra1/+ mice is at the NMJ, variations among laboratories in housing conditions and daily food intake will impact the emergence of this phenotype and may account for some of the divergent reports of phenotypes.
Intriguingly, a recent study by Braunstein et al. finds striatal atrophy, abnormal dopamine signaling, and impaired neurite outgrowth of striatal neurons in Cra1/+ mice. These data suggest that under some conditions dynein mutations may affect multiple circuits in the central as well as the peripheral nervous system [25].
In conclusion, we provide evidence for an early-onset, stable phenotype in Cra1/+ mice. In contrast to previous reports, we find no evidence of either motor or sensory neuron loss at symptomatic ages. Instead, we find evidence of altered topology of NMJs, in that the area within each synapse that is covered by acetylcholine receptors is reduced, together with a reduction in the region of overlap with the presynaptic axon terminal. We propose that disruption of dynein function in Cra1/+ mice results in abnormalities of neuromuscular junction synapse formation, refinement, and/or stabilization. Therefore, Cra1/+ mice may prove to be a useful tool to examine the process of synaptogenesis and maturation at the mammalian NMJ.

Animal Use
Cra1/+ mice were purchased under an MTA agreement from Ingenium Pharmaceuticals. Genotyping for the Cra1 point mutation was performed by Transnetyx, Inc. (Cordova, TN) using a wild type or Cra1 targeting sequence (TGTGTTTCCTTTGCAGGTTTG-GCTTCAGT[A/G]CCAGTGTTTGTGGGATATGCAGGCA-GAAAACATTTACAACAGGCTAGGGGAA). HB9-GFP mice [21] were obtained from Jackson Laboratory. To obtain Cra1/+ mice expressing GFP in motor neurons, HB9-GFP heterozygous mice (HB9-GFP/+) were crossed with Cra1/+ mice to obtain Cra1/+ HB9-GFP/+ mice. Only doubly-heterozygous offspring from the original Cra1/+ and HB9-GFP/+ cross were used. All experimental procedures were done in accordance with the National Institutes of Health guidelines and were approved by the Dana-Farber Cancer Institutional Animal Care and Use Committee. The permit number is 03-054.

Behavior Analyses
Kaplan-Meier survival curves were generated using Prism statistical software, and analyzed using the Curve Comparison analysis for survival data.
Cramping behavior was measured by lifting the animal gently by the tail, suspending the hindlimbs and leaving forelimbs touching the ground. Two independent observers scored the animal as 1 (no hindlimb abnormalities), 2 (cramping of hindlimbs), or 3 (paralysis of one or both hindlimbs). Scorers were blind to genotype.
Grip strength was measured using a digital grip strength meter, which records the maximal strength an animal exerts while trying to resist an opposing pulling force (Columbus Instruments, Columbus, OH). Forelimb grip strength was measured using a mouse tension bar, and hindlimb grip strength was measured using a mouse compression bar. The results of three consecutive trials on the same day were averaged for each animal. Significance was calculated by Student's two-tailed t-test. Statistical tests were performed at every third time point (3 months, 6 months, etc.). Testers were blind to genotype.

Motor Neuron Labeling
Cra1/+ and +/+ mice were euthanized and perfused with PBS, followed by 4% paraformaldehyde (PFA). The lumbar enlargement of the spinal cord was removed, post-fixed in 4% PFA overnight at 4uC, then cryopreserved in 30% sucrose at 4uC overnight. Spinal cords were frozen at 220uC, and sliced at 14 mm. For ChAT staining, slices were blocked in 10% normal donkey serum and 0.1% TritonX-100 in PBS for 1 hour at room temperature, then incubated in goat anti-choline acetyltransferase (ChAT) antibody (1:100, Chemicon, Cat# AB144P) overnight at 4uC. Tissue was then incubated in biotinlyated donkey anti-goat antibody (1:300, Vector Laboratory), followed by streptavidin conjugated Alexa 546 (1:500, Invitrogen), then counterstained with DAPI (1:1000) for 1 minute. Images were obtained with a 40X oil objective using NIS Elements software. ChAT labeled motor neurons were quantified by counting neurons containing a nuclear profile; 8 sections of the spinal cord anterior horn per animal, from three sets (Cra1/+ and +/+) of animals at each age were analyzed. For toluidine blue staining, slices were dried on a hot plate and a drop of filtered toluidine blue solution (1% toluidine blue O (basic blue 17), 1% sodium borate) was added to the slide and then rinsed in water when the solution began to dry. Images were obtained with a 10X objective using NIS Elements software. Toluidine blue labeled motor neurons were quantified by counting only neurons greater than or equal to 600 mm 2 in size; 8 sections of the spinal cord anterior horn per animal, from three pairs (one Cra1/+ and one +/+) of animals at each age were analyzed.
Proprioceptive sensory neurons were quantified by counting 8 sections per animal, from three sets (Cra1/+ and +/+) of animals.  C) Quantification of a-bungarotoxin labeled NMJs demonstrates no difference in total area (as measured by delineating outer edge and measuring area within that boundary, area analyzed is within white trace of 4Ai) of NMJs in Cra1/+ mice (N = 3 animals per genotype). D) NMJ complexity, as measured by the synaptic area of a-bungarotoxin label (4Aii), is significantly decreased in 2 month and 6 month old Cra1/+ mice (*P,0.05; N = 3 mice per genotype). doi:10.1371/journal.pone.0016753.g004 Number of labeled neurons containing a nucleus were counted, and compared to the total number of nuclei in the section to calculate the percentage of sensory neurons that are proprioceptors.
For labeling of proprioceptive fibers within the spinal cord, Cra1/+ and +/+ mice were euthanized and perfused with PBS, followed by 4% PFA, and sections of the lumbar enlargement of the spinal cord were prepared as described above. Sections were blocked in 10% normal goat serum and 0.1% TritonX-100 in PBS for 1 hour at room temperature, then incubated in rabbit antiparvalbumin (Swant, 1:1000, Cat# PV-28) overnight at 4uC. Tissue was then incubated in goat anti-rabbit Alexa-546 (Molecular Probes, 1:1000) for 1 hour at room temperature, and then counterstained with DAPI. Three pairs (one Cra1/+ and one +/+) of animals were examined.

Neuromuscular Junction Labeling
Cra1/+ and +/+ mice were euthanized and perfused with PBS, followed by 4% PFA. The tibialis anterior muscle was removed, post-fixed in 4% PFA overnight at 4uC, then cryopreserved in 30% sucrose at 4uC overnight. Muscles were frozen at 220uC, and sliced longitudinally at 40 mm. Muscle tissue was blocked in 3% bovine serum albumin, 2.5% normal goat serum, and 0.1% TritonX-100 in PBS for 1 hour at room temperature, then incubated in a-bungarotoxin-tetramethylrhodamine (1:500, Molecular Probes, Cat# T1175) overnight at 4uC, then counterstained with DAPI. Confocal images were obtained using a Zeiss LSM 510 META upright confocal microscope, using a 40X air objective.
Measurements of overall NMJ area were determined by delineating the outside edge of the a-bungarotoxin labeled NMJ, then measuring the total area within that outline. For each synapse, measurements of complexity were obtained by measuring the area of fluorescent signal that is above background and is within the outline of the NMJ. NMJ analysis was done using Metamorph software by counting 8 sections of the tibialis anterior muscle per animal from three pairs (one Cra1/+ and one +/+) of animals at each age.

Muscle Fiber Labeling
Cryostat sections of the tibialis anterior muscle from Cra1/+ and +/+ mice were prepared and stained as above. Centrally localized nuclei were counted and compared to the total number of nuclei in the section to calculate the percentage regenerating muscle fibers. Quantification was done by using 8 sections of the tibialis anterior muscle per animal from three pairs (one Cra1/+ and one +/+) of animals.