An increasing number of neurodegenerative diseases are being linked to mutations in genes encoding proteins required for axonal transport and intracellular trafficking. A mutation in p150 Glued , a component of the cytoplasmic dynein/dynactin microtubule motor complex, results in the human neurodegenerative disease distal spinal and bulbar muscular atrophy (dSBMA). We have developed a transgenic mouse model of dSBMA; these mice exhibit late-onset, slowly progressive muscle weakness but do not have a shortened lifespan, consistent with the human phenotype. Examination of motor neurons from the transgenic model reveals the proliferation of enlarged tertiary lysosomes and lipofuscin granules, indicating significant alterations in the cellular degradative pathway. In addition, we observe deficits in axonal caliber and neuromuscular junction (NMJ) integrity, indicating distal degeneration of motor neurons. However, sciatic nerve ligation studies reveal that inhibition of axonal transport is not evident in this model. Together, these data suggest that mutant p150 Glued causes neurodegeneration in the absence of significant changes in axonal transport, and therefore other functions of dynein/dynactin, such as trafficking in the degradative pathway and stabilization of the NMJ are likely to be critical in maintaining the health of motor neurons.
Disruption of cellular transport is implicated in multiple neurodegenerative diseases, an increasing number of which are due to mutations in genes coding for motor and cytoskeletal proteins ( 1 ). Distal spinal and bulbar muscular atrophy (dSBMA) with vocal fold involvement is caused by an autosomal dominant point mutation in the p150 Glued subunit of dynactin, an activator of the retrograde motor protein cytoplasmic dynein ( 2 ). The G59S point mutation occurs in the highly conserved, glycine-rich cytoskeleton-associated protein (CAP-Gly) domain of the p150 Glued polypeptide, which is known to interact directly with microtubules ( 3 ). The predicted structure of the CAP-Gly domain suggests that the introduction of the G59S mutation induces a conformational change leading to an enhanced tendency for the polypeptide to misfold and aggregate ( 2 , 4 ). Indeed, aggregates composed of dynactin and dynein are detected in motor neurons of patients with dSBMA ( 5 ). In addition, both motor neuron loss and decreased neuropil density are observed in patient tissues. Clinically, patients with the G59S mutation present with a slowly progressing phenotype that begins with inspiratory stridor, followed by distal muscle weakness ( 5 ).
Analysis of fibroblasts cultured from patient tissue as well as of mammalian cells transfected with the G59S polypeptide have revealed defects that suggest that the mutation induces both an inhibition of dynein/dynactin function and a toxic gain of function ( 4 ). Mutant G59S p150 Glued has a decreased affinity for microtubules and the microtubule plus-end binding protein EB1. Patient-derived fibroblasts showed delayed recovery after cellular stress induced by microtubule depolymerization, consistent with a loss of dynein/dynactin function. However, the presence of prominent dynein and dynactin-positive aggregates and entrapped mitochondria in transfected cells argues for a toxic gain of function that may disrupt degradative pathways and/or metabolic function ( 4 ). The effects of the G59S mutation are more pronounced in neuronal cell lines, suggesting that an in vivo model may reveal alterations to neuronal health not apparent in cell culture and clarify the relative contributions of loss of normal dynein/dynactin function and toxic gain of function, due for example to protein misfolding/aggregation, to primary pathogenesis.
Numerous lines of evidence support the idea that defects in dynein function can lead to neuronal dysfunction and death. In Cra1 and Loa mice, point mutations in the heavy chain of the retrograde axonal motor cytoplasmic dynein (DHC) lead to a neurodegenerative phenotype ( 6 ). Disruption of the association between dynein and its activator dynactin also results in progressive neurodegeneration in mice ( 7 ). These phenotypes have been interpreted primarily as resulting from induced deficits in retrograde axonal transport, as cytoplasmic dynein is the major molecular motor driving transport from the cell periphery to the cell body. However, dynein and dynactin are essential for multiple cellular functions, including trafficking of endosomes, lysosomes and mitochondria ( 1 ). Dynactin has also been suggested to be essential to maintain the stability of the neuromuscular junction (NMJ) ( 8 ).
Here, we examine the cellular effects of mutant p150 Glued expression in a transgenic mouse model of dSBMA. This model is characterized phenotypically by slowly progressive muscle weakness. At the cellular level, we see the enlargement and proliferation of lysosomes and lipofuscin granules in comparison with littermate controls. In addition, we observe alterations in the axonal caliber of motor neurons and disruptions in the morphology of NMJs, indicating distal changes in motor neurons. Surprisingly, however, we do not see a significant inhibition of retrograde axonal transport, suggesting that other dynein/dynactin-driven processes are critical in maintaining neuronal health.
Low-level expression of mutant p150 Glued expression in a transgenic mouse model leads to slowly progressive muscle weakness
To examine the cellular consequences of mutant p150 Glued expression in a novel model of the human neurodegenerative disease dSBMA, we developed several lines of transgenic mice expressing human p150 Glued with the G59S point mutation, fused to a C-terminal Myc-tag and driven by the Thy 1.2 expression cassette, which drives postnatal, neuronal-specific expression primarily in motor neurons ( 7 ). Of the two transgenic lines identified that stably express the transgene (Fig. 1 A), we focused on the M20 line that displays higher levels of transgene expression in the spinal cord as determined by reverse transcription–polymerase chain reaction (RT–PCR) (data not shown), referred to here as the Tg G59S line.
To verify that the transgenic protein was expressed in brain and spinal cord, western blots from Tg G59S mice and non-transgenic littermate controls were probed with anti-myc antibody (Fig. 1 B). These blots suggest that the transgene is expressed at low levels. Quantitative western blotting was performed to compare overall levels of p150 Glued expression in spinal cord extracts from transgenic and non-transgenic littermates. Two different antibodies were used to differentiate between endogenous and transgenic protein: a monoclonal antibody that does not detect the mutant form of p150 Glued , and a polyclonal antibody that detects both the normal and G59S forms ( 4 ). Overall levels of transgene expression were low in both spinal cord and brain. Using a recombinant myc-tagged p150 Glued polypeptide as a calibration standard, we estimate that in spinal cord extracts expression of the myc-tagged transgene was ∼13% of total levels of p150 Glued ; due to the cell-specific expression of the Thy1.2 promoter, the relative expression level of the transgene is likely to be somewhat higher in motor neurons.
Subunits of dynein and dynactin isolated from wild-type tissue co-sediment at high S values (19–20S), while disruption of this complex results in the partitioning of the subunits into distinct fractions on a sucrose density gradient ( 7 ). To determine if the low level of transgene expression observed in Tg G59S mice disrupts the integrity of the endogenous dynein/dynactin complex, we fractionated high speed cytosolic extracts from the brains of 3-month-old Tg G59S and non-transgenic control mice by sucrose density gradient centrifugation. In extracts from both Tg G59S mice and non-transgenic littermate controls, the dynein subunits DHC (dynein heavy chain) and DIC (dynein intermediate chain) co-fractionated with dynactin subunits p150 Glued and dynamitin (p50), demonstrating that transgene expression does not disrupt the integrity of the dynein/dynactin complex (Fig. 1 C). These results are consistent with previous observations from fibroblasts or lymphoblasts from patients expressing the G59S mutation in dynactin, in which dynein/dynactin also remains intact ( 4 ). Further, the myc-tagged transgene co-fractionates with other subunits of dynein and dynactin (Fig. 1 C, bottom panels), indicating that the mutant p150 Glued is incorporated into endogenous dynein/dynactin complexes.
To characterize the phenotype of Tg G59S mice, we used grip strength assays to look for quantitative evidence of muscle weakness. Forelimb grip strength was assessed monthly from 2 months until 18 months of age in a cohort of age-matched transgenic and non-transgenic littermates. A modest but consistent decline in grip strength was seen as early as 2 months after birth in Tg G59S females; this difference did not reach significance until 10 months of age (Fig. 1 D). Deficits in male Tg G59S mice were more subtle, perhaps because of increased body mass observed in male transgenic mice relative to age-matched controls (data not shown). Less quantitative measures, such as wire-hang tests and foot-printing analyses did not reveal statistically significant differences between Tg G59S mice and non-transgenic littermates, consistent with a relatively mild phenotype of progressive muscle weakness. Notably, lifespan was not shortened in Tg G59S mice. Taken together, these data indicate a mild, slowly progressive, non-fatal motor impairment similar to that observed in human dSBMA patients expressing the G59S mutation ( 5 ).
G59S transgene expression does not disrupt axonal transport but does lead to the proliferation of lysosomes in motor neurons
Cytoplasmic dynein and dynactin drive retrograde axonal transport ( 1 ). To determine if the expression of the mutant transgene impairs axonal transport, we performed a double ligation assay on sciatic nerves of Tg G59S mice and non-transgenic littermates. If axonal transport was functioning normally, kinesin would accumulate proximal to the ligatures while dynein and dynactin would accumulate both proximally and distally ( 9 ). Proximal accumulation of kinesin was clearly seen in both Tg G59S animals and controls (Fig. 2 ). We also observed that dynein and dynactin accumulated on both sides of the ligature site with no significant difference observed between Tg G59S mice and controls (Fig. 2 ), indicating that expression of the G59S transgene does not significantly affect axonal transport in this model. In contrast to these observations, an ∼50% inhibition in retrograde transport has been observed in similar experiments in Loa or Tg dynamitin mice, both models of slowly progressive neuronal degeneration due to either a point mutation in dynein or overexpression of a dynamitin transgene, respectively (Perlson and Holzbaur, submitted for publication). Thus, mouse models with somewhat similar phenotypes of mild motor impairment resulting from defects in related proteins demonstrate significantly different mechanisms at the cellular level.
We examined the expression and localization of the G59S polypeptide at the cellular level in tissue sections from the Tg G59S model, in comparison with sections from age-matched controls. Low-level expression of the myc-tagged transgene was observed in motor neurons from the spinal cord in the Tg G59S model; only background levels of staining are seen in age-matched controls (Fig. 3 A). Double-label immunofluorescence using antibodies to the myc tag and p150 Glued show co-localization throughout the soma and processes of motor neurons (Fig. 3 B). While prominent aggregates of dynein and dynactin are seen in neurons from dSBMA patients ( 5 ), we do not see consistent formation of aggregates in 9 month-old Tg G59S mice. We also examined the localization of the dynactin subunit dynamitin and the dynein subunit DIC (Fig. 3 C), which are uniformly distributed throughout the cell soma and processes in both Tg G59S mice and non-transgenic controls.
Overexpression of mutant G59S p150 Glued in cell lines leads to the formation of perinuclear aggregates of dynein/dynactin not seen at the lower expression levels in the Tg G59S model, as well as the sequestration of mitochondria within those aggregates ( 4 ). To examine the effects of expression of the mutant polypeptide on intracellular organelles in the Tg G59S mouse model, spinal motor neurons were examined at higher resolution using electron microscopy. Motor neurons from transgenic mice were overtly normal, and abnormal sequestration of mitochondria to the perinuclear region was not observed in comparison with age-matched non-transgenic controls.
However, a prominent feature of motor neurons from Tg G59S mice was an increase in the size and complexity of degradative organelles as compared with age-matched controls (Fig. 4 A–D). Lipofuscin granules are a mature form of tertiary lysosome that proliferates with age ( 10 ). Quantification of this observation revealed a significant increase in the number of lipofuscin granules compared with non-transgenic littermate controls ( t -test, P < 0.002; Fig. 4 E). In addition, there is an increase in the average size of the granules compared with those found in cell bodies of motor neurons from age-matched controls, although the difference did not reach statistical significance ( t -test, P < 0.07; Fig. 4 F). Both the increased number and increased size of lipofuscin granules in Tg G59S mice results in an almost two-fold increase in the total cell area occupied by these organelles.
In addition to their increased size, the lipofuscin granules in the Tg G59S animals are more likely to cluster, display greater heterogeneity, and contain greater densities of lamellae (Fig. 4 B and D), suggesting that these lipofuscin granules are more mature ( 10 ) than those in age-matched controls (Fig. 4 A and C). Other organelles in the degradative pathway, such as multivesicular bodies and autophagosomes, were observed in both Tg G59S and controls. However, no change was observed in either the number or the complexity of autophagosomes or multivesicular bodies in Tg G59S mice as compared with age-matched controls.
Cellular studies on the G59S mutation suggested that both loss-of-function and toxic gain-of-function due to protein misfolding/aggregation may both contribute to the observed phenotype in patients expressing the dynactin mutation. The increased number, size, and complexity of lipofuscin granules, and the lack of an effect on axonal transport argue that the gain-of-function mechanism may be most relevant in this model. To further address this question, we compared the electron microscope (EM) analysis of Tg G59S mice with a similar analysis of Loa mice, which express a hypomorphic allele of cytoplasmic DHC ( 6 ). As shown in Fig. 4 G, analysis of motor neurons from heterozygous Loa/+ mice indicates that there is not a significant increase in the number of lipofuscin granules per motor neuron in this line, as compared with age-matched +/+ controls. In striking contrast to the increased size of these granules seen in the Tg G59S mice as compared with littermate controls, in Loa/+ heterozygous mice we see a statistically significant decrease in granule area ( t -test, P < 0.001), again as compared with age-matched littermate controls. Together, these results suggest that a loss-of-function mutation in dynein has a differential effect on cellular degradative pathways than the apparent gain-of-function effect seen in the Tg G59S mouse model described here.
Reduced axonal caliber and disruption of NMJ morphology indicate distal neuronal degeneration in Tg G59S mice
Cross-sections of spinal cord from Tg G59S and non-transgenic mice were stained with hematoxylin in order to visualize neurons. No significant loss of motor neurons was observed in Tg G59S mice as compared with age-matched non-transgenic controls ( t -test, P > 0.25). Although loss of spinal cord motor neurons is not seen, there is a significant change in axonal caliber in mice expressing mutant p150 Glued (Fig. 5 A and B). In Tg G59S mice, there is a clear reduction in the relative number of large caliber motor neurons (K–S test, P « 0.001; Fig. 5 C). We did not see a significant increase in axonal demyelination in the Tg G59S model (Fig. 5 D). While not statistically significant, there is a decrease in the percentage of remyelinated axons in Tg G59S mice (Fig. 5 D), suggesting that expression of the transgene may lead to an impaired ability to remyelinate.
To further investigate the effects of expression of the G59S mutation on the integrity of the distal axon, we examined NMJ morphology in the sternomastoid and soleus muscles from age-matched transgenic and non-transgenic mice. The integrity of pre- and post-synaptic junctions were scored as intact, partially disrupted, or fragmented in 18-month-old mice (Fig. 6 ). In soleus muscle from non-transgenic control mice, a large proportion (45%) of pre-synaptic junctions were clearly intact, while 25% of the junctions were partially disrupted, and 30% were fragmented. In contrast, in NMJs from the soleus of Tg G59S mice, only 31% of pre-synaptic junctions were intact, 23% were partially disrupted, and the majority (46%) were fragmented (Fig. 6 ). Similar observations were seen in sternomastoid muscles from transgenic Tg G59S mice as compared with non-transgenic controls. In all cases, morphology of the post-synaptic junction mirrored that of the pre-synaptic junction. Also, whether intact or fragmented, all junctions examined remain innervated, as assessed by confocal microscopy of muscle preparations stained for neurofilaments.
While significant changes were observed in NMJ morphology in Tg G59S mice, we observed no significant alterations in muscle histology. Hematoxylin staining of quadriceps, soleus, and EDL muscles from 18-month-old Tg G59S mice revealed few central nuclei, regularly shaped fibers, and none of the clustered nuclei indicative of fiber drop out (data not shown). This lack of morphological alterations in muscle is consistent with the relatively mild and slowly progressive muscle weakness observed in grip strength assays, described above.
In this study we describe a mouse model of the human disease dSBMA caused by a G59S mutation in the cytoplasmic dynein interacting protein p150 Glued . These mice model the disease seen in human patients in several key respects, including the relatively mild, slowly progressive phenotype and the underlying degeneration of motor neurons.
Cytoplasmic dynein and dynactin are pleiotropic cellular motors, with roles in intracellular trafficking including ER-to-Golgi motility, endocytosis, lysosome motility, and mitochondrial motility, as well as a critical function in retrograde axonal transport. Further specific roles for dynein and/or dynactin have been proposed, including aggresome formation, autophagy, and stability of the NMJ ( 1 ). Thus, while it is clear that defects in dynein or dynactin result in neuronal degeneration, gaining a mechanistic understanding of the basis for this degeneration is not straightforward.
Further, when expressed in cell culture, the G59S allele of p150 Glued has effects indicative of both the loss of dynein/dynactin function and the toxic gain of function ( 4 ). The relative contribution of these effects to overall pathology may be deduced by comparisons among mouse models showing inhibition of dynein function, such as the Loa and Cra1 lines that express point mutations in DHC ( 6 ), as well as the previously characterized Tg dynamitin mouse ( 7 ), and the Tg G59S model described here. Pathology common to all of these models would suggest effects due to a loss of dynein/dynactin function while deficits seen only in the Tg G59S model would likely be due to toxic gain of function induced by the mutation (Fig. 7 ).
Deficits in axonal transport have been proposed to function as a critical driver of pathology in a number of human neurodegenerative diseases; significant inhibition of axonal transport has been measured in multiple animal models of motor neuron disease, including the Tg dynamitin mouse ( 7 ), the Loa/Loa mouse ( 6 ), and the SOD1 G93A mouse model of familial amyotropic lateral sclerosis ( 11 ). Here, we used ligation experiments to show that low-level expression of the G59S mutation does not induce a measurable inhibition of axonal transport.
In contrast, other signs of neuronal dysfunction are apparent in the Tg G59S mouse. Therefore it is possible that aberrant axonal transport is secondary to other cellular defects, and disruption of other functions of the dynein/dynactin complex, such as trafficking, are contributing to neuronal pathology. Alternatively, the lack of a measurable defect in axonal transport, coupled with biochemical data indicating that expression of mutant p150 Glued in this model does not disrupt the dynein/dynactin complex, suggests that the Tg G59S mouse does not display a loss-of-function phenotype.
Instead, observations from this model are consistent with a toxic gain of function due to protein misfolding leading to enhanced degradation. Specifically, we noted an increase in the number, size, and complexity of lipofuscin granules in the Tg G59S mouse, but not in mice carrying the Loa mutation, which may be considered a hypomorphic dynein allele. Further indication that toxicity caused by misfolding of mutant p150 Glued is critical to the pathogenesis of dSBMA comes from observations of pronounced aggregate formation in motor neurons in vivo seen in immunocytochemistry on tissues from an affected individual ( 5 ), and biochemical and immunocytochemical evidence for aggregate formation in cellular models ( 4 ).
Despite the lack of a clear effect on axonal transport in this model, we see evidence for distal degeneration in motor neurons, including reduced axonal caliber, and fragmentation at the NMJ. Studies in Drosophila have suggested that dynactin has a direct role in maintaining NMJ stability ( 8 ), but a specific mechanism has not been explored. In our EM analysis of motor neurons from the Tg G59S mouse, we found no indication of protein aggregates localized along neuronal processes. Given that axonal transport is unaltered in Tg G59S mice, it is unlikely that axonal strangulation due to protein aggregation along the axon is contributing to peripheral abnormalities, such as loss of axonal caliber and disruption of the NMJ. Instead, expression of misfolded mutant protein may induce both the proliferation of degradative organelles and cellular stress responses that in turn could lead to distal degeneration. It is interesting that both the possible induction of cellular stress pathways and the distal degeneration that we observe are not sufficient to cause motor neuron loss, but instead may represent a chronic response. Again, this may serve as a useful paradigm to understand pathogenesis in dSBMA patients, since this is not a lethal disease in patients that are provided appropriate palliative care.
Recent description of a knock-in model of the G59S mutation by Lai et al . ( 12 ) now allows a useful comparison of multiple models of dSBMA. Heterozygous knock-in mice exhibit a subtle phenotype with no alteration in grip strength or performance on a rotorod test, but with some mild gait abnormalities. Some motor neuron loss per section is observed in the heterozygous knock-in model, as well as apparent destabilization of the NMJ. In comparison with the results from the transgenic model reported here, we can conclude that low-level expression of the transgene in motor neurons only is sufficient to induce a relatively similar phenotype to that observed when the G59S mutation is expressed ubiquitously.
In summary, we hypothesize that multiple factors contribute to the pathogenesis in dSBMA: toxicity of the G59S p150 Glued as well as destabilization of the neuronal periphery, and that both of these pathways are independent of deficits in axonal transport. It may be that more significant defects in axonal transport may lead to a more rapidly progressive motor neuron degeneration, and in the absence of these defects, a milder and more slowly progressive phenotype is observed.
MATERIALS AND METHODS
Observed patterns of inheritance indicate that the insertion site of the transgene is on the X-chromosome in the Tg G59S (M20) mouse. Crude digests from mouse ear punches were diluted and used as template DNA for PCR. PCR was performed using the following primer sequences: 5′-CTGCTCCATCTTCAAATACC-3′ and 5′-GGACTCCAAGCCCTCAAG-3′ (Invitrogen); PCR reactions were analyzed by agarose gel electrophoresis.
Freshly dissected mouse tissue was homogenized in Trizol reagent (Invitrogen) and RNA was extracted according to the manufacturer’s specifications, then treated with RQ1 DNase (Promega) to eliminate residual DNA. RT–PCR was performed using the one-step RT–PCR kit (Qiagen), using the primers described above. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was amplified as a control using primers with the following sequences: 5′-CATGTGGGCCATGAGGTCCACCAC-3′ and 5′-TGAAGGTCGGAGTCAACGGATTTGGT-3′) (Invitrogen). Samples were analyzed by agarose gel electrophoresis. Similar levels of transgene expression were seen in males and females by this assay.
Western blot analysis
Freshly dissected brains and spinal cords from individual mice were homogenized in three volumes of buffer B (20 m m Tris–HCl, pH 7.4, 1 m m ethylene diamine tetraacedic acid, 2 m m ethylene glycol tetraacedic acid) with protease inhibitors phenylmethylsulphonyl fluoride (PMSF), leupeptin, pepstatinA, and N-P-tosyl-L-arginine methyl ester hydrochloride (TAME). Low speed supernatants were collected (10–30 min spin at 11 600× g ) and analyzed by gel electrophoresis and western blotting. Blots were probed using antibodies against dynactin subunit p150 Glued [monoclonal antibody 610474 from BD Biosciences or affinity purified polyclonal antibody UP502 ( 13 )], DIC (DIC monoclonal MAB1618 from Chemicon), DHC (DHC polyclonal R-325 from Santa Cruz), and the myc tag (monoclonal R950-25 from Invitrogen R950-25); actin (monoclonal MAB1501R from Chemicon) was used as a loading control. Detection was performed with Western Lightning (Amersham Biosciences), visualized with the LAS 3000 system, and densitometry of the bands was performed with ImageJ software. Similar levels of transgene expression, as measured by antibody reactivity to the myc-tagged transgene were seen in males and females by this assay (data not shown).
Six 18-month-old Tg G59S and six age-matched control mice were anesthetized using an IP injection of ketamine/xylazine and the sciatic nerve was exposed. Two ligations were placed midlength along the nerve, ∼1 mm apart, and the surgical site was closed. Two hours later, the sciatic nerve was removed and 2 mm nerve segments immediately proximal and distal to the ligation site were removed and gently homogenized in phosphate buffer solution containing the protease inhibitors described above. Proximal and distal nerve lysates were analyzed by gel electrophoresis and immunoblot. Bands were quantified using ImageJ software, and normalized to a GAPDH loading control (monoclonal AB9484 from Abcam). Average accumulation proximal and distal to the ligature site were compared for transgenic Tg G59S mice and non-transgenic littermate control mice using t -test statistics.
Sucrose density centrifugation
Brain homogenates from 3-month-old Tg G59S and non-transgenic control mice in PIPES HEPES EGTA magnesium chloride buffer, pH 7.4, plus protease inhibitors (PMSF, leupeptin, pepstatinA, TAME) were spun at 11 600× g for 10 min to obtain a low-speed supernatant. High-speed supernatants were generated by a second centrifugation step at 80 000× g for 10 min. High speed supernatant fractions were fractionated by centrifugation through 5–25% sucrose density gradient as previously described ( 7 ). Fractions from the gradient were analyzed by SDS–PAGE (sodium dodecyl sulfate polyacrylamide gel electrophoresis) and immunoblot, using antibodies to dynein (DHC and DIC), dynactin subunits (p150 Glued polyclonal UP502) and dynamitin (p50 monoclonal 611003 from BD Biosciences), and actin. To verify that the transgenic G59S p150 Glued myc-tagged protein was co-fractionating with the other components dynein/dynactin, the myc tag was identified on separate blots run with the same sucrose gradient samples and probed with a monoclonal antibody specific for myc.
A grip strength meter (Columbus Instruments) measured the force exerted by a mouse as it was pulled from a grid by the tail while grasping the grid with its forepaws only. An average of three trials per mouse was calculated on a monthly basis. Significance of grip strength comparisons and of body weight comparisons between genotypes were determined by t -test at each time point. Gait analyses were performed by measuring stride length of each foot. Significance between genotypes for stride lengths was determined by student’s t -test.
Spinal cords were fixed in 4% paraformaldehyde, embedded in Optimal Cutting Temperature (OCT) (Tissue-Tek), frozen on dry ice, and then cut into 40 µm sections. Free floating sections were processed for antigen retrieval by incubating in 50 m m Tris, pH 9 at 95°C for 10 min, and then incubated with primary antibodies to dynein [DIC rabbit polyclonal antibody UP1467 ( 14 )], dynactin [polyclonal antibodies UP502 to p150 Glued and UP1097 to dynamitin ( 7 )], and the myc tag (monoclonal antibody from Invitrogen), followed by fluorescently-labeled secondary antibodies. Anti-myc staining was visualized using a fluorescein MOM kit (Vector Labs). Images were acquired on a Leica confocal microscope.
For motor neuron counts in spinal cord, tissues were embedded in OCT (Tissue-Tek) and cooled on dry ice. For muscle, tissues were embedded in OCT snap frozen with isopentane cooled by liquid nitrogen. Frozen sections (10 µm) were fixed in 4% paraformaldehyde for 10 min. Harris’s hematoxylin stain (Vector) was applied for 7 min to visualize neurons and 5–10 min to visualize muscle fibers. Slides were dehydrated briefly in an ethanol series followed by a xylene series and mounted in Vectamount (Vector).
Quantification of motor neurons
Motor neuron number per section was counted in 10 µm sections of cervical spinal cord from 18 month-old mice. A minimum of five sections were counted per animal; scored sections were at least 50 µm apart. Statistical significance was determined using Student’s t-test .
EM of ventral roots and spinal cord
Eighteen month-old Tg G59S and age-matched control mice were anesthetized using an IP injection of ketamine/xylazine. Mice were perfused with 50 ml of 2% glutaraldehyde and 2% paraformaldehyde in 0.1 M cacodylate buffer. The L5 ventral roots and cervical spinal cords were dissected following perfusion. Tissues were post-fixed in 1% osmium tetroxide for 20 min, and processed for embedding in Epon-Araldite resin. Ultrathin sections were cut, mounted on an open grid, and stained with 1% uranyl acetate and 0.5% lead citrate. Grids were analyzed with a 100CX Joel Electron Microscope at 2.5–75 kV. Axonal cross-sectional areas were measured from images taken at 2.5 kV using ImageJ software. At least 10 fields were analyzed per animal and three animals were analyzed per genotype. A K–S test was performed to determine statistical significance. The number of lipofuscin granules was counted and their area was measured from images taken at 5 kV using Image J software. At least 10 fields were analyzed per animal and three animals were analyzed per genotype. Student’s t -test was used to determine statistical significance. To determine the percentage of cell area composed of lipofuscin granules, the total area of lipofuscin granules per cell was taken as a ratio to the total area of that cell; averages were determined for each genotype.
Visualization and scoring of NMJs
Sternomastoid and soleus muscles from 18-month-old Tg G59S mice and age-matched controls were dissected and stained as described ( 15 ) with minor modifications. Briefly, muscles were fixed in 4% paraformaldehyde, permeabilized, and stained with rhodamine-conjugated bungarotoxin to visualize post-synaptic AchR (Molecular Probes), antibodies against phosphorylated neurofilaments (SMI 31 monoclonal – Sternberger Monoclonals Incorporated) to visualize axons, and antibodies to SV2 (Developmental Studies Hybridoma Bank) to visualize pre-synaptic junctions. Primary antibodies were detected using fluorescein isothiocyanate-conjugated anti-mouse secondary antibodies (Jackson ImmunoResearch). Muscles were whole mounted in Vectashield mounting media (Vector). To visualize, count, and score NMJs, a Leica microscope (DMIRBE) with a 40× dry objective was used. Images were acquired with OpenLab software (Improvision) and an OrcaER camera (Hamamatsu). NMJ morphology was scored as intact, partially disrupted, or fragmented; a minimum of 150 NMJs were scored per animal. Three 18-month-old mice per genotype were scored.
The Amyotrophic Lateral Sclerosis Association and the National Institutes of Health [GM48661 to E.H.], and a NIH postdoctoral fellowship to E.C.-L.
The Loa mice used in this study were generously provided by Elizabeth Fisher, Institute of Neurology, University College London. The Thy 1.2 promoter was a kind gift from Pico Caroni, Friedrich Miescher Institute, Switzerland. The authors gratefully acknowledge Mariko K. Tokito for generation of the transgene construct, Jennifer Levy for expert assistance with confocal microscopy, Jean Richa and the Transgenic Mouse Facility at the University of Pennsylvania for creation of founder animals and Neelima Shaw and the Biomedical Core Imaging Facility for their assistance with EM studies. The SV2 antibody developed by Kathleen M. Buckley was obtained from the Developmental Studies Hybridoma Bank at the University of Iowa. We would also like to thank Michael Marks, Nicholas Gonatas, and Steven Shearer for sharing their expertise.
Conflict of Interest statement . The authors declare that they have no conflicts of interest regarding this work.